Massive migration from the steppe is a source

for Indo-European languages in Europe

Wolfgang Haak1,*, Iosif Lazaridis2,3,*, Nick Patterson3, Nadin Rohland2,3, Swapan Mallick2,3,4,
Bastien Llamas1, Guido Brandt5, Susanne Nordenfelt2,3, Eadaoin Harney2,3,4, Kristin
Stewardson2,3,4, Qiaomei Fu2,3,6,7, Alissa Mittnik8, Eszter Bánffy9,10, Christos Economou11,
Michael Francken12, Susanne Friederich13, Rafael Garrido Pena14, Fredrik Hallgren15, Valery
Khartanovich16, Aleksandr Khokhlov17, Michael Kunst18, Pavel Kuznetsov17, Harald
Meller13, Oleg Mochalov17, Vayacheslav Moiseyev16, Nicole Nicklisch5,13,19, Sandra L.
Pichler20, Roberto Risch21, Manuel A. Rojo Guerra22, Christina Roth5, Anna Szécsényi-
Nagy5,9, Joachim Wahl23, Matthias Meyer6, Johannes Krause8,12,24, Dorcas Brown25, David
Anthony25, Alan Cooper1, Kurt Werner Alt5,13,19,20 and David Reich2,3,4

*
Contributed equally to this work

1
Australian Centre for Ancient DNA, School of Earth and Environmental Sciences &
Environment Institute, University of Adelaide, Adelaide, South Australia, SA 5005,
Australia
2
Department of Genetics, Harvard Medical School, Boston, MA, 02115, USA
3
Broad Institute of Harvard and MIT, Cambridge, MA, 02142, USA
4
Howard Hughes Medical Institute, Harvard Medical School, Boston, MA, 02115, USA
5
Institute of Anthropology, Johannes Gutenberg University of Mainz, D-55128 Mainz,
Germany
6
Max Planck Institute for Evolutionary Anthropology, Leipzig, 04103, Germany
7
Key Laboratory of Vertebrate Evolution and Human Origins of Chinese Academy of
Sciences, IVPP, CAS, Beijing, 100049, China
8
Institute for Archaeological Sciences, University of Tübingen, Tübingen, 72074, Germany
9
Institute of Archaeology, Research Centre for the Humanities, Hungarian Academy of
Science, H-1014 Budapest, Hungary
10
Römisch Germanische Kommission (RGK) Frankfurt, D-60325 Frankfurt, Germany
11
Archaeological Research Laboratory, Stockholm University, 114 18, Sweden
12
Department of Paleoanthropology, Senckenberg Center for Human Evolution and
Paleoenvironment, University of Tübingen, Tübingen, D-72070, Germany
13
State Office for Heritage Management and Archaeology Saxony-Anhalt and State Heritage
Museum, D-06114 Halle, Germany
14
Departamento de Prehistoria y Arqueología, Facultad de Filosofía y Letras, Universidad
Autónoma de Madrid, E-28049 Madrid, Spain,
15
The Cultural Heritage Foundation, Västerås, 722 12, Sweden
16
Peter the Great Museum of Anthropology and Ethnography (Kunstkamera) RAS, St.
Petersburg, Russia
17
Volga State Academy of Social Sciences and Humanities, 443099 Russia, Samara, M.
Gor'kogo, 65/67
18
Deutsches Archaeologisches Institut, Abteilung Madrid, E-28002 Madrid, Spain
19
Danube Private University, A-3500 Krems, Austria

1
20
Institute for Prehistory and Archaeological Science, University of Basel, CH-4003 Basel,
Switzerland
21
Departamento de Prehistòria, Universitat Autonoma de Barcelona, E-08193 Barcelona,
Spain
22
Departamento de Prehistòria y Arqueolgia, Universidad de Valladolid, E-47002 Valladolid,
Spain
23
State Office for Cultural Heritage Management Baden-Württemberg, Osteology, Konstanz,
D-78467, Germany
24
Max Planck Institute for the Science of Human History, D-07745 Jena, Germany
25
Anthropology Department, Hartwick College, Oneonta, NY

To whom correspondence should be addressed: David Reich

([email protected])

2
We generated genome-wide data from 69 Europeans who lived between 8,000-3,000
years ago by enriching ancient DNA libraries for a target set of almost four hundred
thousand polymorphisms. Enrichment of these positions decreases the sequencing
required for genome-wide ancient DNA analysis by a median of around 250-fold,
allowing us to study an order of magnitude more individuals than previous studies1-8
and to obtain new insights about the past. We show that the populations of western and
far eastern Europe followed opposite trajectories between 8,000-5,000 years ago. At the
beginning of the Neolithic period in Europe, ~8,000-7,000 years ago, closely related
groups of early farmers appeared in Germany, Hungary, and Spain, different from
indigenous hunter-gatherers, whereas Russia was inhabited by a distinctive population
of hunter-gatherers with high affinity to a ~24,000 year old Siberian6. By ~6,000-5,000
years ago, a resurgence of hunter-gatherer ancestry had occurred throughout much of
Europe, but in Russia, the Yamnaya steppe herders of this time were descended not
only from the preceding eastern European hunter-gatherers, but from a population of
Near Eastern ancestry. Western and Eastern Europe came into contact ~4,500 years
ago, as the Late Neolithic Corded Ware people from Germany traced ~3/4 of their
ancestry to the Yamnaya, documenting a massive migration into the heartland of
Europe from its eastern periphery. This steppe ancestry persisted in all sampled central
Europeans until at least ~3,000 years ago, and is ubiquitous in present-day Europeans.
These results provide support for the theory of a steppe origin9 of at least some of the
Indo-European languages of Europe.

3
Genome-wide analysis of ancient DNA has emerged as a transformative technology for
studying prehistory, providing information that is comparable in power to archaeology and
linguistics. Realizing its promise, however, requires collecting genome-wide data from an
adequate number of individuals to characterize population changes over time, which means
not only sampling a succession of archaeological cultures2, but also multiple individuals per
culture. To make analysis of large numbers of ancient DNA samples practical, we used in-
solution hybridization capture10,11 to enrich next generation sequencing libraries for a target
set of 394,577 single nucleotide polymorphisms (SNPs) (“390k capture”), 354,212 of which
are autosomal SNPs that have also been genotyped using the Affymetrix Human Origins
array in 2,345 humans from 203 populations4,12. This reduces the amount of sequencing
required to obtain genome-wide data by a minimum of 45-fold and a median of 262-fold
(Online Table 1). This strategy allows us to report genomic scale data on more than twice the
number of ancient Eurasians as the entire preceding literature1-8 (Extended Data Table 1).

We used this technology to study population transformations in Europe. We began by

preparing 212 DNA libraries from 119 ancient samples in dedicated clean rooms, and testing
these by light shotgun sequencing and mitochondrial genome capture (SI1, Online Table 1).
We restricted to libraries with molecular signatures of authentic ancient DNA (elevated
damage in the terminal nucleotide), negligible evidence of contamination based on
mismatches to the mitochondrial consensus13, and, where available, a mitochondrial DNA
haplogroup that matched previous results using PCR4,14,15 (SI2). For 123 libraries prepared in
the presence of Uracil-DNA-glycosylase16 to reduce errors due to ancient DNA damage17,
we performed 390k capture, carried out paired end sequencing, and mapped to the human
genome. We restricted analysis to 95 libraries from 69 samples that had at least 0.06-fold
average target coverage (average of 3.8-fold), and used majority rule to call an allele at each
SNP covered at least once (Online Table 1). After combining our data (SI3) with 25 ancient
samples from the literature — three Upper Paleolithic samples from Russia1,7,6, seven people
of European hunter gatherer ancestry4,5,8,2, and fifteen European farmers2,8,4,3, — we had data
from 94 ancient Europeans. Geographically, these came from Germany (n=41), Spain (n=10),
Russia (n=14), Sweden (n=12), Hungary (n=15), Italy (n=1) and Luxembourg (n=1)
(Extended Data Table 2). Following the central European chronology, these included 19
hunter-gatherers (>5,500 BCE), 28 Early Neolithic farmers (EN: ~6,000-4,000 BCE), 11
Middle Neolithic farmers (MN: ~4,000-3,000 BCE) including the Tyrolean Iceman3, 9 Late
Copper/Early Bronze Age individuals (Yamnaya: ~3,300-2,700 BCE), 15 Late Neolithic
4
individuals (LN: ~2,500-2,200 BCE), 9 Early Bronze Age individuals (~2,200-1,500 BCE),
two Late Bronze Age individuals (~1,200-1,100 BCE) and one Iron Age individual (~900
BCE). Two individuals were excluded from analyses as they were related to others from the
same population. The average number of SNPs covered at least once was 212,375 and the
minimum was 22,869 (Fig. 1).

We determined that 34 of the 69 newly analyzed individuals were male and used 2,258 Y
chromosome SNPs targets included in the capture to obtain high resolution Y chromosome
haplogroup calls (SI4). Outside Russia, and before the Late Neolithic period, only a single
R1b individual was found (early Neolithic Spain) in the combined literature (n=70). By
contrast, haplogroups R1a and R1b were found in 60% of Late Neolithic/Bronze Age
Europeans outside Russia (n=10), and in 100% of the samples from European Russia from all
periods (7,500-2,700 BCE; n=9). R1a and R1b are the most common haplogroups in many
European populations today18,19, and our results suggest that they spread into Europe from the
East after 3,000 BCE. Two hunter-gatherers from Russia included in our study belonged to
R1a (Karelia) and R1b (Samara), the earliest documented ancient samples of either
haplogroup discovered to date. These two hunter-gatherers did not belong to the derived
lineages M417 within R1a and M269 within R1b that are predominant in Europeans
today18,19, but all 7 Yamnaya males did belong to the M269 subclade18 of haplogroup R1b.

Principal components analysis (PCA) of all ancient individuals along with 777 present-day
West Eurasians4 (Fig. 2a, SI5) replicates the positioning of present-day Europeans between
the Near East and European hunter-gatherers4,20, and the clustering of early farmers from
across Europe with present day Sardinians3,4,27, suggesting that farming expansions across the
Mediterranean to Spain and via the Danubian route to Hungary and Germany descended from
a common stock. By adding samples from later periods and additional locations, we also
observe several new patterns. All samples from Russia have affinity to the ~24,000 year old
MA16, the type specimen for the Ancient North Eurasians (ANE) who contributed to both
Europeans4 and Native Americans4,6,8. The two hunter-gatherers from Russia (Karelia in the
northwest of the country and Samara on the steppe near the Urals) form an “eastern European
hunter-gatherer” (EHG) cluster at one end of a hunter-gatherer cline across Europe; people of
hunter-gatherer ancestry from Luxembourg, Spain, and Hungary sit at the opposite “western
European hunter-gatherer4” (WHG) end, while the hunter-gatherers from Sweden4,8 (SHG)
are intermediate. Against this background of differentiated European hunter-gatherers and
5
homogeneous early farmers, multiple events transpired in all parts of Europe included in our
study. Middle Neolithic Europeans from Germany, Spain, Hungary, and Sweden from the
period ~4,000-3,000 BCE are intermediate between the earlier farmers and the WHG,
suggesting an increase of WHG ancestry throughout much of Europe. By contrast, in Russia,
the later Yamnaya steppe herders of ~3,000 BCE plot between the EHG and the Near East /
Caucasus, suggesting a decrease of EHG ancestry during the same time period. The Late
Neolithic and Bronze Age samples from Germany and Hungary2 are distinct from the
preceding Middle Neolithic and plot between them and the Yamnaya. This pattern is also
seen in ADMIXTURE analysis (Fig. 2b, SI6), which implies that the Yamnaya have ancestry
from populations related to the Caucasus and South Asia that is largely absent in 38 Early or
Middle Neolithic farmers but present in all 25 Late Neolithic or Bronze Age individuals. This
ancestry appears in Central Europe for the first time in our series with the Corded Ware
around 2,500 BCE (SI6, Fig. 2b, Extended Data Fig. 1). The Corded Ware shared elements of
material culture with steppe groups such as the Yamnaya although whether this reflects
movements of people has been contentious21. Our genetic data provide direct evidence of
migration and suggest that it was relatively sudden. The Corded Ware are genetically closest
to the Yamnaya ~2,600 kilometers away, as inferred both from PCA and ADMIXTURE (Fig.
2) and FST (0.011 ± 0.002) (Extended Data Table 3). If continuous gene flow from the east,
rather than migration, had occurred, we would expect successive cultures in Europe to
become increasingly differentiated from the Middle Neolithic, but instead, the Corded Ware
are both the earliest and most strongly differentiated from the Middle Neolithic population.

“Outgroup” f3-statistics6 (SI7), which measure shared genetic drift between a pair of
populations (Extended Data Fig. 2), support the clustering of hunter-gatherers, Early/Middle
Neolithic, and Late Neolithic/Bronze Age populations into different groups as in the PCA
(Fig. 2a). We also analyzed f4-statistics, which allow us to test whether pairs of populations
are consistent with descent from common ancestral populations, and to assess significance
using a normally distributed Z-score. Early European farmers from the Early and Middle
Neolithic were closely related but not identical. This is reflected in the fact that Loschbour
shared more alleles with post-4,000 BCE European farmers from Germany, Spain, Hungary,
Sweden, and Italy than with early farmers of Germany, Spain, and Hungary, documenting an
increase of hunter-gatherer ancestry in multiple regions of Europe during the course of the
Neolithic. The two EHG form a clade with respect to all other present-day and ancient
populations (|Z|<1.9), and MA1 shares more alleles with them (|Z|>4.7) than with other
6
ancient or modern populations, suggesting that they may be a source for the ANE ancestry in
present Europeans4,12,22 as they are geographically and temporally more proximate than
Upper Paleolithic Siberians. The Yamnaya differ from the EHG by sharing fewer alleles with
MA1 (|Z|=6.7) suggesting a dilution of ANE ancestry between 5,000-3,000 BCE on the
European steppe. This was likely due to admixture of EHG with a population related to
present-day Near Easterners, as the most negative f3-statistic in the Yamnaya (giving
unambiguous evidence of admixture) is observed when we model them as a mixture of EHG
and present-day Near Eastern populations like Armenians (Z = -6.3; SI7). The Late Neolithic
/ Bronze Age groups of central Europe share more alleles with Yamnaya than the Middle
Neolithic populations do (|Z|=12.4) and more alleles with the Middle Neolithic than the
Yamnaya do (|Z|=12.5), and have a negative f3-statistic with the Middle Neolithic and
Yamnaya as references (Z=-20.7), proving that they were descended from a mixture of the
local European populations and new migrants from the east. Moreover, the Yamnaya share
more alleles with the Corded Ware (|Z|≥3.6) than with any other Late Neolithic/ Early Bronze
age group with at least two individuals (SI7), proving that they had more eastern ancestry,
consistent with the PCA and ADMIXTURE patterns (Fig. 2).

Modeling of the ancient samples shows that while Karelia is genetically intermediate between
Loschbour and MA1, the topology that considers Karelia as a mixture of these two elements
is not the only one that can fit the data (SI8). To avoid biasing our inferences by fitting an
incorrect model, we developed new statistical methods that are substantial extensions of a
previously reported approach4, which allow us to obtain precise estimates of the proportion of
mixture in later Europeans without requiring a formal model for the relationship among the
ancestral populations. The method (SI9) is based on the idea that if a Test population has
ancestry related to reference populations Ref1, Ref2, …, RefN in proportions α1, α2, ...,αN, and
the references are themselves differentially related to a triple of outgroup populations A, B, C,
then 𝑓4 (𝑇𝑒𝑠𝑡, 𝐴; 𝐵, 𝐶) = ∑𝑁
𝑖=1 𝛼𝑖 𝑓4 (𝑅𝑒𝑓𝑖 , 𝐴; 𝐵, 𝐶). By using a large number of outgroup

populations we can fit the admixture coefficients αi and estimate mixture proportions (SI9,
Extended Data Fig. 3). Using 15 outgroups from Africa, South/East/North Asia, and the
Americas, we obtain good fits as assessed by a formal test (SI10), and estimate that the
Middle Neolithic populations of Germany and Spain have ~18-34% more WHG-related
ancestry than Early Neolithic ones and that the Late Neolithic and Early Bronze Age
populations of Germany have ~22-39% more EHG-related ancestry than the Middle

7
Neolithic ones (SI9). If we model them as mixtures of Yamnaya-related and Middle Neolithic
populations, the inferred degree of population turnover is doubled to 48-80% (SI9, SI10).

To distinguish whether a Yamnaya or an EHG source fits the data better, we added ancient
samples as outgroups (SI9). Adding any Early or Middle Neolithic farmer results in EHG-
related genetic input into Late Neolithic populations being a poor fit to the data (SI9); thus,
Late Neolithic populations have ancestry that cannot be explained by a mixture of EHG and
Middle Neolithic. When using Yamnaya instead of EHG, however, we obtain a good fit (SI9,
SI10). These results can be explained if the new genetic material that arrived in Germany was
a composite of two elements: EHG and a type of Near Eastern ancestry different from that
which was introduced by early farmers (also suggested by PCA and ADMIXTURE; Fig. 2,
SI5, SI6). We estimate that these two elements each contributed about half the ancestry each
of the Yamnaya (SI6, SI9), explaining why the population turnover inferred using Yamnaya
as a source is about twice as high compared to the undiluted EHG. The estimate of Yamnaya-
related ancestry in the Corded Ware is consistent when using either present populations or
ancient Europeans as outgroups (SI9, SI10), and is 73.1 ± 2.2% when both sets are combined
(SI10). The best proxies for ANE ancestry in Europe4 were initially Native Americans12,22,
and then the Siberian MA16, but both are geographically and temporally too remote for what
appears to be a recent migration into Europe4. We can now add three new pieces to the puzzle
of how ANE ancestry was transmitted to Europe: first the EHG, then the Yamnaya formed by
mixture between EHG and a Near Eastern related population, and then the Corded Ware who
were formed by a mixture of the Yamnaya with Middle Neolithic Europeans. We caution that
the sampled Yamnaya individuals from Samara might not be directly ancestral to Corded
Ware individuals from Germany. It is possible that a more western Yamnaya population, or
an earlier (pre-Yamnaya) steppe population may have migrated into central Europe, and
future work may uncover more missing links in the chain of transmission of steppe ancestry.

By extending our model to a three way mixture of WHG, Early Neolithic and Yamnaya, we
estimate that the ancestry of the Corded Ware was 79% Yamnaya-like, 4% WHG, and 17%
Early Neolithic (Fig. 3). A small contribution of the first farmers is also consistent with
uniparentally inherited DNA: e.g. mitochondrial DNA haplogroup N1a and Y chromosome
haplogroup G2a, common in early central European farmers14,23, almost disappear during the
Late Neolithic and Bronze Age, when they are largely replaced by Y haplogroups R1a and
R1b (SI4) and mtDNA haplogroups I, T1, U2, U4, U5a, W, and subtypes of H14,23,24 (SI2).
8
The uniparental data not only confirm a link to the steppe populations but also suggest that
both sexes participated in the migrations (SI2, SI4, Extended Data Table 3). The magnitude
of the population turnover that occurred becomes even more evident if one considers the fact
that the steppe migrants may well have mixed with eastern European agriculturalists on their
way to central Europe. Thus, we cannot exclude a scenario in which the Corded Ware
arriving in today’s Germany had no ancestry at all from local populations.

Our results support a view of European pre-history punctuated by two major migrations: first,
the arrival of first farmers during the Early Neolithic from the Near East, and second of
Yamnaya pastoralists during the Late Neolithic from the steppe (Extended Data Fig. 5). Our
data further show that both migrations were followed by resurgences of the previous
inhabitants: first, during the Middle Neolithic, when hunter-gatherer ancestry rose again after
its Early Neolithic decline, and then between the Late Neolithic and the present, when farmer
and hunter-gatherer ancestry rose after its Late Neolithic decline. This second resurgence
must have started during the Late Neolithic/Bronze Age period itself, as the Bell Beaker and
Unetice groups had reduced Yamnaya ancestry compared to the earlier Corded Ware, and
comparable levels to that in some present-day Europeans (Fig. 3). Today, Yamnaya related
ancestry is lower in southern Europe and higher in northern Europe. Further data are needed
to determine whether the steppe ancestry arrived in southern Europe at the time of the Late
Neolithic / Bronze Age, or is due to migrations in historical times from northern Europe25,26.

Our results provide new data relevant to debates on the origin and expansion of Indo-
European languages in Europe (SI11). Although ancient DNA is silent on the question of the
languages spoken by preliterate populations, it does carry evidence about processes of
migration which are invoked by theories on Indo-European language dispersals. Such
theories make predictions about movements of people to account for the spread of languages
and material culture. The technology of ancient DNA makes it possible to reject or confirm
the proposed migratory movements, as well as to identify new movements that were not
previously known. The best argument for the “Anatolian hypothesis27” that Indo-European
languages arrived in Europe from Anatolia ~8,500 years ago is that major language
replacements are thought to require major migrations, and that after the Early Neolithic when
farmers established themselves in Europe, the population base was likely to have been so
large as to be impervious to subsequent turnover27,28. However, our study shows that a later
major turnover did occur, and that steppe migrants replaced ~3/4 of the ancestry of central
9
Europeans. An alternative theory is the “Steppe hypothesis”, which proposes that early Indo-
European speakers were pastoralists of the grasslands north of the Black and Caspian Seas,
and that their languages spread into Europe after the invention of wheeled vehicles9. Our
results make a compelling case for the steppe as a source of at least some of the Indo-
European languages in Europe by documenting a massive migration ~4,500 years ago
associated with the Yamnaya and Corded Ware cultures, which are identified by proponents
of the Steppe hypothesis as vectors for the spread of Indo-European languages into Europe.
These results challenge the Anatolian hypothesis by showing that not all Indo-European
languages in Europe can plausibly derive from the first farmer migrations thousands of years
earlier (SI11). We caution that the location of the Proto-Indo-European9,27,29,30 homeland that
also gave rise to the Indo-European languages of Asia, as well as the Indo-European
languages of southeastern Europe, cannot be determined from the data reported here (SI11).
Studying the mixture in the Yamnaya themselves, and understanding the genetic relationships
among a broader set of ancient and present-day Indo-European speakers, may lead to new
insight about the shared homeland.

10
Acknowledgments
We thank Peter Bellwood, Joachim Burger, Paul Heggarty, Mark Lipson, Colin Renfrew,
Jared Diamond, Svante Pääbo, Ron Pinhasi and Pontus Skoglund for critical comments. We
thank Svante Pääbo for support for establishing the ancient DNA facilities in Boston, and
Pontus Skoglund for detecting the presence of two related individuals in our dataset. We
thank Ludovic Orlando, Thorfinn S. Korneliussen, and Cristina Gamba for help in obtaining
data. We thank Agilent Technologies and Götz Frommer for help in developing the capture
reagents. We thank Clio Der Sarkissian, Guido Valverde, Luka Papac, and Birgit Nickel for
wet lab support. We thank archaeologists Veit Dresely, Robert Ganslmeier, Oleg Balanvosky,
José Ignacio Royo Guillén, Anett Osztás, Vera Majerik, Tibor Paluch, Krisztina Somogyi and
Vanda Voicsek for sharing samples and discussion about archaeological context. This
research was supported by an Australian Research Council grant to W.H. and B.L.
(DP130102158), and German Research Foundation grants to K.W.A. (Al 287/7-1 and 7-3, Al
287/10-1 and Al 287/14-1) and to H.M. (Me 3245/1-1 and 1-3). D.R. was supported by U.S.
National Science Foundation HOMINID grant BCS-1032255, U.S. National Institutes of
Health grant GM100233, and the Howard Hughes Medical Institute.

Author contributions
WH, NP, NR, JK, KWA and DR supervised the study. WH, EB, CE, MF, SF, RGP, FH, VK,
AK, MK, PK, HM, OM, VM, NN, SP, RR, MARG, CR, ASN, JW, JKr, DB, DA, AC, KWA
and DR assembled archaeological material, WH, IL, NP, NR, SM, AM and DR analyzed
genetic data. WH, NR, BL, GB, SN, EH, KS and AM performed wet laboratory ancient DNA
work. NR, QF, MM and DR developed the 390k capture reagent. WH, IL and DR wrote the
manuscript with help from all co-authors.

11
References

1. Fu, Q. et al. Genome sequence of a 45,000-year-old modern human from western

Siberia. Nature 514, 445-449, (2014).
2. Gamba, C. et al. Genome flux and stasis in a five millennium transect of European
prehistory. Nat. Commun. 5, 5257 (2014).
3. Keller, A. et al. New insights into the Tyrolean Iceman's origin and phenotype as
inferred by whole-genome sequencing. Nat. Commun. 3, 698, (2012).
4. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for
present-day Europeans. Nature 513, 409-413, (2014).
5. Olalde, I. et al. Derived immune and ancestral pigmentation alleles in a 7,000-year-
old Mesolithic European. Nature 507, 225-228, (2014).
6. Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of
Native Americans. Nature 505, 87-91, (2014).
7. Seguin-Orlando, A. et al. Genomic structure in Europeans dating back at least 36,200
years. Science 346, 1113-1118, (2014).
8. Skoglund, P. et al. Genomic Diversity and Admixture Differs for Stone-Age
Scandinavian Foragers and Farmers. Science 344, 747-750, (2014).
9. Anthony, D. W. The Horse, the Wheel, and Language: How Bronze-Age Riders from
the Eurasian Steppes Shaped the Modern World. (Princeton University Press, 2007).
10. Fu, Q. et al. DNA analysis of an early modern human from Tianyuan Cave, China.
Proc. Natl. Acad. Sci. USA 110, 2223–2227, (2013).
11. Rohland, N., Harney, E., Mallick, S., Nordenfelt, S. & Reich, D. Partial uracil–DNA–
glycosylase treatment for screening of ancient DNA. Philosophical Transactions of
the Royal Society of London B: Biological Sciences 370, (2014).
12. Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065-1093,
(2012).
13. Fu, Q. et al. A Revised Timescale for Human Evolution Based on Ancient
Mitochondrial Genomes. Curr. Biol. 23, 553-559, (2013).
14. Brandt, G. et al. Ancient DNA reveals key stages in the formation of central European
mitochondrial genetic diversity. Science 342, 257-261, (2013).
15. Der Sarkissian, C. et al. Ancient DNA reveals prehistoric gene-flow from Siberia in
the complex human population history of North East Europe. PLoS Genet. 9,
e1003296, (2013).
16. Briggs, A. W. et al. Removal of deaminated cytosines and detection of in vivo
methylation in ancient DNA. Nucleic Acids Res. 38, e87-e87, (2010).
17. Briggs, A. W. et al. Patterns of damage in genomic DNA sequences from a
Neandertal. Proceedings of the National Academy of Sciences 104, 14616-14621,
(2007).
18. Myres, N. M. et al. A major Y-chromosome haplogroup R1b Holocene era founder
effect in Central and Western Europe. Eur. J. Hum. Genet. 19, 95-101, (2011).
19. Underhill, P. A. et al. The phylogenetic and geographic structure of Y-chromosome
haplogroup R1a. Eur. J. Hum. Genet., (2014).
20. Skoglund, P. et al. Origins and genetic legacy of Neolithic farmers and hunter-
gatherers in Europe. Science 336, 466-469, (2012).
21. Czebreszuk, J. in Ancient Europe, 8000 B.C. to A.D. 1000: Encyclopedia of the
Barbarian World (eds Bogucki PI and & Crabtree PJ) 467-475 (Charles Scribners &
Sons, 2003).
22. Lipson, M. et al. Efficient moment-based inference of admixture parameters and
sources of gene flow. Mol. Biol. Evol. 30, 1788-1802, (2013).

12
23. Szécsényi-Nagy, A. et al. Tracing the genetic origin of Europe's first farmers reveals
insights into their social organization. bioRxiv, DOI: 10.1101/008664, (2014).
24. Haak, W. et al. Ancient DNA from European early Neolithic farmers reveals their
Near Eastern affinities. PLoS Biol. 8, e1000536, (2010).
25. Hellenthal, G. et al. A genetic atlas of human admixture history. Science 343, 747-
751, (2014).
26. Ralph, P. & Coop, G. The geography of recent genetic ancestry across Europe. PLoS
Biol. 11, e1001555, (2013).
27. Renfrew, C. Archaeology and Language: The Puzzle of Indo-European Origins.
(Pimlico, 1987).
28. Bellwood, P. First Farmers: The Origins of Agricultural Societies. (Wiley-
Blackwell, 2004).
29. Gamkrelidze, T. V. & Ivanov, V. V. The Early History of Indo-European Languages.
Sci. Am. 262, 110-116, (1990).
30. Mallory, J. P. In Search of the Indo-Europeans: Language, Archaeology and Myth.
(Thames and Hudson, 1991).

13
Figure Legends

Figure 1: Location and SNP coverage of samples included in this study. (a) Geographic
location and time-scale (central European chronology) of the 69 newly typed ancient
individuals from this study (black outline) and 25 from the literature for which shotgun
sequencing data was available (no outline). (b) Number of SNPs covered at least once in the
analysis dataset of 94 individuals.

Figure 2: Population transformations in Europe. (a) PCA analysis, (b) ADMIXTURE

analysis. The full ADMIXTURE analysis including present-day humans is shown in
Extended Data Fig. 1.

Figure 3: Admixture proportions. We estimate mixture proportions using a method that

gives unbiased estimates even without an accurate model for the relationships between the
test populations and the outgroup populations (SI9). Population samples are grouped
according to chronology (ancient) and Yamnaya ancestry (present-day humans).

14
Online Methods

Screening of libraries (shotgun sequencing and mitochondrial capture)

The 212 libraries screened in this study (SI1) from a total of 119 samples (SI3) were
produced at Adelaide (n=151), Tübingen (n=16), and Boston (n=45) (Online Table 1).

The libraries from Adelaide and Boston had internal barcodes directly attached to both sides
of the molecules from the DNA extract so that each sequence begins with the barcode10. The
Adelaide libraries had 5 base pair (bp) barcodes on both sides, while the Boston libraries had
7 bp barcodes. Libraries from Tübingen had no internal barcodes, but were differentiated by
the sequence of the indexing primer31.

We adapted a reported protocol for enriching for mitochondrial DNA10, with the difference
that we adjusted the blocking oligonucleotides and PCR primers to fit our libraries with
shorter adapters. Over the course of this project, we also lowered the hybridization
temperature from 65°C to 60°C and performed stringent washes at 55°C instead of 60°C32.

We used an aliquot of approximately 500ng of each library for target enrichment of the
complete mitochondrial genome in two consecutive rounds32, using a bait set for human
mtDNA32. We performed enrichment in 96-well plates with one library per well, and used a
liquid handler (Evolution P3, Perkin Elmer) for the capture and washing steps33. We used
blocking oligonucleotides in hybridization appropriate to the adapters of the truncated
libraries. After either of the two enrichment rounds, we amplified the enriched library
molecules with the primer pair that keeps the adapters short (PreHyb) using Herculase Fusion
II PCR Polymerase. We performed an indexing PCR of the final capture product using one or
two indexing primers31. We cleaned up all PCR’s using SPRI technology34 and the liquid
handler. Libraries from Tübingen were amplified with the primer pair IS5/IS631.

For libraries from Boston and Adelaide, we used a second aliquot of each library for shotgun
sequencing after performing an indexing PCR31. We used unique index combinations for
each library and experiment, allowing us to distinguish shotgun sequencing and
mitochondrial DNA capture data, even if both experiments were in the same sequencing run.
We sequenced shotgun libraries and mtDNA captured libraries from Tübingen in independent
sequencing runs since the index was already attached at the library preparation stage.

15
We quantified the sequencing pool with the BioAnalyzer (Agilent) and/or the KAPA Library
Quantification kit (KAPA biosystems) and sequenced on Illumina MiSeq, HiSeq2500 or
NextSeq500 instruments for 2×75, 2×100 or 2×150 cycles along with the indexing read(s).

Enrichment for 394,577 SNP targets (“390k capture”)

The protocol for enrichment for SNP targets was similar to the mitochondrial DNA capture,
with the exception that we used another bait set (390k) and about twice as much library (up to
1000ng) compared to the mtDNA capture.

The specific capture reagent used in this study is described for the first time here. To target
each SNP, we used a different oligonucleotide probe design compared to ref. 1. We used four
52 base pair probes for each SNP target. One probe ends just before the SNP, and one starts
just after. Two probes are centered on the SNP, and are identical except for having the
alternate alleles. This probe design avoids systematic bias toward one SNP allele or another.
For the template sequence for designing the San and Yoruba panels baits, we used the
sequence that was submitted for these same SNPs during the design of the Affymetrix Human
Origins SNP array12. For SNPs that were both in the San and Yoruba panels, we used the
Yoruba template sequence in preference. For all other SNPs, we used the human genome
reference sequence as a template. Online Table 2 gives the list of SNPs that we targeted,
along with details of the probes used. The breakdown of SNPs into different classes is:

124,106 “Yoruba SNPs”: All SNPs in “Panel 5” of the Affymetrix Human Origins array
(discovered as heterozygous in a Yoruba male: HGDP00927)12 that passed the
probe design criteria specified in ref. 11.

146,135 “San SNPs”: All SNPs in “Panel 4” of the Affymetrix Human Origins array
(discovered as heterozygous in a San male: HGDP01029)12 that passed probe
design criteria11. The full Affymetrix Human Origins array Panel 4 contains
several tens of thousands of additional SNPs overlapping those from Panel 5, but
we did not wish to redundantly capture Panel 5 SNPs.

98,166 “Compatibility SNPs”: SNPs that overlap between the Affymetrix Human Origins
the Affymetrix 6.0, and the Illumina 610 Quad arrays, which are not already
16
 included in the “Yoruba SNPs” or “San SNPs” lists12 and that also passed the
probe design design criteria11.

26,170 “Miscellaneous SNPs”: SNPs that did not overlap the Human Origins array. The
subset analyzed in this study were 2,258 Y chromosome SNPs
(http://isogg.org/tree/ISOGG_YDNA_SNP_Index.html) that we used for Y
haplogroup determination.

Processing of sequencing data

We restricted analysis to read pairs that passed quality control according to the Illumina
software (“PF reads”).

We assigned read pairs to libraries by searching for matches to the expected index and
barcode sequences (if present, as for the Adelaide and Boston libraries). We allowed no more
than 1 mismatch per index or barcode, and zero mismatches if there was ambiguity in
sequence assignment or if barcodes of 5 bp length were used (Adelaide libraries).

We used Seqprep (https://github.com/jstjohn/SeqPrep) to search for overlapping sequence

between the forward and reverse read, and restricted to molecules where we could identify a
minimum of 15 bp of overlap. We collapsed the two reads into a single sequence, using the
consensus nucleotide if both reads agreed, and the read with higher base quality in the case of
disagreement. For each merged nucleotide, we assigned the base quality to be the higher of
the two reads. We further used Seqprep to search for the expected adapter sequences at either
ends of the merged sequence, and to produce a trimmed sequence for alignment.

We mapped all sequences using BWA-0.6.135. For mitochondrial analysis we mapped to the
mitochondrial genome RSRS36. For whole genome analysis we mapped to the human
reference genome hg19. We restricted all analyses to sequences that had a mapping quality of
MAPQ≥37.

We sorted all mapped sequences by position, and used a custom script to search for mapped
sequences that had the same orientation and start and stop positions. We stripped all but one
of these sequences (keeping the best quality one) as duplicates.

17
Mitochondrial sequence analysis and assessment of ancient DNA authenticity
For each library for which we had average coverage of the mitochondrial genome of at least
10-fold after removal of duplicated molecules, we built a mitochondrial consensus sequence,
assigning haplogroups for each library as described in SI2.

We used contamMix-1.0.9 to search for evidence of contamination in the mitochondrial

DNA13. This software estimates the fraction of mitochondrial DNA sequences that match the
consensus more closely than a comparison set of 311 worldwide mitochondrial genomes.
This is done by taking the consensus sequence of reads aligning to the RSRS mitochondrial
genome, and requiring a minimum coverage of 5 after filtering bases where the quality was
<30. Raw reads are then realigned to this consensus. In addition, the consensus is multiply
aligned with the other 311 mitochondrial genomes using kalign (2.0.4)37 to build the
necessary inputs for contamMix, trimming the first and last 5 bases of every read to mitigate
against the confounding factor of ancient damage. This software had difficulty running on
datasets with higher coverage, and for these datasets, we down-sampled to 50,000 reads,
which we found produced adequate contamination estimation.

For all sequences mapping to the mitochondrial DNA that had a cytosine at the terminal
nucleotide, we measured the proportion of sequences with a thymine at that position. For
population genetic analysis, we only used partially UDG-treated libraries with a minimum of
33
3% C→T substitutions as recommended by ref. . In cases where we used a fully UDG-
treated library for 390k analysis, we examined mitochondrial capture data from a non-UDG-
treated library made from the same extract, and verified that the non-UDG library had a
38
minimum of 10% C→T at the first nucleotide as recommended by ref. . Metrics for the
mitochondrial DNA analysis on each library are given in Online Table 1.

390k capture, sequence analysis and quality control

For 390k analysis, we restricted to reads that not only mapped to the human reference
genome hg19 but that also overlapped the 354,212 autosomal SNPs genotyped on the Human
Origins array4. We trimmed the last two nucleotides from each sequence because we found
that these are highly enriched in ancient DNA damage even for UDG-treated libraries. We
further restricted analyses to sites with base quality≥30.

18
We made no attempt to determine a diploid genotype at each SNP in each sample. Instead,
we used a single allele – randomly drawn from the two alleles in the individual – to represent
the individual at that site20,39. Specifically, we made an allele call at each target SNP using
majority rule over all sequences overlapping the SNP. When each of the possible alleles was
supported by an equal number of sequences, we picked an allele at random. We set the allele
to “no call” for SNPs at which there was no read coverage.

We restricted population genetic analysis to libraries with a minimum of 0.06-fold average

coverage on the 390k SNP targets, and for which there was an unambiguous sex
determination based on the ratio of X to Y chromosome reads (SI4) (Online Table 1). For
individuals for whom there were multiple libraries per sample, we performed a series of
quality control analysis. First, we used the ADMIXTURE software40,41 in supervised mode,
using Kharia, Onge, Karitiana, Han, French, Mbuti, Ulchi and Eskimo as reference
populations. We visually inspected the inferred ancestry components in each individual, and
removed individuals with evidence of heterogeneity in inferred ancestry components across
libraries. For all possible pairs of libraries for each sample, we also computed statistics of the
form D(Library1, Library2; Probe, Mbuti), where Probe is any of a panel of the same set of
eight reference populations), to determine whether there was significant evidence of the
Probe population being more closely related to one library from an ancient individual than
another library from that same individual. None of the individuals that we used had strong
evidence of ancestry heterogeneity across libraries. For samples passing quality control for
which there were multiple libraries per sample, we merged the sequences into a single BAM.

We called alleles on each merged BAM using the same procedure as for the individual
libraries. We used ADMIXTURE41 as well as PCA as implemented in EIGENSOFT42 (using
the lsqproject: YES option to project the ancient samples) to visualize the genetic
relationships of each set of samples with the same culture label with respect to 777 diverse
present-day West Eurasians4. We visually identified outlier individuals, and renamed them
for analysis either as outliers or by the name of the site at which they were sampled
(Extended Data Table 1). We also identified two pairs of related individuals based on the
proportion of sites covered in pairs of ancient samples from the same population that had
identical allele calls using PLINK43. From each pair of related individuals, we kept the one
with the most SNPs.

19
Population genetic analyses
We determined genetic sex using the ratio of X and Y chromosome alignments44 (SI4), and
mitochondrial haplogroup for all samples (SI2), and Y chromosome haplogroup for the male
samples (SI4). We studied population structure (SI5, SI6). We used f-statistics to carry out
formal tests of population relationships (SI6) and built explicit models of population history
consistent with the data (SI7). We estimated mixture proportions in a way that was robust to
uncertainty about the exact population history that applied (SI8). We estimated the minimum
number of streams of migration into Europe needed to explain the data (SI9, SI10). The
estimated mixture proportions shown in Fig. 3 were obtained using the lsqlin function of
Matlab and the optimization method described in SI9 with 15 world outgroups.

31. Kircher, M., Sawyer, S. & Meyer, M. Double indexing overcomes inaccuracies in
multiplex sequencing on the Illumina platform. Nucleic Acids Res. 40, e3-e3, (2012).
32. Meyer, M. et al. A mitochondrial genome sequence of a hominin from Sima de los
Huesos. Nature 505, 403-406, (2014).
33. Rohland, N., Harney, E., Mallick, S., Nordenfelt, S. & Reich, D. Partial uracil–DNA–
glycosylase treatment for screening of ancient DNA. Philosophical Transactions of
the Royal Society of London B: Biological Sciences 370, DOI:
10.1098/rstb.2013.0624 (2014).
34. Rohland, N. & Reich, D. Cost-effective, high-throughput DNA sequencing libraries
for multiplexed target capture. Genome Res. 22, 939-946, (2012).
35. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler
transform. Bioinformatics 25, 1754-1760, (2009).
36. Behar, D. M. et al. A “Copernican” Reassessment of the Human Mitochondrial DNA
Tree from its Root. Am. J. Hum. Genet. 90, 675-684, (2012).
37. Lassmann, T. & Sonnhammer, E. L. L. Kalign – an accurate and fast multiple
sequence alignment algorithm. BMC Bioinformatics 6, 298-298, (2005).
38. Sawyer, S., Krause, J., Guschanski, K., Savolainen, V. & Pääbo, S. Temporal patterns
of nucleotide misincorporations and DNA fragmentation in ancient DNA. PLoS ONE
7, e34131, (2012).
39. Green, R. E. et al. A Draft Sequence of the Neandertal Genome. Science 328, 710-
722, (2010).
40. Alexander, D. & Lange, K. Enhancements to the ADMIXTURE algorithm for
individual ancestry estimation. BMC Bioinformatics 12, 246, (2011).
41. Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry
in unrelated individuals. Genome Res. 19, 1655-1664, (2009).
42. Reich, D., Price, A. L. & Patterson, N. Principal component analysis of genetic data.
Nat. Genet. 40, 491-492, (2008).
43. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-
based linkage analyses. Am. J. Hum. Genet. 81, 559-575, (2007).
44. Skoglund, P., Storå, J., Götherström, A. & Jakobsson, M. Accurate sex identification
of ancient human remains using DNA shotgun sequencing. J. Archaeol. Sci. 40, 4477-
4482, (2013).

20
Figure 1

21
Figure 2

22
Figure 3

23
Extended Data Table 1: Number of ancient Eurasian modern human samples screened in genome-wide studies to date.
Only studies that produced at least one sample at ≥0.05× coverage are listed.

First author Description No. samples at ≥0.05× No. samples at >0.25×

coverage (enough for coverage (enough to
Procrustes analysis) analyze in pairs)
3
Keller Tyrolean Iceman 1 1
6
Raghavan Upper Paleolithic Siberians 2 1
5
Olalde Mesolithic Iberian from LaBrana 1 1
8
Skoglund Farmers and hunter-gatherers from Sweden 5 2
4
Lazaridis Early European farmer from Germany & Mesolithic hunter-gatherers from 7 4
Luxembourg and Sweden
2
Gamba Neolithic, Bronze Age, Iron Age Hungary 13 9
1
Fu Upper Paleolithic Siberian from Ust-Ishim 1 1
7
Seguin-Orlando Upper Paleolithic European from Kostenki 1 1
Total before study 31 20

This study Hunter-gatherers and pastoralists from Russia, Mesolithic hunter-gatherers from 69 58
Sweden, Early Neolithic from Germany, Hungary, and Spain, Middle Neolithic
from Germany & Spain, Late Neolithic / Bronze Age from Germany

24
Extended Data Table 2: The 69 samples newly reported on in this study. Samples with direct radiocarbon dates are indicated by a calibrated date “cal
BCE” along with associated laboratory numbers. Dates that are estimated based on faunal elements associated with the samples are not indicated with “cal”
(although they are still calibrated, absolute dates).
Reich ID Pop Label for Analysis Culture Group Location and sample details (e.g. sample, grave and museum ID) Date (lab no.) Country Sex mt-hg Y-hg Autosomal SNPs
I0061 Karelia_HG Russian Mesolithic EHG Yuzhnyy Oleni Ostrov, Karelia, Russia; UzOO74, grave 142, MAE RAS 5773-74 5500-5000 BCE Russia M C1g (formerly C1f) R1a1 341554
I0124 Samara_HG Russian Neolithic HG EHG Sok River, Samara, Russia; SVP44 5650-5555 cal BCE (Beta – 392490) Russia M U5a1d R1b1a 206748
I0011 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 1 5898-5531 cal BCE Sweden F U5a1 228271
I0012 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 2 5898-5531 cal BCE Sweden M U2e1 I2c2 292853
I0013 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 3 5898-5531 cal BCE Sweden M U5a1 I2a1b 251108
I0014 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 4 5898-5531 cal BCE Sweden F U5a2d 311299
I0015 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 6 5898-5531 cal BCE Sweden M U5a2d I2a1 285307
I0016 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 9 5898-5531 cal BCE Sweden M U5a2 I2a1 275233
I0017 Motala_HG Swedish Mesolithic SHG Motala, Sweden; Motala 12 5898-5531 cal BCE Sweden M U2e1 I2a1b 337794
I0174 Starcevo_EN Starcevo EN Alsónyék-Bátaszék, Mérnöki telep, Hungary; BAM25a, feature 1532 5710-5550 cal BCE (MAMS 11939 ) Hungary M N1a1a1b H2 101653
I0176 LBKT_EN LBKT EN Szemely-Hegyes, Hungary: SZEH4b, feature 1001 5210-4940 cal BCE (Beta - 310038) Hungary N1a1a1a3 30718
I0046 LBK_EN LBK EN Halberstadt-Sonntagsfeld, Germany: HAL5, grave 2, feature 241.1 5206-5004 cal BCE (MAMS 21479) Germany F T2c1d'e'f 266764
I0048 LBK_EN LBK EN Halberstadt-Sonntagsfeld, Germany: HAL25, grave 28, feature 861 5206-5052 cal BCE (MAMS 21482) Germany M K1a G2a2a 123828
I0056 LBK_EN LBK EN Halberstadt-Sonntagsfeld, Germany; HAL14, grave 15, feature 430 5206-5052 cal BCE (MAMS 21480) Germany M T2b(8) G2a2a 136578
I0057 LBK_EN LBK EN Halberstadt-Sonntagsfeld; HAL34, grave 38, feature 992 5207-5067 cal BCE (MAMS 21483) Germany F N1a1a1 55802
I0100 LBK_EN LBK EN Halberstadt-Sonntagsfeld; HAL4, grave 1, feature 139 5032-4946 cal BCE (KIA40341) Germany F N1a1a1a 342342
I0659 LBK_EN LBK EN Halberstadt-Sonntagsfeld, Germany; HAL2, grave 35, feature 999 5079-4997 cal BCE (KIA40350) Germany M N1a1a G2a2a1 191007
5066-4979 cal BCE (KIA30408)
0821 LBK_EN LBK EN Halberstadt-Sonntagsfeld, Germany; HAL24, grave 27, feature 867 5034-4942 cal BCE (KIA40348) Germany M Pre-X2d1 G2a2a1 55914
I0795 LBK_EN LBK EN Karsdorf, Germany; KAR6a, feature 170 5207-5070 cal BCE (MAMS 22823) Germany M H1 T1a 47804
I0054 LBK_EN LBK EN Oberwiederstedt-Unterwiederstedt, UWS4, Germany, grave 6, feature 1 14 5209-5070 cal BCE (MAMS 21485) Germany F J1c17 337625
I0022 LBK_EN LBK EN Viesenhäuser Hof, Stuttgart-Mühlhausen, Germany; LBK1976 5500-4800 BCE Germany F T2e 160852
I0025 LBK_EN LBK EN Viesenhäuser Hof, Stuttgart-Mühlhausen, Germany; LBK1992 5500-4800 BCE Germany F T2b 307686
I0026 LBK_EN LBK EN Viesenhäuser Hof, Stuttgart-Mühlhausen, Germany; LBK2155 5500-4800 BCE Germany F T2b 315484
I0409 Spain_EN Els_Trocs EN Els Trocs, Spain; Troc1 5311-5218 cal BCE (MAMS 16159) Spain F J1c3 172903
I0410 Spain_EN Els_Trocs EN Els Trocs, Spain; Troc3 5178-5066 cal BCE (MAMS 16161) Spain M pre-T2c1d2 R1b1 297595
I0411 Spain_EN_relative_of_I0410 Els_Trocs EN Els Trocs, Spain; Troc4 5177-5068 cal BCE (MAMS 16162) Spain M K1a2a F* 31507
I0412 Spain_EN Els_Trocs EN Els Trocs, Spain; Troc 5 5310-5206 cal BCE (MAMS 16164) Spain M N1a1a1 I2a1b1 333940
I0413 Spain_EN Els_Trocs EN Els Trocs, Spain; Troc7 5303-5204 cal BCE (MAMS 16166) Spain F V 295844
I0405 Spain_MN La_Mina MN La Mina, Spain; Mina3 3900-3600 BCE Spain M K1a1b1 I2a1a1/H2? 133230
I0406 Spain_MN La_Mina MN La Mina, Spain; Mina4 3900-3600 BCE Spain M H1 I2a2a1 324169
I0407 Spain_MN La_Mina MN La Mina, Spain; Mina6b 3900-3600 BCE Spain F K1b1a1 236225
I0408 Spain_MN La_Mina MN La Mina, Spain; Mina18a 3900-3600 BCE Spain F pre-U5b1i 321761
I0172 Esperstedt_MN Salzmünde/Bernburg MN Esperstedt, Germany; ESP24, feature 1841 3360-3086 cal BCE (Erl8699) Germany M T2b I2a1b1a 279147
I0559 Baalberge_MN Baalberge MN Quedlinburg, Germany; QLB15D, feature 21033 3645-3537 cal BCE (MAMS 22818) Germany M HV6’17 R*? 64304
I0560 Baalberge_MN Baalberge MN Quedlinburg, Germany; QLB18A, feature 21039 3640-3510 cal BCE (Er7856) Germany F T2e1 133305
I0807 Baalberge_MN Baalberge MN Esperstedt, Germany; ESP30, feature 6220 3887-3797 cal BCE (Er7784) Germany M H1e1a F* 33481
I0231 Yamnaya Yamnaya EBA Ekaterinovka,_Southern Steppe, Samara , Russia, SVP3 2910-2875 cal BCE (Beta 392487) Russia M U4a1 R1b1a2a2 348142
I0357 Yamnaya Yamnaya EBA Lopatino I, Sok_River, Samara, Russia; SVP5 same sample as SVP37 3090-2910 BCE (Beta 392489) Russia F W6 163845
I0370 Yamnaya Yamnaya EBA Ishkinovka I, Eastern Orenburg, Pre-Ural steppe, Samara, Russia: SVP10 3300-2700 BCE Russia M H13a1a1a R1b1a2a2 199345
I0429 Yamnaya Yamnaya EBA Lopatino I, Sok River, Samara, Russia, SVP38 3339-2917 cal BCE (AA47804) Russia M T2c1a2 R1b1a2a2 217664
I0438 Yamnaya Yamnaya EBA Luzhki I, Samara River, Samara, Russia; SVP50 3021-2635 cal BCE (AA47807) Russia M U5a1a1 R1b1a2a2 213493
I0439 Yamnaya Yamnaya EBA Lopatino I, Sok River, Samara, Russia, SVP52 3305-2925 cal BCE (Beta 392491) Russia M U5a1a1 R1b1a 98900
I0441 Yamnaya Yamnaya EBA Kurmanaevskii III, Buzuluk, Samara, Russia; SVP54 3010-2622 cal BCE (AA47805) Russia F H2b 51326
I0443 Yamnaya Yamnaya EBA Lopatino II, Sok River, Samara, Russia; SVP57 3300-2700 BCE Russia M W3a1a R1b1a2a 343890
I0444 Yamnaya Yamnaya EBA Kutuluk I, Kutuluk River, Samara, Russia; SVP58 3300-2700 BCE Russia M H6a1b R1b1a2a2 187126
I0550 Karsdorf_LN unknown LN Karsdorf, Germany; KAR22a, feature 191 3950-3400 BCE?, C14C pending Germany F T1a1 59907
I0103 Corded_Ware_LN Corded Ware LN Esperstedt, Germany; ESP16, feature 6236 2566-2477 cal BCE (MAMS 21488) Germany F W6a 336918
I0049 Corded_Ware_LN Corded Ware LN Esperstedt, Germany; ESP22, feature 6140 2454-2291 cal BCE (MAMS 21489) Germany F X2b4 167170
I0106 Corded_Ware_LN Corded Ware LN Esperstedt, Germany; ESP26, feature 6233.1 2454-2291 cal BCE (MAMS 21490) Germany F T2a1b1 69886
I0104 Corded_Ware_LN Corded Ware LN Esperstedt, Germany; ESP11, feature 6216 2473-2348 cal BCE (MAMS 21487) Germany M U4b1a1a1 R1a1a1 336637
I0059 BenzigerodeHeimburg_LN Bell Beaker? LN Benzingerode-Heimburg, Germany; BZH6, grave 2, feature/find 1287/1036 2286-2153 cal BCE (MAMS 21486) Germany F H1 /H1b'ad 241081
I0058 BenzigerodeHeimburg_LN Bell Beaker LN Benzingerode-Heimburg, Germany; BZH4, grave 7, feature 4607 2283-2146 cal BCE (MAMS 21491) Germany F H1e 246728
I0171 BenzigerodeHeimburg_LN Bell Beaker? LN Benzingerode-Heimburg, Germany; BZH12, grave 3 ,feature 6256 2204-2136 cal BCE (KIA27952 ) Germany F U5a1a2a 66800
I0112 Bell_Beaker_LN Bell Beaker LN Quedlinburg XII, Germany; QUEXII6, feature 6256 2340-2190 cal BCE (Er7038) Germany F H13a1a2 341003
I0113 Bell_Beaker_LN Bell Beaker LN Quedlinburg XII, Germany; QUEXII4, feature 6255.1 2290-2130 cal BCE (Er7283) Germany F J1c5 190352
I0108 Bell_Beaker_LN Bell Beaker LN Rothenschirmbach, Germany; ROT6, feature 10044 2497-2436 cal BCE (Er8710) Germany F H5a3 260528
I0111 Bell_Beaker_LN Bell Beaker LN Rothenschirmbach, Germany; ROT4, feature 10142 2414-2333 cal BCE (Er8712) Germany F H3new 208256
I0060 Bell_Beaker_LN Bell Beaker LN Rothenschirmbach, Germany; ROT3, feature 10011 2294-2206 cal BCE (MAMS 22819) Germany F K1a2c 47085
I0806 Bell_Beaker_LN Bell Beaker LN Quedlinburg VII 2, Germany; QLB28b, feature 19617 2296-2206 cal BCE (MAMS 22820) Germany M H1 R1b1a2a1a2 91757
I0118 Alberstedt_LN unknown LN Alberstedt, Germany; ALB3, feature 7144.2 2459-2345 cal BCE (MAMS 21492) Germany F HV6'17 349956
I0114 Unetice_EBA_relative_of_I0117 Unetice EBA Esperstedt, Germany; ESP2, feature 3340.1 2131-1979 cal BCE (MAMS 21493) Germany M I3a I2a2 217031
I0115 Unetice_EBA Unetice EBA Esperstedt, Germany; ESP3, feature 1559.1 1931-1780 cal BCE (MAMS 21494) Germany F U5a1 123744
I0116 Unetice_EBA Unetice EBA Esperstedt, Germany; ESP4, feature 3322/3323 2118-1961 cal BCE (MAMS 21495) Germany M W3a1 I2c2 308158
I0117 Unetice_EBA Unetice EBA Esperstedt, Germany; ESP29, feature 3332/3333 2199-2064 cal BCE (MAMS 21496) Germany F I3a 279996
I0164 Unetice_EBA Unetice EBA Quedlinburg VIII, Germany; QUEVII6, feature 3580 2012-1919 cal BCE (MAMS 21497) Germany F pre-U5b2a1b 332832
I0803 Unetice_EBA Unetice EBA Eulau, Germany; EUL41A, feature 882 2115-1966 cal BCE (MAMS 22822) Germany F H4a1a1 144186
I0804 Unetice_EBA Unetice EBA Eulau, Germany; EUL57B, feature1911.5 2131-1982 cal BCE (MAMS 22821) Germany M H3 I2 22869
I0047 Unetice_EBA Unetice EBA Halberstadt-Sonntagsfeld, Germany; HAL16, grave 19, feature 613.1 2022-1937 cal BCE (MAMS 21481) Germany F V 288353
I0099 Halberstadt_LBA Late Bronze Age LBA Halberstadt-Sonntagsfeld, Germany; HAL36C, grave 40, feature 1114 1113-1021 cal BCE (MAMS 21484) Germany M H23 R1a1a1b1a2 337566

25
Extended Data Table 3: Pairwise FST for all ancient groups with ≥2 individuals, present-day Europeans with ≥10 individuals, and
selected other groups. FST (below the diagonal), standard deviation (above the diagonal).

BenzigerodeHeimburg_LN

SwedenSkoglund_NHG
HungaryGamba_EN
HungaryGamba_BA
Corded_Ware_LN
Bell_Beaker_LN
Baalberge_MN

Unetice_EBA
Motala_HG

Norwegian
Belarusian

Spain_MN
Mordovian
Lithuanian
Hungarian

Spain_EN
BedouinB
Armenian

Iraqi_Jew

Yamnaya
Bulgarian

Sardinian
Bergamo

Orcadian
Icelandic

Karitiana
Estonian

LBK_EN
Croatian

Russian

Spanish
Papuan
Basque

English

Turkish

Yoruba
Sicilian
French

Lezgin
Czech

Sindhi
Greek

WHG
Onge
Mala
EHG

Han
Armenian 0.004 0.000 0.001 0.001 0.001 0.002 0.000 0.000 0.001 0.000 0.000 0.002 0.000 0.001 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.001 0.001 0.001 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Baalberge_MN 0.023 0.003 0.004 0.004 0.005 0.006 0.004 0.004 0.005 0.004 0.003 0.007 0.004 0.004 0.003 0.003 0.004 0.003 0.008 0.005 0.004 0.004 0.005 0.004 0.004 0.004 0.004 0.004 0.004 0.003 0.004 0.003 0.004 0.003 0.003 0.003 0.003 0.005 0.005 0.003 0.013 0.003 0.004 0.005 0.004 0.004
Basque 0.017 0.019 0.001 0.000 0.001 0.002 0.000 0.000 0.001 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
BedouinB 0.025 0.047 0.041 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.000 0.000 0.001 0.001 0.002 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Belarusian 0.014 0.026 0.012 0.041 0.001 0.002 0.000 0.000 0.001 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Bell_Beaker_LN 0.017 0.020 0.012 0.042 0.008 0.002 0.001 0.001 0.002 0.001 0.001 0.003 0.001 0.001 0.001 0.001 0.002 0.001 0.003 0.002 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.002 0.001 0.002 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.004 0.001 0.001 0.002 0.001 0.001
BenzigerodeHeimburg_LN 0.019 0.021 0.017 0.047 0.011 0.004 0.002 0.002 0.003 0.002 0.002 0.004 0.002 0.002 0.002 0.002 0.002 0.002 0.004 0.003 0.002 0.002 0.003 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.003 0.002 0.003 0.002 0.002 0.002 0.002 0.003 0.003 0.002 0.006 0.002 0.002 0.003 0.002 0.002
Bergamo 0.007 0.015 0.008 0.032 0.007 0.008 0.012 0.000 0.001 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Bulgarian 0.006 0.019 0.009 0.031 0.003 0.008 0.010 0.002 0.001 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Corded_Ware_LN 0.023 0.032 0.025 0.052 0.015 0.008 0.007 0.019 0.017 0.001 0.001 0.003 0.001 0.001 0.001 0.001 0.002 0.001 0.003 0.002 0.001 0.002 0.002 0.002 0.001 0.001 0.002 0.001 0.002 0.001 0.002 0.001 0.002 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.004 0.001 0.002 0.002 0.002 0.002
Croatian 0.009 0.019 0.010 0.035 0.002 0.007 0.011 0.004 0.001 0.015 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Czech 0.011 0.024 0.009 0.037 0.001 0.007 0.010 0.005 0.002 0.015 0.001 0.002 0.000 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
EHG 0.067 0.082 0.060 0.093 0.042 0.041 0.038 0.057 0.051 0.034 0.048 0.043 0.002 0.002 0.002 0.002 0.003 0.002 0.004 0.003 0.002 0.003 0.003 0.003 0.002 0.002 0.002 0.002 0.003 0.002 0.003 0.002 0.003 0.002 0.002 0.002 0.002 0.003 0.003 0.002 0.006 0.002 0.003 0.003 0.002 0.002
English 0.011 0.020 0.008 0.038 0.004 0.006 0.009 0.004 0.003 0.014 0.003 0.002 0.045 0.000 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Estonian 0.017 0.027 0.014 0.045 0.002 0.009 0.011 0.010 0.006 0.016 0.005 0.002 0.040 0.005 0.000 0.000 0.001 0.000 0.002 0.001 0.000 0.001 0.002 0.001 0.001 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
French 0.009 0.018 0.006 0.035 0.004 0.006 0.008 0.002 0.002 0.015 0.002 0.001 0.048 0.001 0.005 0.000 0.001 0.000 0.002 0.001 0.000 0.000 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Greek 0.004 0.018 0.010 0.028 0.007 0.010 0.014 0.002 0.000 0.019 0.003 0.004 0.057 0.005 0.009 0.003 0.001 0.000 0.002 0.001 0.000 0.000 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Han 0.108 0.136 0.120 0.128 0.110 0.114 0.117 0.113 0.108 0.118 0.110 0.111 0.135 0.113 0.110 0.110 0.110 0.001 0.003 0.002 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.003 0.001 0.001 0.002 0.001 0.001
Hungarian 0.009 0.019 0.009 0.036 0.001 0.007 0.009 0.003 0.001 0.014 0.002 0.000 0.046 0.001 0.003 0.001 0.003 0.108 0.002 0.001 0.000 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
HungaryGamba_BA 0.016 0.021 0.010 0.039 0.005 0.008 0.010 0.008 0.005 0.016 0.005 0.005 0.046 0.007 0.006 0.005 0.007 0.115 0.005 0.003 0.002 0.002 0.003 0.002 0.002 0.002 0.002 0.002 0.003 0.002 0.003 0.002 0.003 0.002 0.002 0.002 0.002 0.003 0.003 0.002 0.006 0.002 0.003 0.003 0.003 0.002
HungaryGamba_EN 0.016 0.009 0.018 0.037 0.021 0.020 0.025 0.009 0.012 0.037 0.015 0.018 0.081 0.018 0.026 0.014 0.010 0.129 0.016 0.014 0.001 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.002 0.001 0.002 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.004 0.001 0.002 0.002 0.002 0.001
Icelandic 0.015 0.025 0.011 0.041 0.005 0.008 0.011 0.006 0.005 0.014 0.006 0.003 0.045 0.002 0.006 0.003 0.008 0.113 0.004 0.007 0.021 0.001 0.002 0.001 0.000 0.000 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Iraqi_Jew 0.005 0.031 0.021 0.026 0.020 0.022 0.024 0.011 0.010 0.029 0.014 0.017 0.073 0.016 0.023 0.014 0.009 0.112 0.015 0.018 0.020 0.021 0.002 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Karitiana 0.210 0.234 0.218 0.230 0.205 0.208 0.208 0.212 0.206 0.209 0.207 0.206 0.206 0.209 0.205 0.208 0.209 0.180 0.205 0.214 0.229 0.209 0.214 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.004 0.002 0.002 0.003 0.002 0.001
LBK_EN 0.023 0.009 0.024 0.043 0.029 0.021 0.025 0.017 0.018 0.035 0.022 0.024 0.084 0.024 0.033 0.020 0.016 0.137 0.022 0.018 0.006 0.029 0.026 0.237 0.001 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.003 0.001 0.001 0.002 0.001 0.001
Lezgin 0.005 0.033 0.018 0.034 0.012 0.015 0.015 0.010 0.007 0.017 0.009 0.010 0.052 0.011 0.014 0.009 0.007 0.103 0.009 0.017 0.025 0.013 0.011 0.201 0.032 0.001 0.001 0.000 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Lithuanian 0.019 0.030 0.014 0.046 0.001 0.010 0.010 0.011 0.006 0.016 0.005 0.003 0.043 0.006 0.002 0.006 0.010 0.114 0.004 0.007 0.026 0.007 0.024 0.209 0.034 0.015 0.001 0.000 0.001 0.000 0.002 0.000 0.002 0.000 0.000 0.000 0.001 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Mala 0.045 0.074 0.061 0.068 0.051 0.055 0.057 0.053 0.048 0.058 0.051 0.051 0.081 0.053 0.053 0.051 0.049 0.070 0.050 0.056 0.070 0.055 0.049 0.187 0.077 0.043 0.056 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.000 0.002 0.001 0.001 0.003 0.001 0.001 0.002 0.001 0.001
Mordovian 0.014 0.028 0.014 0.041 0.002 0.010 0.011 0.009 0.005 0.013 0.005 0.003 0.040 0.005 0.003 0.005 0.008 0.100 0.003 0.007 0.025 0.006 0.020 0.196 0.032 0.011 0.004 0.047 0.001 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Motala_HG 0.088 0.083 0.072 0.113 0.060 0.057 0.053 0.073 0.069 0.060 0.067 0.062 0.048 0.063 0.058 0.064 0.075 0.166 0.062 0.058 0.091 0.062 0.095 0.254 0.093 0.080 0.058 0.112 0.060 0.001 0.002 0.001 0.002 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.004 0.001 0.001 0.002 0.001 0.001
Norwegian 0.014 0.024 0.010 0.041 0.004 0.007 0.009 0.006 0.005 0.014 0.003 0.002 0.044 0.002 0.005 0.002 0.007 0.112 0.003 0.009 0.021 0.002 0.020 0.207 0.028 0.012 0.006 0.053 0.006 0.062 0.002 0.000 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Onge 0.168 0.191 0.181 0.186 0.171 0.174 0.177 0.174 0.170 0.180 0.171 0.172 0.197 0.174 0.173 0.172 0.170 0.155 0.171 0.176 0.187 0.175 0.172 0.283 0.194 0.167 0.177 0.129 0.166 0.227 0.174 0.001 0.002 0.001 0.001 0.001 0.001 0.002 0.002 0.001 0.004 0.001 0.002 0.002 0.002 0.001
Orcadian 0.014 0.021 0.011 0.041 0.006 0.009 0.010 0.007 0.006 0.016 0.006 0.004 0.047 0.003 0.007 0.004 0.008 0.114 0.004 0.008 0.021 0.005 0.020 0.209 0.029 0.013 0.008 0.054 0.007 0.063 0.004 0.175 0.002 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Papuan 0.186 0.206 0.199 0.202 0.190 0.194 0.194 0.193 0.187 0.199 0.190 0.191 0.218 0.192 0.192 0.191 0.189 0.180 0.189 0.192 0.204 0.195 0.190 0.304 0.212 0.185 0.196 0.151 0.185 0.244 0.193 0.232 0.193 0.002 0.002 0.002 0.002 0.002 0.002 0.002 0.004 0.002 0.002 0.002 0.002 0.002
Russian 0.015 0.027 0.014 0.042 0.002 0.010 0.010 0.009 0.005 0.015 0.005 0.003 0.039 0.005 0.002 0.005 0.008 0.099 0.003 0.007 0.024 0.006 0.021 0.196 0.032 0.012 0.004 0.047 0.002 0.058 0.005 0.166 0.007 0.185 0.000 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Sardinian 0.014 0.014 0.013 0.036 0.017 0.016 0.022 0.007 0.009 0.033 0.011 0.013 0.074 0.012 0.020 0.009 0.009 0.123 0.012 0.014 0.010 0.016 0.017 0.225 0.015 0.021 0.021 0.064 0.020 0.085 0.016 0.183 0.016 0.201 0.020 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Sicilian 0.003 0.017 0.011 0.027 0.009 0.011 0.013 0.003 0.002 0.022 0.004 0.006 0.059 0.006 0.012 0.004 0.002 0.109 0.005 0.010 0.010 0.010 0.008 0.209 0.016 0.009 0.013 0.048 0.010 0.078 0.007 0.169 0.009 0.187 0.010 0.008 0.000 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Sindhi 0.016 0.047 0.032 0.043 0.023 0.026 0.028 0.023 0.020 0.029 0.022 0.022 0.057 0.023 0.025 0.022 0.020 0.081 0.021 0.027 0.042 0.025 0.020 0.185 0.049 0.014 0.028 0.014 0.021 0.088 0.024 0.142 0.026 0.162 0.021 0.035 0.021 0.001 0.001 0.000 0.003 0.000 0.001 0.002 0.001 0.001
Spain_EN 0.034 0.021 0.030 0.053 0.037 0.028 0.035 0.024 0.028 0.044 0.030 0.032 0.092 0.032 0.040 0.028 0.027 0.144 0.031 0.028 0.018 0.036 0.037 0.245 0.015 0.042 0.041 0.085 0.039 0.099 0.036 0.202 0.036 0.219 0.041 0.021 0.026 0.058 0.002 0.001 0.004 0.001 0.002 0.002 0.002 0.002
Spain_MN 0.033 0.014 0.025 0.054 0.032 0.022 0.026 0.023 0.026 0.035 0.027 0.029 0.077 0.027 0.036 0.024 0.025 0.138 0.028 0.028 0.023 0.031 0.037 0.241 0.016 0.040 0.035 0.079 0.035 0.081 0.030 0.196 0.031 0.214 0.035 0.021 0.025 0.052 0.009 0.001 0.004 0.001 0.002 0.002 0.002 0.002
Spanish 0.008 0.014 0.005 0.032 0.006 0.007 0.011 0.002 0.002 0.018 0.004 0.003 0.054 0.003 0.008 0.001 0.003 0.111 0.003 0.006 0.011 0.005 0.013 0.210 0.017 0.011 0.009 0.052 0.008 0.069 0.005 0.172 0.005 0.190 0.008 0.007 0.003 0.024 0.024 0.020 0.003 0.000 0.001 0.002 0.001 0.001
SwedenSkoglund_NHG 0.071 0.066 0.058 0.094 0.047 0.050 0.048 0.056 0.053 0.054 0.052 0.047 0.058 0.050 0.050 0.049 0.058 0.153 0.050 0.041 0.068 0.051 0.076 0.244 0.078 0.066 0.049 0.101 0.047 0.034 0.049 0.211 0.053 0.230 0.047 0.067 0.061 0.074 0.084 0.074 0.052 0.003 0.004 0.004 0.003 0.003
Turkish 0.001 0.023 0.014 0.026 0.010 0.013 0.016 0.005 0.004 0.021 0.006 0.008 0.058 0.009 0.013 0.006 0.003 0.095 0.007 0.013 0.016 0.012 0.006 0.198 0.022 0.005 0.015 0.040 0.010 0.081 0.011 0.160 0.012 0.178 0.011 0.013 0.003 0.013 0.032 0.030 0.006 0.064 0.001 0.002 0.001 0.001
Unetice_EBA 0.016 0.016 0.014 0.042 0.007 0.002 0.003 0.008 0.007 0.005 0.007 0.005 0.034 0.005 0.006 0.005 0.009 0.112 0.006 0.007 0.022 0.007 0.021 0.206 0.022 0.013 0.007 0.053 0.007 0.051 0.005 0.173 0.007 0.190 0.008 0.019 0.011 0.024 0.030 0.023 0.007 0.043 0.013 0.002 0.001 0.001
WHG 0.086 0.073 0.062 0.106 0.054 0.055 0.056 0.066 0.065 0.070 0.061 0.056 0.078 0.058 0.053 0.058 0.070 0.162 0.058 0.049 0.080 0.057 0.091 0.260 0.091 0.082 0.053 0.110 0.056 0.053 0.057 0.220 0.059 0.239 0.054 0.074 0.072 0.087 0.092 0.070 0.060 0.050 0.078 0.057 0.002 0.002
Yamnaya 0.030 0.054 0.034 0.060 0.020 0.016 0.013 0.028 0.023 0.011 0.022 0.020 0.028 0.021 0.020 0.022 0.026 0.118 0.020 0.023 0.050 0.020 0.037 0.204 0.052 0.019 0.022 0.056 0.018 0.063 0.021 0.178 0.022 0.197 0.019 0.044 0.030 0.028 0.060 0.053 0.027 0.062 0.026 0.012 0.076 0.001
Yoruba 0.142 0.163 0.155 0.141 0.150 0.151 0.153 0.148 0.145 0.158 0.148 0.149 0.179 0.150 0.152 0.147 0.145 0.177 0.148 0.151 0.156 0.152 0.141 0.270 0.163 0.143 0.155 0.139 0.148 0.200 0.151 0.222 0.152 0.231 0.148 0.153 0.139 0.133 0.169 0.167 0.144 0.185 0.137 0.151 0.195 0.158

26
Extended Data Figure Legends

Extended Data Figure 1: ADMIXTURE analysis.

Extended Data Figure 2: Outgroup f3-statistic f3(Dinka; X, Y), measuring the degree of
shared drift among pairs of ancient individuals.

Extended Data Figure 3: Modeling Corded Ware as a mixture of N=1, 2, or 3 ancestral

populations.
(a) We present in the left column a histogram of raw f4-statistic residuals and on the right Z-
scores for the best-fitting (lowest squared 2-norm of the residuals, or resnorm) model at each
N. (b), we show on the left how resnorm and on the right how the maximum |Z| score change
for different N. (c) resnorm of different N=2 models. The set of outgroups used in this
analysis, in the terminology of SI9, is “World Foci 15 + Ancients”.

Extended Data Fig. 4: Modeling Europeans as mixtures of increasing complexity: N=1

(EN), N=2 (EN, WHG), N=3 (EN, WHG, Yamnaya), N=4 (EN, WHG, Yamnaya,
Nganasan), N=5 (EN, WHG, Yamnaya, Nganasan, BedouinB). The residual norm of the
fitted model (SI9) and its changes are indicated.

Extended Data Fig. 5: Geographic distribution of archaeological cultures and graphic

illustration of proposed population movements / turnovers discussed in the main text
(symbols of samples are identical to Figure 1):
(a) proposed routes of migration by early farmers into Europe ~9,000-7000 years ago, (b)
resurgence of hunter-gatherer ancestry during the Middle Neolithic 7,000-5,000 years ago, (c)
arrival of steppe ancestry in central Europe during the Late Neolithic ~4,500 years ago. White
arrows indicate the two possible scenarios of the arrival of Indo-European language groups.

27
Extended Data Figure 1 (high-resolution version uploaded separately).

28
Extended Data Figure 2

29
Extended Data Figure 3

30
Extended Data Figure 4

31
Extended Data Figure 5

32
Supplementary Information
Massive migration from the steppe is a source
for Indo-European languages in Europe
Table of Contents 1

SI 1 – Protocols for DNA extraction and library preparation 2-4

SI 2 – Mitochondrial genome analysis 5-21

SI 3 – Description of ancient samples and archaeological context 22-37

SI 4 – Sex determination and Y chromosome analysis 38-51

SI 5 – Principal Component Analysis 52-54

SI 6 – ADMIXTURE analysis 55-59

SI 7 – Population relationships and admixture using f-statistics 60-77

SI 8 – Phylogenetic relationships of ancient Eurasians 78-92

SI 9 – Inference of admixture proportions without detailed phylogenetic modeling 93-126

SI 10 – Number of streams of migration into ancient Europe 127-133

SI 11 – Relevance of ancient DNA to the problem of Indo-European language dispersals 134-140

1
Supplementary Information 1
Protocols for DNA extraction and library preparation

Nadin Rohland*, Wolfgang Haak*, Alissa Mittnik, Bastien Llamas, Eadaoin Harney, Susanne
Nordenfelt, Kristin Stewardson, David Reich

* To whom correspondence should be addressed ([email protected] or

[email protected])

Strategy
The 69 samples newly reported in this study reflect data from 95 ancient DNA libraries. They
are a subset of 119 samples and 212 libraries that were screened in the course of this study
(Online Table 1).
Briefly, our screening strategy included the following steps:
x Silica-based DNA extraction1,2
x Double-stranded library preparation with truncated barcoded Illumina adapters, with
either no DNA damage repair3, full USER-treatment4 or partial UDG repair5
x Shallow shotgun sequencing
x Mitochondrial DNA (mtDNA) capture and sequencing

We then selected samples for 390k SNP capture based on the following criteria:
x DNA damage patterns characteristic of ancient DNA
x Low estimates of contamination based on the mtDNA data6
x Evidence of substantial complexity in the library

Libraries with no UDG treatment were only used for screening (shallow shotgun sequencing
and mtDNA capture) to assess the authenticity of the DNA extracted from these samples. We
then prepared partial or full UDG treated libraries for the 390k SNP capture. All such libraries
also were subjected to their own round of screening (mtDNA capture and shotgun) prior to
390k capture.

In what follows we present details on the methods used for DNA extraction and library
preparation at the University of Adelaide’s Australian Centre for Ancient DNA (ACAD)
(n=151 libraries), the University of Tübingen Germany (n=16 libraries), and the Harvard
Medical School Boston USA (n=45 libraries).

Adelaide lab (n=151 libraries)

DNA extraction was carried out in the clean-room facilities at ACAD, using an in-solution
silica-based protocol2. Briefly, 200 mg of tooth or bone powder were demineralized in 4 mL
0.5M EDTA and 60 μL of Proteinase K at 37°C overnight, followed by an additional
incubation with 60 μL of Proteinase K at 55°C for 1 h. DNA was subsequently bound to silica
particles using a custom binding buffer [90% Qiagen QG buffer (i.e. 5 M guanidinium
thiocyanate), 220 mM sodium acetate, 25 mM sodium chloride, 1% Triton X 100] at a 1:4
ratio of lysate: binding buffer (vol:vol). DNA binding for 1h was followed by two washes in
80% ethanol and a final elution in 200 μL TE buffer (10mM Tris, 0.1mM EDTA, pH8.0)
containing 0.05% Tween 20.

2
DNA libraries were prepared in the clean-room facility following the preparation of the
extract. For a subset of samples, double-stranded libraries were prepared with truncated
barcoded adapters using the library preparation protocol by Meyer and Kircher (2010)3. This
protocol started with 20 μL of extract and did not involve UDG treatment to remove
characteristic ancient DNA damage. For libraries that had characteristics of authentic ancient
DNA as determined by shallow shotgun sequencing and mtDNA capture, we prepared one or
more double-stranded libraries with truncated barcoded adapters and a UDG and
endonuclease VIII damage repair treatment (USER enzyme, NEB)4.

DNA libraries were amplified by PCR in quintuplicates for an initial 12-13 cycles (AmpliTaq
Gold, Life Technologies), and PCR replicates were pooled and purified with the Agencourt
AMPure XP system. DNA libraries were then re-amplified for 12-13 cycles in quintuplicates
or sextuplicates, followed by pooling and purification, visual inspection on a 3.5% agarose
gel, and final quantification using a NanoDrop 2000c spectrophotometer.

This protocol evolved over the course of the study. For the last set of ancient DNA libraries
built for this study, partial DNA damage removal5 was performed using truncated barcoded
adapters. Briefly, this method removes DNA damage except for the first and last positions of
the reads, making it possible to test for the presence of characteristic DNA damage patterns
while making the DNA libraries more robust for population genetic analysis. These partially
UDG-treated libraries were amplified using isothermal amplifications using the commercial
TwistAmp® Basic kit (TwistDx Ltd). The amplification followed the manufacturer’s
recommendations and used 13.4 μL of libraries after the Bst amplification step, as well as an
increased incubation time of the isothermal reaction to 44 min at 37°C, followed by gel
electrophoresis and Nanodrop quantification. Approximately 1-2 μg DNA was shipped to
Harvard Medical School in Boston for further analysis.

Tübingen lab (n=16 libraries)

The preparation of Motala samples was described in Lazaridis et al. (2014)7.

For the 9 Stuttgart Linienbandkeramik (LBK) samples newly reported in this study, 100 μL
DNA extracts were prepared from 39-69 mg of dentin powder in the clean room facilities of
the University of Tübingen using the silica-based protocol by Dabney et al. (2013)1. A 40 μL
aliquot of each of these extracts was turned into a double-stranded library using the protocol
of Meyer and Kircher (2010)3, and indices were then added to both ends of the library
molecules8. An aliquot of each LBK extract was also sent to Boston for preparation of partial
UDG-treated libraries.

Boston lab (n=45 libraries)

DNA was extracted from approximately 75mg of bone powder following the extraction
method by Dabney et al. 20131 with the exception that the MinElute column Zymo extender
assembly was replaced by the High Pure Extender Assembly from the Roche High Pure Viral
Nucleic Acid Large Volume Kit (M. Meyer, personal communication). The elution volume
varied between 32 and 90 uL.

3
Library preparation was performed starting from between 20 and 30 uL of extract. We used a
double stranded library preparation that is an adaption of Meyer and Kircher (2010)3 and
Kircher, Sawyer and Meyer (2012)8, resulting in barcoded libraries with truncated adapters
following (Rohland et al. 2014)5. This library preparation, involving different types of UDG-
treatment, was use for all extracts prepared in Boston (from the Samara Valley Project
samples) as well as on nine DNA extracts prepared in Tübingen (from Stuttgart LBK
samples). Online Table 1 shows which library preparation method was used (i.e. no [minus]
or partial [half] or full [plus] UDG treatment); UDG minus was only used to assess the
authenticity of extracted DNA from the samples; nuclear data only originates from plus or
half libraries

A subset of libraries from Adelaide were amplified further in Boston to achieve sufficient
material for enrichment (mtDNA and 390k). We used Herculase II Fusion DNA Polymerase
(Agilent) and the primers PreHyb-F and PreHyb-R5 for these re-amplifications, and cleaned
the products with the MinElute PCR Purification System.

Summary information on all libraries made for this study is given in Online Table 1.

References

1. Dabney, J. et al. Complete mitochondrial genome sequence of a Middle Pleistocene

cave bear reconstructed from ultrashort DNA fragments. Proc Natl Acad Sci U S A
110, 15758-15763, (2013).
2. Brotherton, P. et al. Neolithic mitochondrial haplogroup H genomes and the genetic
origins of Europeans. Nat Commun 4, 1764, (2013).
3. Meyer, M. & Kircher, M. Illumina sequencing library preparation for highly
multiplexed target capture and sequencing. Cold Spring Harb Protoc 2010, pdb
prot5448, (2010).
4. Briggs, A. W. & Heyn, P. Preparation of next-generation sequencing libraries from
damaged DNA. Methods Mol Biol 840, 143-154, (2012).
5. Rohland, N., Harney, E., Mallick, S., Nordenfelt, S. & Reich, D. Partial UDG
treatment for screening of ancient DNA. Phil. Trans. R. Soc. B 370, 20130624
(2014).
6. Fu, Q. et al. A revised timescale for human evolution based on ancient mitochondrial
genomes. Current Biology 23, 553-559, (2013).
7. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for
present-day Europeans. Nature 513, 409-413, (2014).
8. Kircher, M., Sawyer, S. & Meyer, M. Double indexing overcomes inaccuracies in
multiplex sequencing on the Illumina platform. Nucleic Acids Research 40, e3,
(2012).

4
Supplementary Information 2
Mitochondrial genome analysis

Wolfgang Haak*, Swapan Mallick, Guido Brandt and Bastien Llamas

* To whom correspondence should be addressed ([email protected]),

We used an automated procedure to obtain an initial mitochondrial DNA haplogroup

determination. Following this, we double-checked reads for each sample visually. The final
set of haplogroup determinations is given in Online Table 1.

We imported bam files (deduplicated and quality filtered) into Geneious (Geneious version
7.1.3 created by Biomatters. Available from http://www.geneious.com/) and re-assembled the
reads against the Revised Sapiens Reference Sequence (RSRS)1. SNPs were called in
Geneious for all polymorphisms with a minimum coverage of 2 and a minimum variant
frequency of 0.7. The assembly and the resulting list of SNPs were double-checked against
SNPs reported in phylotree.org (mtDNA tree Build 16 (19 Feb 2014))2. Following
recommendations in van Oven et al. 2010, we excluded common indels at nucleotide
positions 309.1C(C), 315.1C, AC indels at 515-522, 16182C, 16183C, 16193.1C(C), but
report C16519T. Haplotype calls are given in Table S2.1, reporting all SNPs compared to
RSRS and following the hierarchical structure until the most derived SNP shared with an
existing sequence. Additional SNPs were considered to be private mutations and are reported
in a separate column.

We report 105 complete mitochondrial genomes from the 119 individuals that were analysed
for this study.

Overall, we found no inconsistencies between consensus calls from different library

architectures (UDG treated and untreated), between libraries from different extracts from the
same individual, or from mitochondrial sequences that had been generated previously using a
different technique3,4.

Sample HAL14 appears to harbor two potentially heteroplasmic sites at np 3338, where we
find that 60% of reads show a T and 40% a C in a UDG-treated library and 70/30% T/C ratio
in an untreated library, and at 16093 (51/49% C/T ratio UDG-treated and 17/83% C/T ratio
untreated). Since we observe both alleles of both sites in the UDG-treated library, DNA
damage is unlikely to be an explanation. The fact that all other SNP calls are unambiguous,
i.e. the variant is observed in >99.9% of all reads, also suggests that contamination is an
unlikely explanation. Of note, T3338C has been reported as a sub-haplogroup defining SNP
in T2b8, immediately derived from T2b, while 16093 has been described as recurrent
mutation5. Another heteroplasmic site was found in sample La Mina 2 at np 16274, where
54% of the reads show and A and 46% a G in a UDG-treated libraries, whereas all
neighbouring D-loop SNP are consistent throughout all reads, again rendering contamination
unlikely.

The majority of samples presented here were previously typed by sequencing of the
hypervariable segment I (HVS I) and/or by typing of 22 haplogroup-defining coding region

5
SNPs6 as part of large-scale diachronic studies7-9. The fact that previous PCR-based
mitochondrial analysis of these samples was successful likely explains the high success rate
of our mtDNA capture (88%). Haplogroup calls from PCR-based typing techniques were
consistent in all cases with our results from complete sequencing of the mitochondrial
genomes, expect for individual BENZ18 due to a sample mixup with another sample from the
same site. Haplogroup calls were also consistent with previously reported results using a
Mitochip v2 typing assay3. The full mitochondrial sequencing provided higher resolution
haplogroup determination, but did not alter our previous findings about the history of the
Mittelelbe-Saale sample series. Briefly10:
1. Hunter-gatherers carry predominantly U-haplotypes.
2. Early farmers predominantly carry haplotypes T2, K, HV, H, V, X and N1a.
3. Middle Neolithic farmers exhibit a similar mitochondrial profile as early farmers and
show an increase in the frequency of U-haplotypes, providing mitochondrial DNA
based confirmation of the resurgence of hunter-gatherer lineages that is also
documented from our autosomal data.
4. Late Neolithic individuals mark a second major turnover, highlighted by the
appearance of I, T1, and increase of U2, U4, U5a, H and W haplotypes.

We used the mitochondrial haplotype frequencies in Brandt et al. 201410, and in a recent study
of the Yamnaya and Catacomb cultures11 together with the new data (10 Yamnaya) reported
here to explore the maternal relationships between the Late Cooper Age/early Bronze Age
groups from the steppes and Late Neolithic groups from Central Europe (Table S2.2). We
added newly reported mitochondrial genomes from this study to existing population data sets,
notably eight novel sequences to the early Neolithic LBK set, and also removed two samples
that radiocarbon dating in this study showed were from more recent time periods (HAL36 and
HAL16). We added the Samara hunter-gatherer to the Eastern hunter-gather group (HGE) and
two new samples to the Middle Neolithic Baalberge group (BAC). The ten new samples from
Els Trocs (5) and La Mina (5) were pooled with the Epi-Cardial (CAR) and Middle Neolithic
(MNS) data from Spain, respectively. We did not include the three N1a genomes from
Hungary as these were chosen non-randomly for a phylogeographic study.

We performed principal component analysis (PCA) based on frequencies from the 22 most
common West Eurasia mitochondrial haplogroups (Table S2.2; Figure S2.1). The results
show that Bronze Age Kurgan samples from the steppe (BAK) and Siberia (BAS) as well as
the Catacomb culture (CAT) are linked to the hunter-gatherer samples from eastern Europe
(HGE) and northern Europe (PWC), and separated from central European early and Middle
Neolithic samples by the first PC (20%).

The Late Copper Age/Early Bronze Age Yamnaya fall closest to the central European Late
Neolithic Corded Ware (CWC) and Bell Beaker groups (BBC), and in particular to samples
from the early Bronze Age Unetice culture (UC). This supports a genetic contribution of
eastern groups to central Europe during the Late Neolithic ~4,500 BCE, consistent with the
autosomal SNP data. These results indicate that the Late Neolithic migration into central
Europe was not entirely male in origin.

6
 T2

T1
K
2 Z BAS U2
N1a W
Principal component 2 - 11.3 % variance

I U4 U5a
LBK HV U8 C HGE
BAC
SCG BAK
X UC
J YAM
CWC R
FBC
RSC
U3 PWC CAT
0

MNS
SMC V BBC
BEC
TRE
CAR
HGC
-2

NPO
NBQ
N* U

MNG

H
-4

HGS

U5b

-4 -2 0 2 4 6
Principal component 1 - 20 % variance

Figure S2.1. PCA based on mitochondrial DNA haplogroup frequencies from 24

prehistoric populations. Temporal groups are color-coded: hunter-gatherer (blue), Early
Neolithic (brown), Middle Neolithic (orange), Late Neolithic/Early Bronze Age (yellow), and
Late Neolithic/Early Bronze Age East (green). For full length names and details of samples
and populations please refer to Table S2.2.

7
Table S2.1. Mitochondrial sequence haplotypes of the 105 ancient samples in this study.
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
I0061 UzOO74 C1g (formerly A8577G, A11605t, C146T, C152T, C195T, A247G, A249d, 290-291d, T489C, A769G, A825t, A1018G, A2758G, C2885T, T3552a, T3594C, G4104A,
C1f) A12217G, T4312C, A4715G, G7146A, C7196a, T7256C, A7521G, T8468C, A8577G, G8584A, T8655C, A9545G, C10400T, T10664C, A10688G,
T16189C! C10810T, C10915T, A11605t, A12217G, G13105A, A13263G, G13276A, T13506C, T13650C, T14318C, T14783C, G15043A,
G15301A, A15487t, A16129G, T16187C, G16230A, T16278C, T16298C, C16311T, T16325C, C16327T, C16519T
I0124 SVP44 U5a1d A16241C, C16519T C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3027C, T3197C, T3594C, G4104A, T4312C, G7146A,
T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G, A15218G,
A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, A16241C, C16256T, C16270T, T16278C, C16311T, A16399G, C16519T
I0011 Motala1 U5a1 G5460A, T16093C, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, G5460A, G7146A,
C16519T T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G, A15218G,
T16093C, A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, A16399G, C16519T
I0017 Motala12 U2e1 C16527T C146T, C195T, T217C, A247G, C340T, A508G, A769G, A825t, A1018G, A1811G, A2758G, C2885T, T3594C, A3720G, G4104A,
T4312C, A5390G, T5426C, C6045T, T6152C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C,
A10688G, C10810T, C10873T, A10876G, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, T13020C, G13105A,
G13276A, T13506C, T13650C, T13734C, A15907G, A16051G, G16129c, T16187C, T16223C, G16230A, T16278C, C16311T, T16362C,
C16527T
I0012 Motala2 U2e1 C16527T C146T, C195T, T217C, A247G, C340T, A508G, A769G, A825t, A1018G, A1811G, A2758G, C2885T, T3594C, A3720G, G4104A,
T4312C, A5390G, T5426C, C6045T, T6152C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C,
A10688G, C10810T, C10873T, A10876G, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, T13020C, G13105A,
G13276A, T13506C, T13650C, T13734C, A15907G, A16051G, G16129c, T16187C, T16223C, G16230A, T16278C, C16311T, T16362C,
C16527T
I0013 Motala3 U5a1 G5460A, G8860A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, G5460A, G7146A,
A9389G, C16519T T7256C, A7521G, T8468C, T8655C, G8701A, G8860A, A9389G, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C,
A14793G, A15218G, A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T,
A16399G, C16519T
I0014 Motala4 U5a2d A13158G, C16519T C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, G7146A, T7256C,
A7521G, A7843G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, A13158G, G13276A, T13506C, T13617C, T13650C, A14793G,
A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, C16519T, G16526A
I0015 Motala6 U5a2d T152C!, G6480A, C146T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, G6480A, G7146A, T7256C,
C16519T A7521G, A7843G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G, A16129G,
T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, C16519T, G16526A
I0016 Motala9 U5a2 T195C!, G228A, C146T, C152T, G228A, A247G, C514T, A769G, A825t, A1018G, G1888A, A2246G, A2758G, C2885T, T3197C, T3594C, A3756T,
C514T, G1888A, G4104A, T4312C, G6917A, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G8854A, G9477A, A9531G, C9540T, G10398A,
A2246G, A3756T, T10664C, A10688G, C10810T, C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A,
G6917A, G8854A, T13506C, T13617C, T13650C, A14793G, A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T, T16278C,
A9531G, C16519T C16311T, C16519T, G16526A

8
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
I0101 DEB36 U5a1a'g C150T, G1007A, C146T, C150T, C152T, C195T, A247G, A769G, A825t, G1007A, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C,
T16093C, C16519T G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G,
A15218G, T16093C, A16129G, T16187C, C16189T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, A16399G, C16519T
I0056 HAL14 T2b(8?) heteroplasmy at np C146T, C152T, C195T, A247G, G709A, A769G, A825t, G930A, A1018G, G1888A, A2758G, C2885T, T3338C (40%), T3594C,
3338 (60% T, 40% G4104A, T4216C, T4312C, A4917G, G5147A, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A,
C) and 16093 (51% T10463C, T10664C, A10688G, C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A,
C, 49% T) G13368A, T13506C, T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16294T, C16296T, T16304C, C16311T
I0102 HAL15 N1a1a1a3 T4772C, G16129A! C146T, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T, T3336C,
T3594C, G4104A, T4312C, T4772C, A5315G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, A8901G, C9540T, T10238C,
T10664C, A10688G, C10810T, C10873T, C10915T, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C, A13780G,
G15043A, C16147A, T16154C, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, C16311T, C16320T, C16355T
I0821 HAL24 pre-X2d1 Topologically C146T, C152T, G207A, A247G, A769G, A825t, A1018G, G1719A, A2758G, C2885T, T3594C, G4104A, T4312C, T6221C, C6371T,
missing: T204C, A6791G, G7146A, T7256C, A7521G, T8468C, T8503C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T,
A5186G C10873T, C10915T, T11878C, A11914G, G13105A, G13276A, T13506C, T13650C, G13708A, A13966G, T14470C, A16129G,
T16187C, G16230A, C16311T
I0048 HAL25 K1a 195C!, T8433g, C146T, C152T, A247G, C497T, A769G, A825t, A1018G, T1189C, A1811G, A2758G, C2885T, A3480G, T3594C, G4104A, T4312C,
T16093C G7146A, T7256C, A7521G, T8433g, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, A10550G, T10664C, A10688G, C10810T,
C10873T, C10915T, T11299C, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13650C,
C14167T, T14798C, T16093C, A16129G, T16187C, C16189T, T16223C, T16224C, G16230A, T16278C
I0659 HAL2A N1a1a1a2 Topologically C146T, C195T, T199C, T204C, G207A, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T,
missing: T15299C, T3336C, T3594C, G4104A, T4312C, A5315G, G7146A, T7256C, A7521G, T8468C, G8485A, T8655C, G8701A, A8901G, C9540T,
C16172T! T10238C, T10664C, A10688G, C10810T, C10873T, C10915T, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C,
A13780G, G15043A, T16086C, A16129G, C16147A, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, C16311T,
C16320T, C16355T
I0057 HAL34 N1a1a1 C152T!!, A11884G, C146T, C152T!!, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T,
G14259A T3336C, T3594C, G4104A, T4312C, A5315G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, A8901G, C9540T, T10238C,
T10664C, A10688G, C10810T, C10873T, C10915T, A11884G, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C,
A13780G, G14259A, G15043A, A16129G, C16147A, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, C16311T,
C16355T
I0820 HAL37 W1c'i none T119C, C146T, C152T, A189G, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, A2758G, C2885T, A3505G,
T3594C, G4104A, T4312C, G5046A, G5460A, G7146A, T7256C, A7521G, C7864T, G8251A, T8468C, T8655C, G8701A, G8994A,
C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G, A11947G, T12414C, G13105A, G13276A,
T13506C, T13650C, G15884c, A16129G, T16187C, C16189T, G16230A, T16278C, C16292T, C16311T
I0100 HAL4 N1a1a1a T15900C C146T, C152T!!, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T,
T3336C, T3594C, G4104A, T4312C, A5315G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, A8901G, C9540T, T10238C,
T10664C, A10688G, C10810T, C10873T, C10915T, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C, A13780G,
G15043A, T15900C, A16129G, C16147A, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, C16311T, C16320T, C16355T
I0046 HAL5 T2c1d'e'f C9651a C152T, C195T, A247G, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C, T4312C, A4917G,
G6261A, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, C9651a, G10398A, T10463C, T10664C, A10688G,
C10810T, C10822T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C,

9
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16292T, C16294T, C16296T, C16311T
I0796 KAR11B H T152C! G73A, C146T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C, G7146A,
T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G,
A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16311T
I0797 KAR16A H46b C152T!! G73A, C146T, C152T!!, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2772T, C2885T, T3594C, G4104A, T4312C,
T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11893G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, C16311T
I0795 KAR6A H1 G1719A, C14380T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G1719A, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A,
T4312C, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, C14380T, T14766C, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, C16311T
I0019 LBK1254 HV6'17 C16519T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A, T4312C, G7146A, T7256C,
A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G,
T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16519T
I0020 LBK1577 T2e T195C! C146T, C150T, C152T, A247G, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C, T4312C,
A4917G, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C, A10688G,
C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C, T13650C,
A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, G16153A, T16187C, C16189T, T16223C, G16230A,
T16278C, C16294T, C16296T, C16311T
I0021 LBK1581 T2b none C146T, C152T, C195T, A247G, G709A, A769G, A825t, G930A, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, G5147A, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C,
T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16294T, T16304C, C16311T
I0022 LBK1976 T2e none C146T, C150T, C152T, C195T, A247G, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C, A10688G,
C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C, T13650C,
A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, G16153A, T16187C, C16189T, T16223C, G16230A,
T16278C, C16294T, C16296T, C16311T
I0023 LBK1979 H C16186T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, C16186T, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0024 LBK1988 W1c'i none T119C, C146T, C152T, A189G, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, A2758G, C2885T, A3505G,
T3594C, G4104A, T4312C, G5046A, G5460A, G7146A, T7256C, A7521G, C7864T, G8251A, T8468C, T8655C, G8701A, G8994A,
C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G, A11947G, T12414C, G13105A, G13276A,
T13506C, T13650C, G15884c, A16129G, T16187C, C16189T, G16230A, T16278C, C16292T, C16311T
I0025 LBK1992 T2b none C146T, C152T, C195T, A247G, G709A, A769G, A825t, G930A, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, G5147A, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C,

10
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
A10688G, C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C,
T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16294T, C16296T, T16304C, C16311T
I0026 LBK2155 T2b none C146T, C152T, C195T, A247G, G709A, A769G, A825t, G930A, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, G5147A, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C,
T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16294T, C16296T, T16304C, C16311T
I0027 LBK2172 H40 none G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C,
G7146A, T7256C, A7521G, T7621C, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0054 UWS4 J1c17 C9302T, A15467G C146T, C152T, G185A, C195T, G228A, A247G, C295T, C462T, T489C, A769G, A825t, A1018G, A2758G, C2885T, G3010A, T3594C,
T3847C, G4104A, T4216C, T4312C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9302T, C9540T, T10664C, A10688G,
C10810T, C10873T, C10915T, A11251G, A11914G, A12612G, T12705C, G13105A, G13276A, T13506C, T13650C, G13708A,
T14798C, C15452a, A15467G, C16069T, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T
I0176 SZEH4b N1a1a1a3 A16300G C146T, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T, T3336C,
T3594C, G4104A, T4312C, A5315G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, A8901G, C9540T, T10238C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C, A13780G, G15043A,
A16129G, C16147A, T16154C, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, A16300G, C16311T, C16320T, C16355T
I0409 Troc1 J1c3 C16519T C146T, C152T, G185A, C195T, G228A, A247G, C295T, C462T, T489C, A769G, A825t, A1018G, A2758G, C2885T, G3010A, T3594C,
G4104A, T4216C, T4312C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, T10664C, A10688G, C10810T, C10873T,
C10915T, A11251G, A11914G, A12612G, T12705C, G13105A, G13276A, T13506C, T13650C, G13708A, C13934T, T14798C,
C15452a, C16069T, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16519T
I0410 Troc3 pre-T2c1d2 none C195T, A247G, T279C, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C, T4312C, A4917G,
C5187T, G6261A, G7146A, T7256C, A7521G, C7873T, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C,
A10688G, C10810T, C10822T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A,
T13506C, T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T, T16223C,
G16230A, T16278C, C16292T, C16294T, C16311T
I0412 Troc5 N1a1a1 G5460A C146T, C152T!!, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T,
T3336C, T3594C, G4104A, T4312C, A5315G, G5460A, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, A8901G, C9540T,
T10238C, T10664C, A10688G, C10810T, C10873T, C10915T, A11884G, A11914G, G12501A, G13105A, G13276A, T13506C,
T13650C, A13780G, G15043A, A16129G, C16147A, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, C16311T, C16355T
I0413 Troc7 V C16519T T72C, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A, T4312C, G4580A,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, C15904T, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, T16298C, C16311T, C16519T
I0411 Troc4 K1a2a none C146T, C152T, C195T, A247G, C497T, A769G, A825t, A1018G, T1189C, A1811G, A2758G, C2885T, A3480G, T3594C, G4104A,
T4312C, G5773A, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, A10550G, T10664C, A10688G,
C10810T, C10873T, C10915T, T11025C, T11299C, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A,
T13506C, T13650C, C14167T, T14798C, A16129G, T16187C, C16189T, T16223T, T16224C, G16230A, T16278C
I0174 BAM25a N1a1a1 C16193T C146T, C152T!!, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T,

11
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
T3336C, T3594C, G4104A, T4312C, A5315G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, A8901G, C9540T, T10238C,
T10664C, A10688G, C10810T, C10873T, C10915T, A11884G, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C,
A13780G, G15043A, A16129G, C16147A, T16172C, C16189T, C16193T, G16230A, C16248T, T16278C, C16311T, C16355T
I0166 HAL13 V1a C7331T, T16311C!, T72C, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A, T4312C, G4580A,
C16519T T4639C, G7146A, T7256C, C7331T, A7521G, T8468C, T8655C, G8701A, A8869G, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, C15904T, A16129G,
T16187C, C16189T, T16223C, G16230A, T16278C, T16298C, C16519T
I0162 OSH1 H16a'c'd none G73A, C146T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C, G7146A,
T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, C10394T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T, T16223C,
G16230A, T16278C, C16311T
I0163 OSH7 H5b none G73A, C146T, C152T, C195T, A247G, C456T, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C,
G5471A, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, T16304C, C16311T, C16519T
I0165 OSH9 U5b1b T152C!, C6644T, C146T, C150T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, A5656G, C6644T,
A14053G, G7146A, T7256C, A7521G, A7768G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T,
T16192C! C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, G12618A, T12705C, G13105A, G13276A, T13506C, T13617C,
T13650C, A14053G, T14182C, A16129G, T16187C, T16223C, G16230A, C16270T, T16278C, C16311T
I0170 OAW1 HV6'17 C16519T G73A, C146T, C152T, C195T, A247G, A769G, A1018G, A825t, A2758G, C2885T, T3594C, G4104A, T4312C, G7146A, T7256C,
A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G,
T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16519T
I0807 ESP30 H1e1a C5960T, G8865A G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A, T4312C,
G5460A, C5960T, T7028C, G7146A, T7256C, A7521G, T8468C, A8512G, T8655C, G8701A, G8865A, C9540T, G10398A, T10664C,
A10688G, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C,
C14902T, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T
I0822 HQU3 K1e G1018A!, C10757T C146T, C195T, A247G, A769G, A825t, T1189C, A1811G, T1819C, A2758G, C2885T, A3480G, T3594C, G4104A, T4312C, G7146A,
T7256C, A7521G, G8251A, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, C10478T, A10550G, T10664C, A10688G,
C10757T, C10810T, C10873T, C10915T, T11299C, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A,
T13506C, T13650C, C14142g, C14167T, T14798C, A16129G, T16187C, C16189T, T16223T, T16224C, G16230A, T16278C, T16362C
I0808 HQU4 H7d np 304-311, 3569- G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, A4793G,
3589, 16129-16156, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
16301-16309, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, C15409T, A16129G, T16187C,
16380-16389 no C16189T, T16223C, G16230A, T16278C, C16311T
coverage
I0212 HQU5 T2c1d1 T152C!, C4017T, C195T, A247G, T279C, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, C4017T, G4104A, T4216C, T4312C,
C6340T, G8860A, A4917G, C5187T, G6261A, C6340T, G7146A, T7256C, A7521G, C7873T, T8468C, T8655C, G8697A, G8701A, G8860A, C9540T,
G15110A G10398A, T10463C, T10664C, A10688G, C10810T, C10822T, C10873T, C10915T, A11251G, A11812G, T12705C, G13105A,
G13276A, G13368A, T13506C, T13650C, A14233G, G14905A, G15110A, C15452a, A15607G, G15928A, T16126C, A16129G,
T16187C, C16189T, T16223C, G16230A, T16278C, C16292T, C16294T, C16311T
I0559 QLB15D HV6'17 C16519T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A, T4312C, G7146A, T7256C,
A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G,

12
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16519T
I0560 QLB18A T2e1 none C41T, C146T, C150T, C152T, C195T, A247G, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A,
T4216C, T4312C, A4917G, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C,
T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, G16153A, T16187C, C16189T, T16223C,
G16230A, T16278C, C16294T, C16311T
I0556 QLB2A U8a1 T7660C, T16342C, C146T, C152T, C195T, A247G, T282C, A769G, A825t, A1018G, A1811G, A2758G, C2885T, T3594C, C3738T, G4104A, T4312C,
T16342C T6392C, C6455T, A7055G, G7146A, T7256C, A7521G, T7660C, T8468C, T8655C, G8701A, C9365T, C9540T, T9698C, G10398A,
T10664C, A10688G, C10810T, C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13145A,
G13276A, T13506C, T13650C, A16129G, T16187C, C16189T, T16223T, G16230A, T16278C, C16311T, T16342C, C16519T
I0557 QLB6B U5b2a2 T16192C!, C146T, C150T, C152T, C195T, A247G, A769G, A825t, A1018G, C1721T, A2758G, C2885T, T3197C, C3212T, T3594C, G4104A,
C16519T T4312C, A4732G, G7146A, T7256C, A7521G, A7768G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G,
C10810T, C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C,
A13637G, T13650C, T14182C, A16129G, T16187C, T16223C, G16230A, C16270T, T16278C, C16311T, G16398A, C16519T,
I0548 BENZ14 U5a2b4 C16519T C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, G7146A, T7256C,
A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G9548A, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G, G15301A!!,
A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, C16519T, G16526A
I0549 BENZ18/BEN W1c'i none T119C, C146T, C152T, A189G, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, A2758G, C2885T, A3505G,
Z15? T3594C, G4104A, T4312C, G5046A, G5460A, G7146A, T7256C, A7521G, C7864T, G8251A, T8468C, T8655C, G8701A, G8994A,
C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G, A11947G, T12414C, G13105A, G13276A,
T13506C, T13650C, G15884c, A16129G, T16187C, C16189T, G16230A, T16278C, C16292T, C16311T
I0172 ESP24 T2b T12811C C146T, C152T, C195T, A247G, G709A, A769G, A825t, G930A, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, G5147A, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, T12811C, G13105A, G13276A, G13368A,
T13506C, T13650C, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T, T16223C,
G16230A, T16278C, C16294T, C16296T, T16304C, C16311T
I0175 BAL16 N1a1a1 C152T!!, T8227C, C146T, C152T!!, C195T, T199C, T204C, A247G, 573.XC, T669C, A769G, A825t, A1018G, G1719A, G2702A, A2758G, C2885T,
G10143A, T3336C, T3594C, G4104A, T4312C, A5315G, G7146A, T7256C, A7521G, T8227C, T8468C, T8655C, G8701A, A8901G, C9540T,
A11884G, G10143A, T10238C, T10664C, A10688G, C10810T, C10873T, C10915T, A11884G, A11914G, G12501A, G13105A, G13276A,
T16355C! T13506C, T13650C, A13780G, G15043A, A16129G, C16147A, T16172C, T16187C, C16189T, G16230A, C16248T, T16278C, C16311T
I0551 SALZ3B U3a1 none C146T, C150T, C152T, C195T, A247G, A769G, A825t, A1018G, A1811G, A2294G, A2758G, C2885T, G3010A, T3594C, G4104A,
T4312C, T4703C, C6518T, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9266A, C9540T, G10398A, A10506G, T10664C,
A10688G, C10810T, C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C,
T13650C, C13934T, A14139G, T15454C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, A16343G, G16390A
I0800 SALZ57A H3 T152C!, A5515G G73A, C146T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, A5515G, T6776C,
T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0802 SALZ77A H3 none G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T6776C,
T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,

13
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
T16223C, G16230A, T16278C, C16311T
I0552 SALZ7A H5 C5993T, A14566G, G73A, C146T, C152T, C195T, A247G, C456T, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C,
C16519T C5993T, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, A14566G, T14766C, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, T16304C, C16311T, C16519T
I0554 SALZ88A J1c A235G, T4083C, C146T, C152T, C195T, G228A, A235G, A247G, C295T, C462T, T489C, A769G, A825t, A1018G, A2758G, C2885T, G3010A, T3594C,
A7049G, A15799G, T4083C, G4104A, T4216C, T4312C, A7049G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, T10664C, A10688G,
C16519T C10810T, C10873T, C10915T, A11251G, A11914G, A12612G, T12705C, G13105A, G13276A, T13506C, T13650C, G13708A,
T14798C, C15452a, A15799G, C16069T, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16519T,
I0798 SALZ18A H10e'f'g C13503T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11719G, A11914G, T12705C, G13105A, G13276A, C13503T, T13506C, T13650C, T14470a, T14766C, T16093C, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, C16311T
I0799 SALZ21B H1e T1766C G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, T1766C, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A,
T4312C, G5460A, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, C16311T
I0408 Mina18a pre-U5b1i T13656C, C146T, C150T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, A3105G, T3197C, T3594C, G4104A, T4312C,
T16209C, A16399G A5656G, G7146A, T7256C, A7521G, A7768G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T,
(topologically C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C,
missing C3498T, T13656C, T14182C, A16129G, C16167T, T16187C, C16189T, C16192T, T16209C, T16223C, G16230A, C16270T, T16278C, C16311T,
T6674G, G15777A, A16399G
T16311C, T16356C
I0404 Mina2 J2a1a1 T195C!, A10499G, C146T, C150T, A215G, A247G, C295T, T319C, T489C, G513A, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A,
G10586A, T4216C, T4312C, G7146A, T7256C, C7476T, A7521G, G7789A, T8468C, T8655C, G8701A, C9540T, A10499G, G10586A, T10664C,
G11377A, A10688G, C10810T, C10873T, C10915T, A11251G, G11377A, A11914G, A12612G, T12705C, G13105A, G13276A, T13506C,
C16519T; T13650C, G13708A, A13722G, A14133G, G15257A, C15452a, C16069T, T16126C, A16129G, G16145A, T16187C, C16189T,
heteroplasmy at np T16223C, G16230A, T16231C, C16261T, T16278C, C16311T, C16519T
16274 (54% A, 46%
G)
I0405 Mina3 K1a1b1 C16301T; C114T, C146T, C152T, C195T, A247G, C497T, A769G, A825t, A1018G, T1189C, A1811G, A2758G, C2885T, A3480G, T3594C,
G4104A, T4312C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, A10550G, T10664C, A10688G,
C10810T, C10873T, C10915T, T11299C, A11467G, A11470G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C,
T13650C, C14167T, T14798C, A15924G, T16093C, A16129G, T16187C, C16189T, T16223T, T16224C, G16230A, T16278C, C16301T
I0406 Mina4 H1 C722T G73A, C146T, C152T, C195T, A247G, C722T, A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A,
T4312C, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0407 Mina6b K1b1a1 none C146T, C195T, A247G, A769G, A825t, A1018G, T1189C, A1811G, A2758G, C2885T, A3480G, T3594C, G4104A, T4312C, G5913A,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, G9962A, A10289G, A10550G, T10664C, A10688G,
C10810T, C10873T, C10915T, T11299C, A11467G, A11914G, A11923G, A12308G, G12372A, T12705C, G13105A, G13276A,
T13506C, T13650C, C13967T, C14167T, T14798C, G15257A, C15946T, T16093C, A16129G, T16187C, C16189T, T16223T, T16224C,

14
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
G16230A, T16278C, G16319A, A16463G
I0370 SVP10 H13a1a1 C14809T, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, C2259T, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C,
C16261T, C16519T A4745G, T7028C, G7146A, T7256C, G7337A, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13326C, T13506C, T13650C, C13680T, T14766C,
C14809T, C14872T, A16129G, T16187C, C16189T, T16223C, G16230A, C16261T, T16278C, C16311T, C16519T
I0355 SVP2 K1b2a none C152T, A247G, A769G, A825t, A1018G, T1189C, A1811G, A2758G, C2885T, A3480G, T3594C, G4104A, T4312C, G5913A, G7146A,
T7256C, A7521G, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, A10550G, T10664C, A10688G, C10810T, C10873T,
C10915T, T11299C, A11467G, A11914G, A12308G, G12372A, T12705C, T12738g, G12771A, G13105A, G13276A, T13506C,
T13650C, C14167T, T14798C, A16129G, T16187C, C16189T, T16223T, T16224C, G16230A, T16278C
I0231 SVP3 U4a1 T7153C C146T, A247G, G499A, A769G, A825t, A1018G, A1811G, A2758G, C2885T, T3594C, G4104A, T4312C, T4646C, T5999C, A6047G,
G7146A, T7153C, T7256C, A7521G, T8468C, T8655C, G8701A, C8818T, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, C11332T, A11467G, A11914G, A12308G, G12372A, T12705C, A12937G, G13105A, G13276A, T13506C, T13650C,
C14620T, T15693C, A16129G, C16134T, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, T16356C
I0429 SVP38 T2c1a2 none C146T, C152T, C195T, A247G, 573.XC, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, C5817T, G6261A, G7146A, T7256C, A7521G, C8455T, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A,
T10463C, T10664C, A10688G, C10810T, C10822T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A,
G13276A, G13368A, T13506C, T13650C, A13973t, A14233G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G,
T16187C, C16189T, T16223C, G16230A, T16278C, C16292T, C16294T, C16311T
I0357 SVP5 W6c G7852A C146T, C152T, A189G, C194T, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, A2758G, C2885T, A3505G,
T3594C, A4093G, G4104A, T4312C, G5046A, G5460A, G7146A, T7256C, A7521G, G7852A, C8002T, G8251A, T8468C, T8614C,
T8655C, C8658T, G8701A, G8994A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G,
A11947G, T12414C, G13105A, G13276A, T13506C, T13650C, G15884c, A16129G, T16187C, C16189T, C16192T, G16230A,
T16278C, C16292T, C16311T, T16325C
I0438 SVP50 U5a1a1 C16519T C146T, C152T, C195T, A247G, A769G, A825t, A1018G, T1700C, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, T5495C,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G,
A15218G, A15924G, A16129G, T16187C, C16189T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, A16399G, C16519T
I0439 SVP52 U5a1a1 C16519T C146T, C152T, C195T, A247G, A769G, A825t, A1018G, T1700C, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, T5495C,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, A14793G,
A15218G, A15924G, A16129G, T16187C, C16189T, T16223C, G16230A, C16256T, C16270T, T16278C, C16311T, A16399G, C16519T
I0441 SVP54 H2b C16519T G73A, C146T, C195T, A247G, A769G, A825t, A1018G, G1438A, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C,
G7146A, T7256C, A7521G, T8468C, T8598C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16519T
I0443 SVP57 W3a1a A15951G C146T, C152T, A189G, C194T, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, T1406C, A2758G, C2885T,
A3505G, T3594C, G4104A, T4312C, G5046A, G5460A, G7146A, C7151T, T7256C, A7521G, G8251A, T8468C, T8655C, G8701A,
G8994A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G, A11947G, T12414C, G13105A,
A13263G, G13276A, T13506C, T13650C, T15784C, G15884c, A15951G, A16129G, T16187C, C16189T, G16230A, T16278C,
C16292T, C16311T
I0444 SVP58 H6a1b C16519T G73A, C146T, C152T, C195T, T239C, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G3915A, G4104A,
T4312C, A4727G, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9380A, C9540T, G10398A, G10589A, T10664C,

15
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
A10688G, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C,
A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, T16362C, A16482G, C16519T
I0051 ALB2 H3b np 2118-2125, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2581G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C,
4771, 10026-10069 T6776C, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
no coverage C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0118 ALB3 HV6'17 C12112T, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A, T4312C, G7146A, T7256C,
C16519T, G16526A A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G,
C12112T, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16519T, G16526A
I0805 QLB26 H1 none G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A, T4312C,
T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0806 QLB28 H1 none G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A, T4312C,
T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0113 QUEXII4 J1c5 T14470C, C16519T C146T, C152T, G185A, C195T, G228A, A247G, C295T, C462T, T489C, A769G, A825t, A1018G, A2758G, C2885T, G3010A, T3594C,
G4104A, T4216C, T4312C, A5198G, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, T10664C, A10688G, C10810T,
C10873T, C10915T, A11251G, A11914G, A12612G, T12705C, G13105A, G13276A, T13506C, T13650C, G13708A, T14470C,
T14798C, C15452a, C16069T, T16126C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, , C16311T, C16519T,
I0112 QUEXII6 H13a1a2 G9025A, C16519T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, C2259T, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C,
A4745G, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9025A, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, A13542G, T13650C, C13680T, T14766C,
C14872T, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16519T
I0060 ROT3 K1a2c T146C!, T152C!, C195T, A214G, A247G, C497T, A769G, A825t, A1018G, A1811G, T1189C, A2758G, C2885T, A3480G, T3594C, G4104A, T4216C,
A214G, G14305A, T4312C, G7146A, T7256C, A7521G, G7775A, T8468C, T8655C, G8701A, G9055A, C9540T, T9698C, A10550G, T10664C, A10688G,
T14311C, T16086C C10810T, C10873T, C10915T, T11025C, T11299C, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A,
T13506C, T13650C, C14167T, G14305A, T14311C, T14798C, T16086C, A16129G, T16187C, C16189T, T16223C, T16224C, G16230A,
T16278C
I0111 ROT4 H3new C4577T, C16256T G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, C4577T,
T6776C, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, C16256T, T16278C, C16311T
I0108 ROT6 H5a3 C16519T G73A, C146T, C152T, C195T, A247G, C456T, G513A, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A,
T4312C, T4336C, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T,
C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, G15884A, A16129G,
T16187C, C16189T, T16223C, G16230A, T16278C, T16304C, C16311T, C16519T
I0171 BZH12 U5a1a2a C16519T C146T, C152T, C195T, A247G, 573.XC, A769G, A825t, A1018G, T1700C, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C,
A5319G, A6629G, T6719C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G,
C10810T, C10873T, C10915T, A11467G, A11914G, A12308G, C12346T, G12372A, T12705C, G13105A, G13276A, T13506C,

16
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
T13617C, T13650C, A14793G, A15218G, A16129G, T16187C, C16189T, T16223C, G16230A, C16256T, C16270T, T16278C,
C16311T, A16399G, C16519T
I0058 BZH4 H1e A15220G, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A, T4312C,
A15401G, G5460A, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T,
A16293G C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A15220G, A15401G,
A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, A16293G, C16311T
I0059 BZH6 H1/H1b'ad A8149G, A9377G, G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3010A, T3594C, G4104A, T4312C,
T9467C, A13671G, T7028C, G7146A, T7256C, A7521G, A8149G, T8468C, T8655C, G8701A, A9377G, T9467C, C9540T, G10398A, T10664C, A10688G,
T14319C, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, A13671G, T14319C,
T16189C! T14766C, A16129G, T16187C, T16223C, G16230A, T16278C, C16311T
I0104 ESP11 U4b1a1a1 none C146T, C152T, A247G, A499G, A769G, A825t, A1018G, A1811G, T2083C, A2758G, C2885T, T3594C, A3672G, G4104A, T4312C,
T4646C, T5999C, A6047G, G7146A, T7256C, A7521G, T7705C, T8468C, A8642G, T8655C, G8701A, C9540T, G10398A, T10664C,
A10688G, C10810T, C10873T, C10915T, C11332T, T11339C, A11467G, A11914G, T12297C, A12308G, G12372A, T12705C,
G13105A, G13276A, T13506C, T13650C, C14620T, T15693C, C15789T, A16129G, T16187C, C16189T, T16223C, G16230A,
T16278C, C16311T, T16356C, T16362C
I0103 ESP16 W6a none C146T, C152T, A189G, C194T, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, A2758G, C2885T, A3505G,
T3594C, A4093G, G4104A, T4312C, G5046A, G5460A, G7146A, T7256C, A7521G, G8251A, T8468C, T8610C, T8614C, T8655C,
G8701A, G8994A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G, A11947G, T12414C,
G13105A, G13276A, T13506C, T13650C, G15884c, A16129G, T16187C, C16189T, C16192T, G16230A, T16278C, C16292T, C16311T,
T16325C
I0049 ESP22 X2b4 none C146T, C152T, A153G, G225A, T226C, A247G, A769G, A825t, A1018G, G1719A, A2758G, C2885T, T3594C, G3705A, G4104A,
T4312C, T6221C, C6371T, G7146A, T7256C, A7521G, C8393T, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G,
C10810T, C10873T, C10915T, A11914G, G13105A, G13276A, T13506C, T13650C, G13708A, A13966G, T14470C, G15927A,
A16129G, T16187C, G16230A, C16311T
I0106 ESP26 T2a1b1 none C146T, C152T, C195T, A247G, G709A, A769G, A825t, A1018G, G1888A, T2141C, A2758G, C2885T, T3594C, G4104A, T4216C,
T4312C, A4917G, G7146A, T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, G10398A, T10463C, T10664C, A10688G,
C10810T, C10873T, C10915T, A11251G, A11812G, A11914G, T12705C, G13105A, G13276A, G13368A, T13506C, T13650C,
T13965C, A13966G, A14233G, A14687G, G14905A, C15452a, A15607G, G15928A, T16126C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16294T, C16311T, T16324C
I0050 ESP5 U5a2d T5999C, A9389G, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, T5999C, G7146A,
T13474C, T7256C, A7521G, A7843G, T8468C, T8655C, G8701A, A9389G, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T,
G14544A, C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13474C, T13506C, T13617C,
A14587G, T13650C, G14544A, A14587G, A14793G, C14854T, A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T,
C14854T, C16519T C16270T, T16278C, C16311T, C16519T
I0550 KAR22A T1a1 none C146T, A247G, G709A, A769G, A825t, A1018G, G1888A, A2758G, C2885T, T3594C, G4104A, T4216C, T4312C, A4917G, G7146A,
T7256C, A7521G, T8468C, T8655C, G8697A, G8701A, C9540T, T9899C, G10398A, T10463C, T10664C, A10688G, C10810T,
C10873T, C10915T, A11251G, A11914G, C12633a, T12705C, G13105A, G13276A, G13368A, T13506C, T13650C, G14905A,
C15452a, A15607G, G15928A, T16126C, A16129G, A16163G, C16186T, T16187C, T16223C, G16230A, T16278C, C16294T, C16311T
I0117 ESP29 I3a none C146T, C195T, T199C, T204C, G207A, T239C, A247G, T250C, 573.XC, A769G, A825t, A1018G, G1719A, A2758G, C2885T, T3594C,
G4104A, T4312C, A4529t, G7146A, T7256C, A7521G, G8251A, T8468C, T8655C, G8701A, C9540T, T10034C, T10238C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C, A13780G, G15043A,
A15924G, T16086C, T16187C, C16189T, G16230A, T16278C, C16311T, G16391A

17
Lab ID Individual ID mtDNA-hg call Private mutations SNPs against RSRS
I0115 ESP3 U5a1 C4796T, C12103A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C, C4796T, G7146A,
C14003T, T7256C, A7521G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A14893G, A11467G, A11914G, C12103A, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C, T13650C, C14003T,
T14971C, C16519T A14793G, A14893G, T14971C, A15218G, A16129G, T16187C, C16189T, C16192T, T16223C, G16230A, C16256T, C16270T,
T16278C, C16311T, A16399G, C16519T
I0116 ESP4 W3a1 C5211T, G6267A, C146T, C152T, A189G, C194T, T204C, G207A, A247G, G709A, A769G, A825t, A1018G, T1243C, T1406C, A2758G, C2885T,
T14025C, C16147G A3505G, T3594C, G4104A, T4312C, G5046A, C5211T, G5460A, G6267A, G7146A, T7256C, A7521G, G8251A, T8468C, T8655C,
G8701A, G8994A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, C11674T, A11914G, A11947G, T12414C,
G13105A, A13263G, G13276A, T13506C, T13650C, T14025C, T15784C, G15884c, A16129G, C16147G, T16187C, C16189T,
G16230A, T16278C, C16292T, C16311T
I0803 EUL41A H4a1a1 C13545T, C16519T C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, C3992T, A4024G, G4104A, T4312C,
T5004C, T7028C, G7146A, T7256C, A7521G, G8269A, T8468C, T8655C, G8701A, G9123A, C9540T, A10044G, G10398A, T10664C,
A10688G, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, C13545T, T13650C,
C14365T, A14582G, T14766C, A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16519T
I0804 EUL57B H3 T152C! G73A, T152C!, C146T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T6776C,
T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T
I0047 HAL16 V G207A, A14841G T72C, G73A, C146T, C152T, C195T, G207A, A247G, A769G, A825t, A1018G, A2758G, C2885T, T3594C, G4104A, T4312C, 4580A,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T,
A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A14841G, C15904T, A16129G, T16187C,
C16189T, T16223C, G16230A, T16278C, T16298C, C16311T, C16519T,
I0164 QUEVIII6 U5b2a1 T5918C, T16192C!, C146T, C150T, C152T, C195T, A247G, A769G, A825t, A1018G, C1721T, A2758G, C2885T, T3197C, T3594C, G4104A, T4312C,
(pre-U5b2a1b) C16519T A4732G, T5918C, G7146A, T7256C, A7521G, A7768G, T8468C, T8655C, G8701A, G9477A, C9540T, G10398A, T10664C, A10688G,
C10810T, C10873T, C10915T, A11467G, A11914G, A12308G, G12372A, T12705C, G13105A, G13276A, T13506C, T13617C,
A13637G, T13650C, T14182C, A16129G, T16187C, T16223C, G16230A, T16278C, C16311T, C16519T
I0114 ESP2 I3a none C146T, C195T, T199C, T204C, G207A, T239C, A247G, T250C, 573.XC, A769G, A825t, A1018G, G1719A, A2758G, C2885T, T3594C,
G4104A, T4312C, A4529t, G7146A, T7256C, A7521G, G8251A, T8468C, T8655C, G8701A, C9540T, T10034C, T10238C, T10664C,
A10688G, C10810T, C10873T, C10915T, A11914G, G12501A, G13105A, G13276A, T13506C, T13650C, A13780G, G15043A,
A15924G, T16086C, T16187C, C16189T, G16230A, T16278C, C16311T, G16391A
I0099 HAL36C H23 none G73A, C146T, C152T, C195T, A247G, A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G4104A, T4312C, T7028C,
G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, C10211A, G10398A, T10664C, A10688G, C10810T, C10873T,
C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13650C, T14766C, A16129G, T16187C, C16189T,
T16223C, G16230A, T16278C, C16311T

18
Table S2.2. Haplogroup frequencies of 24 ancient populations used for principal component analysis.
Population abbr. n C Z N* N1a I W X R HV V H T1 T2 J U U2 U3 U4 U5a U5b U8 K Publication (number of samples)
Bramanti et al. 2009 (11)12, Fu et al. 2013 (5)13,
Hunter-gatherer central HGC 28 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 14.3 10.7 0.0 7.1 32.1 32.1 3.6 0.0 Bollongino et al. 2013 (5)14, Lazaridis et al. 2014
(7)15
Haak et al. 2005 (14)16, Bramanti 200817, Haak et
Linear Pottery culture LBK 115 0.0 0.0 0.0 11.3 0.0 3.5 0.9 0.0 5.2 3.5 16.5 0.0 25.2 11.3 0.0 0.0 0.9 0.0 1.7 0.9 0.0 19.1 al. 2010 (22)6, Brandt et al. 2013 (64)8, Lazaridis et
al. 2014 (1)15, this study (8)
Rössen culture RSC 17 0.0 0.0 0.0 5.9 0.0 0.0 5.9 0.0 23.5 5.9 29.4 0.0 11.8 0.0 0.0 0.0 0.0 0.0 0.0 5.9 0.0 11.8 Brandt et al. 2013 (11)8, Lee et al. 201318

Schöningen group SCG 33 0.0 0.0 0.0 3.0 0.0 9.1 3.0 0.0 3.0 0.0 15.2 0.0 12.1 15.2 0.0 0.0 0.0 0.0 0.0 6.1 3.0 30.3 Brandt et al. 2013 (33)8

Baalberge culture BAC 21 0.0 0.0 0.0 4.8 0.0 0.0 9.5 0.0 4.8 0.0 23.8 4.8 23.8 4.8 0.0 0.0 0.0 0.0 0.0 4.8 4.8 14.3 Brandt et al. 2013 (19)8, this study (2)

Salzmünde culture SMC 29 0.0 0.0 0.0 6.9 0.0 0.0 3.4 0.0 3.4 3.4 31.0 0.0 6.9 20.7 0.0 0.0 10.3 0.0 0.0 3.4 0.0 10.3 Brandt et al. 2013 (29)8

Bernburg culture BEC 17 0.0 0.0 0.0 0.0 0.0 5.9 5.9 0.0 0.0 5.9 23.5 0.0 11.8 0.0 0.0 0.0 0.0 0.0 11.8 17.6 0.0 17.6 Brandt et al. 2013 (17)8

Corded Ware culture CWC 44 0.0 0.0 0.0 0.0 2.3 2.3 6.8 0.0 2.3 0.0 22.7 6.8 11.4 9.1 0.0 2.3 0.0 6.8 9.1 4.5 0.0 13.6 Haak et al. 2008 (9)19, Brandt et al. 2013 (35)8

Bell Beaker culture BBC 35 0.0 0.0 0.0 0.0 2.9 8.6 0.0 0.0 0.0 0.0 40.0 5.7 5.7 2.9 0.0 2.9 0.0 5.7 14.3 5.7 0.0 5.7 Brandt et al. 2013 (29)8

Unetice culture UC 94 0.0 0.0 0.0 0.0 12.8 4.3 5.3 1.1 2.1 3.2 21.3 2.1 6.4 6.4 2.1 6.4 0.0 1.1 12.8 2.1 3.2 7.4 Brandt et al. 2013 (94)8

Middle Neolithic Germany MNG 16 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 37.5 0.0 0.0 6.3 6.3 0.0 0.0 0.0 0.0 50.0 0.0 0.0 Bollongino et al. 2013 (16)14

Malmström et al. 200920, Skoglund et al. 2012

Funnel Beaker culture FBC 10 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10.0 0.0 30.0 20.0 0.0 0.0 0.0 0.0 10.0 20.0 0.0 10.0
(3)21, Bramanti et al. 2009 (7)12
Malmström et al. 200920, Skoglund et al. 2012
Pitted Ware culture PWC 19 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10.5 5.3 0.0 0.0 5.3 0.0 0.0 0.0 0.0 42.1 15.8 15.8 0.0 5.3
(19)21
Chandler 200322, Chandler et al. 2005 (7)23,
Hunter-Gatherer south HGS 13 0.0 0.0 15.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 38.5 0.0 0.0 0.0 0.0 0.0 0.0 7.7 0.0 38.5 0.0 0.0 Hervella et al. 2012 (4)24, Sánchez-Quinto et al.
2012 (2)25
Gamba et al. 2011 (11)26, Lacan 2011, Lacan et al.
(Epi)Cardial CAR 23 0.0 0.0 13 4.3 0.0 0.0 4.3 0.0 0.0 0.0 26.1 0.0 13 4.3 0.0 0.0 0.0 0.0 0.0 4.3 0.0 30.4
2011b (7)27, this study (5)

Neolithic Portugal NPO 17 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.9 70.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 17.6 5.9 0.0 0.0 Chandler 2003, Chandler et al. 2005 (17)22,23

Neolithic Basque Country &

NBQ 43 0.0 0.0 0.0 0.0 2.3 0.0 2.3 0.0 2.3 0.0 44.2 0.0 2.3 4.7 25.6 0.0 0.0 0.0 0.0 7.0 0.0 9.3 Hervella 2010, Hervella et al. 2012 (43)24
Navarre

Middle Neolithic Spain MNS 16 0.0 0.0 0.0 0.0 6.3 6.3 0.0 0.0 0.0 0.0 31.3 0.0 12.5 18.8 0.0 0.0 0.0 6.3 0.0 6.3 0.0 12.5 Sampietro et al. 2007 (11)28, this study (5)

Treilles TRE 29 0.0 0.0 0.0 0.0 0.0 0.0 13.8 0.0 6.9 3.4 20.7 0.0 6.9 20.7 3.4 0.0 0.0 0.0 3.4 13.8 0.0 6.9 Lacan 2011, Lacan et al. 2011a (29)29

Bramanti et al. 2009 (2)12, Krause et al. 2010 (1)30,

Hunter-gatherer east HGE 15 20 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.7 0.0 0.0 0.0 0.0 20 0.0 26.7 26.7 0.0 0.0 0.0
Der Sarkissian et al. in 2013 (11)7, this study (1)

Bronze Age Siberia BAS 11 0.0 9.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 9.1 9.1 9.1 0.0 0.0 9.1 0.0 27.3 18.2 0.0 0.0 9.1 Keyser et al. 2009 (11)31

Bronze Age Kazakhstan BAK 8 0.0 0.0 0.0 0.0 12.5 0.0 0.0 0.0 12.5 0.0 12.5 25.0 12.5 0.0 0.0 0.0 0.0 0.0 12.5 12.5 0.0 0.0 Lalueza-Fox et al. 2004 (8)32

Catacomb culture CAT 24 0.0 0.0 0.0 0.0 4.2 0.0 0.0 8.3 4.2 0.0 25 8.3 0.0 8.3 0.0 0.0 0.0 29.2 12.5 0.0 0.0 0.0 Wilde et al. 2014 (24)11

Bronze Age Yamnaya YAM 36 0.0 0.0 0.0 2.8 2.8 8.3 2.8 0.0 0.0 0.0 25 5.6 11.1 5.6 2.8 2.8 0.0 5.6 16.7 0.0 0.0 8.3 Wilde et al. 2014 (26)11, this study (10)

19
References

1 Behar, D. M. et al. A "Copernican" Reassessment of the Human Mitochondrial DNA

Tree from its Root. American Journal of Human Genetics 90, 675-684,
doi:10.1016/j.ajhg.2012.03.002 (2012).
2 van Oven, M. & Kayser, M. Updated comprehensive phylogenetic tree of global
human mitochondrial DNA variation. Human Mutation 30, E386-394,
doi:10.1002/humu.20921 (2009).
3 Brotherton, P. et al. Neolithic mitochondrial haplogroup H genomes and the genetic
origins of Europeans. Nat Commun 4, 1764, doi:10.1038/ncomms2656 (2013).
4 Der Sarkissian, C. et al. Mitochondrial Genome Sequencing in Mesolithic North East
Europe Unearths a New Sub-Clade within the Broadly Distributed Human
Haplogroup C1. PLoS ONE 9, e87612, doi:10.1371/journal.pone.0087612 (2014).
5 Bandelt, H. J., Kong, Q. P., Richards, M. & Macaulay, V. in Human Mitochondrial
DNA and the Evolution of Homo Sapiens (eds H. J. Bandelt, V. Macaulay, & M.
Richards) 47-90 (Springer, 2006).
6 Haak, W. et al. Ancient DNA from European early Neolithic farmers reveals their
near eastern affinities. PLoS Biology 8, e1000536 (2010).
7 Der Sarkissian, C. et al. Ancient DNA Reveals Prehistoric Gene-Flow From Siberia
in the Complex Human Population History of North East Europe. PLoS Genetics 9,
e1003296 (2013).
8 Brandt, G. et al. Ancient DNA reveals key stages in the formation of central
European mitochondrial genetic diversity. Science 342, 257-261,
doi:10.1126/science.1241844 (2013).
9 Szécsényi-Nagy, A. et al. Tracing the genetic origin of Europe's first farmers reveals
insights into their social organization. bioRxiv, doi:http://dx.doi.org/10.1101/008664
(2014).
10 Brandt, G., Szécsényi-Nagy, A., Roth, C., Alt, K. W. & Haak, W. Human
paleogenetics of Europe – The known knowns and the known unknowns. Journal of
Human Evolution available online 13 November 2014,
doi:10.1016/j.jhevol.2014.06.017 (2014).
11 Wilde, S. et al. Direct evidence for positive selection of skin, hair, and eye
pigmentation in Europeans during the last 5,000 y. P Natl Acad Sci USA 111, 4832-
4837, doi:10.1073/pnas.1316513111 (2014).
12 Bramanti, B. et al. Genetic discontinuity between local hunter-gatherers and central
Europe's first farmers. Science 326, 137-140, doi:10.1126/science.1176869 (2009).
13 Fu, Q. et al. A revised timescale for human evolution based on ancient mitochondrial
genomes. Current Biology 23, 553-559, doi:10.1016/j.cub.2013.02.044 (2013).
14 Bollongino, R. et al. 2000 years of parallel societies in Stone Age Central Europe.
Science 342, 479-481, doi:10.1126/science.1245049 (2013).
15 Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for
present-day Europeans. Nature 513, 409-413, doi:10.1038/nature13673 (2014).
16 Haak, W. et al. Ancient DNA from the first European farmers in 7500-year-old
Neolithic sites. Science 310, 1016-1018, doi:10.1126/science.1118725 (2005).
17 Bramanti, B. Genetic analysis of aDNA from sixteen skeletons of the Vedrovice
collection. Anthropologie 46, 153-160 (2008).
18 Lee, E. J. et al. Ancient DNA insights from the Middle Neolithic in Germany.
Archaeological and Anthropological Sciences 6, 199-204, doi:DOI 10.1007/s12520-
013-0173-0 (2014).
19 Haak, W. et al. Ancient DNA, Strontium isotopes, and osteological analyses shed
light on social and kinship organization of the Later Stone Age. P Natl Acad Sci USA
105, 18226-18231, doi:10.1073/pnas.0807592105 (2008).
20 Malmström, H. et al. Ancient DNA reveals lack of continuity between neolithic
hunter-gatherers and contemporary Scandinavians. Current biology 19, 1758-1762,
doi:10.1016/j.cub.2009.09.017 (2009).

20
21 Skoglund, P. et al. Origins and genetic legacy of Neolithic farmers and hunter-
gatherers in Europe. Science 336, 466-469, doi:10.1126/science.1216304 (2012).
22 Chandler, H. Using ancient DNA to link culture and biology in human populations,
University of Oxford, (2003).
23 Chandler, H., Sykes, B. & Zilhao, J. in Actas del III Congreso del Neolitico en la
Peninsula Iberica, Santander, Monografias del Instituto Internacional de
Investigaciones Prehistoricas de Cantabria. (eds P Arias, R Ontanon, & C Garcia-
Monco) 781-786.
24 Hervella, M. et al. Ancient DNA from hunter-gatherer and farmer groups from
Northern Spain supports a random dispersion model for the Neolithic expansion into
Europe. PLoS ONE 7, e34417, doi:10.1371/journal.pone.0034417 (2012).
25 Sanchez-Quinto, F. et al. Genomic affinities of two 7,000-year-old iberian hunter-
gatherers. Current biology 22, 1494-1499, doi:10.1016/j.cub.2012.06.005 (2012).
26 Gamba, C. et al. Ancient DNA from an Early Neolithic Iberian population supports a
pioneer colonization by first farmers. Molecular Ecology 21, 45-56,
doi:10.1111/j.1365-294X.2011.05361.x (2012).
27 Lacan, M. et al. Ancient DNA suggests the leading role played by men in the
Neolithic dissemination. P Natl Acad Sci USA 108, 18255-18259,
doi:10.1073/pnas.1113061108 (2011).
28 Sampietro, M. L. et al. Palaeogenetic evidence supports a dual model of Neolithic
spreading into Europe. Proceedings. Biological sciences / The Royal Society 274,
2161-2167, doi:10.1098/rspb.2007.0465 (2007).
29 Lacan, M. et al. Ancient DNA reveals male diffusion through the Neolithic
Mediterranean route. P Natl Acad Sci USA 108, 9788-9791,
doi:10.1073/pnas.1100723108 (2011).
30 Krause, J. et al. A complete mtDNA genome of an early modern human from
Kostenki, Russia. Current biology 20, 231-236, doi:10.1016/j.cub.2009.11.068
(2010).
31 Keyser, C. et al. Ancient DNA provides new insights into the history of south
Siberian Kurgan people. Human genetics 126, 395-410, doi:10.1007/s00439-009-
0683-0 (2009).
32 Lalueza-Fox, C. et al. Unravelling migrations in the steppe: mitochondrial DNA
sequences from ancient central Asians. Proc Biol Sci 271, 941-947,
doi:10.1098/rspb.2004.2698 (2004).

21
Supplementary Information 3
Description of ancient samples and archaeological context

Wolfgang Haak*, Guido Brandt, Christina Roth, Anna Szécsényi-Nagy, Susanne Friederich,
Harald Meller, Michael Francken, Joachim Wahl, Roberto Risch, Michael Kunst, Manuel A.
Rojo Guerra, Rafael Garrido-Pena, Eszter Bánffy, Kurt W. Alt, Dorcas Brown, David
Anthony and David Reich

* To whom correspondence should be addressed ([email protected])

Overview
The archaeology of prehistoric Europe has revealed numerous material cultures that differed
from each other over space and time. Some cultures left behind artifacts similar to those that
preceded them. Other cultures left behind artifacts that mark a clear break from preceding
local traditions.

The largest single group of samples analyzed in this study was from the transect through time
study from the Early Neolithic to the Early Bronze Age reported in Brandt et al. 20131. We
selected samples from that study based on two criteria: We predominantly chose samples
based on overall DNA quality, which we assessed based on performance in our previous
study (largely PCR success rates, replication of results, and completeness of SNP typing).
From the subset of samples that performed well, we chose ones that could be assigned most
confidently to an archaeological culture via accompanying grave goods/context, orientation of
the burials, available radiocarbon dates, and seemingly typical mtDNA haplotype. In addition,
we added to this study some individuals for which the cultural attribution was less clear, in
that there were no characteristic grave goods or context. These individuals were included to
test the ability of the 390k SNP capture strategy to assign a potentially unknown or
unprovenanced sample to a candidate population in combination with absolute dating. For the
latter set of samples, we obtained 19 new radiocarbon dates at the Curt-Engelholm-Centre for
Archaeometry, Mannheim, Germany.

In what follows we provide archaeological context information for the 69 samples for which
genome-wide data is newly presented in this study (a total of 119 individuals were screened
as detailed in Online Table 1, but here we discuss only the subset that produced genome-wide
data). We present the samples in chronological order, largely following the central European
terminology, i.e. Mesolithic (>6000-5000 BCE), Early Neolithic (6000-4000 BCE), Middle
Neolithic (4000-2800 BCE), Late Neolithic (2800-2200 BCE), and Early Bronze Age (2200-
1600 BCE; Figure 1a). We note that there are different traditional terminology for samples
from the Russian steppe and regions of southern Europe.

We use “BCE” (Before Common Era) to refer to a calibrated date that is estimated based on
archaeological context (e.g. radiocarbon dates from associated faunal remains), and “calBCE”
to refer to a calibrated radiocarbon date range (1σ) directly from human skeletal remains,
followed by the respective dating laboratory number.

22
Hunter-gatherer samples
We generated data on two hunter-gatherer samples from western Russia (the easternmost part
of Europe) to bridge the geographical gap between the West European site Loschbour
(WHG)2 and the Siberian site Mal’ta in Siberia, Russia3, from which hunter-gatherer genetic
data are available from the literature.

The individual we refer to as ‘Karelia’ in this study is

x UZ0074/I0061 (MAE RAS collection number 5773-74, grave number 142)
from the ~5500 BCE Mesolithic site Yuzhnyy Oleni Ostrov (island in Lake Onega) in
Karelia, Western Russia, for which a complete mitochondrial genome was published
recently4. Mitochondrial HVS I data from eight other individuals from the same site have also
been described5.

The individual we refer to as ‘Samara hunter-gatherer’

x I0124/SVP44 (5640-5555 calBCE, Beta-392490)
is an adult male from grave 1 in a Neolithic-Eneolithic settlement producing artifacts from the
Elshanka, Samara, and Repin cultures. The specific site is Lebyazhinka IV, on the Sok River,
Samara oblast, Russia. (‘Neolithic’ here refers to the presence of ceramics, not to
domesticated animals or plants.) The radiocarbon date of this individual, based on a femur, is
centuries before the appearance of domesticated animals in the middle Volga region.
Lebyazhinka IV and the neighboring Lebyazhinka V site were occupied seasonally by
multiple cultures between 7000-3500 BCE; a few graves were found in the settled areas6.

Early Neolithic
The Early Neolithic in Europe in this study is represented by new samples reported from sites
in Hungary, Germany and Spain.

The central European Neolithic (6200–3950 BCE) is first manifested in the Starčevo culture,
the Transdanubian Linienbandkeramik (LBKT), and the central European distribution of the
Linienbandkeramik (LBK); these were the first people in the region to exploit agriculture and
animal husbandry7. The Linienbandkeramik appears earliest in the archaeological record of
western Hungary (Transdanubia), where it incorporated novel technologies and cultural
elements from the preceding, southeastern Starčevo culture, which in turn also shows
similarities in material culture to early farming groups further southeast, including Anatolia,
the Levant and the Near East. During the Neolithic transition, the LBK expanded relatively
rapidly along the major waterways and fertile Loess plains towards central Europe, extending
as far as the Paris Basin in the West and Ukraine in the East8,9.

Starčevo in Hungary: Alsónyék-Bátaszék, Mérnöki telep

The Alsónyék-Bátaszék, Mérnöki telep site was excavated by the Institute of Archaeology,
Hungarian Academy of Sciences, between 2006 and 2009 (excavators: A. Osztás, I. Zalai-
Gaál). This site is a part of the site complex “Alsónyék”, which forms the largest
agglomeration of Neolithic settlements and cemeteries in Hungary. Alsónyék is situated along
the Tolna county part of the M6 motorway track in south Hungary. At Mérnöki telep, traces
of the Early Neolithic Starčevo culture and other Neolithic features, including LBK and the
Late Neolithic Lengyel culture, came to light across a total of 80 hectares. Out of the 1568
archaeological features, more than 400 could be associated with the Starčevo culture
including 26 burials. The majority of features were pits, in various shapes and sizes; however,

23
ditches and various oven types were also found. The skeletal remains were all found inside
ovens or in pits, and were all in flexed positions. Only one burial contained grave goods10. All
26 Starčevo skeletons were investigated as part of the German Research Foundation (DFG)
funded bioarchaeological (ancient DNA and stable isotopes) project entitled “The population
history of the Carpathian Basin in the Neolithic period and its influence on the colonization
of central Europe”. Skeletons in uncertain stratigraphic positions, and buried without any
grave goods, were radiocarbon dated at the CEZA laboratory in Mannheim, Germany. Of
these, individual
x BAM25a/I0174 (feature 1532, 5710-5550 calBCE, MAMS 11939)
was included in this study.

Transdanubian LBK in Hungary: Szemely-Hegyes

The south-Hungarian site Szemely-Hegyes is located in Baranya county along the M6
motorway. The site was excavated between 2006 and 2007 by a commercial archaeology
contractor (Ásatárs Ltd; excavators T. Paluch and K. Somogyi). Approximately 1400
archaeological features were documented across 4 hectares, including finds from the LBK,
Sopot culture, Copper Age Balaton-Lasinja, and Furchenstich groups11 (T. Paluch, K.
Somogyi, J. Jakucs personal communication), including ten Neolithic graves. An LBK
settlement with 20 houses, several pits, ditches and ovens was also uncovered. Six out of the
ten Neolithic graves could be assigned to either the Transdanubian LBK or to the Sopot
period, and the skeletons were radiocarbon dated at Beta Analytic in Miami, USA. Sample
x SZEH4b/I0176 (feature 1001, 5210-4940 calBCE, Beta - 310038)
associated with the Transdanubian LBK was analyzed here.

LBK in Germany: Halberstadt-Sonntagsfeld

The most intensively studied region in this study is Mittelelbe-Saale (MES) in Saxony-
Anhalt, Germany. MES is situated within a tight network of waterways along ancient
established trade routes. The region is climatically and geologically well characterized and
benefits from its location in the rain-shadow of the nearby Harz Mountains, its deposits of
ores and salts, and its fertile black and para-brown soils12-14. The earliest MES culture
sampled in this study is the LBK.
The LBK Halberstadt-Sonntagsfeld site was discovered during construction of a new
development and was excavated between 1999 and 2002, uncovering 1324 archaeological
features across an area of 9947m2. The majority of the finds could be attributed to the LBK,
with contours and remnants of seven long houses, several pits and 42 graves15. The remaining
finds could be attributed to the Middle Neolithic Bernburg culture as well as the Unetice and
Urnfield cultures of the Bronze Age. The site is a classical example of LBK settlement
burials, where the majority of graves were grouped around five long houses, and one group of
six graves located inside the central yard area of the settlement. Skeletal material from the site
is well preserved and led to mitochondrial DNA results for 39 individuals1. We included
seven individuals clearly associated with an LBK house:
x HAL5/I0046 (grave 2, feature 241.1, 5206-5004 calBCE, MAMS 21479)
x HAL25/I0048 (grave 28, feature 861, 5206-5052 calBCE, MAMS 21482)
x HAL14/I0056 (grave 15, feature 430, 5206-5052 calBCE, MAMS 21480)
x HAL34/I0057 (grave 38, feature 992, 5207-5067 calBCE, MAMS 21483)
x HAL4/I0100 (grave 1, feature 139, 5032-4946 calBCE, KIA40341)

24
 x HAL2/I0659 (grave 35, feature 999, 5079-4997 calBCE, KIA40350 and 5066-4979
calBCE, KIA30408; Fig. S3.1)
x HAL24/I0821 (grave 27, feature 867, 5034-4942 calBCE, KIA40348)
Radiocarbon dates obtained from these samples confirmed the attribution to the older and
intermediate phase of the LBK period.

Figure S3.1. Skeleton HAL2 (I0659 in this study) from grave 35 at the Halberstadt-
Sonntagsfeld site (Photo: J. Lipták, LDA Sachsen-Anhalt, Germany).

LBK in Germany: Oberwiederstedt-Unterwiederstedt

Several sites around Wiederstedt, Mansfeld-Südharz, and Saxony-Anhalt were excavated in
the late 20th century. Promontory areas above the Wipper River in the Harz region in
Hettstedt were repeatedly used during LBK times, as reflected in traces of settlements. In
addition, excavations in 2000 revealed a mass grave of 12 individuals of which individual
x UWS4/I0054 (grave 6, feature 1 14, 5209-5070 calBCE, MAMS 21485)
was included in this study16.

LBK in Germany: Karsdorf

The site of Karsdorf is located in the valley of Unstrut, Burgenlandkreis, Saxony-Anhalt,
Germany. The slope on which Karsdorf is located is characterized by alluvial loess. The place
itself was settled intensively since the earliest phase of the LBK in the region. The settlement
area is at least 50 acres in size and nearly 30 houses have been excavated. So-called
‘settlement burials’ were regularly found in pits in the center of the settlement area, of which
individual
x KAR6/I0795 (feature 170, 5207-5070 calBCE, MAMS 22823)
was sampled for this study.

25
LBK in Germany: Stuttgart-Mühlhausen
The site Stuttgart-Mühlhausen “Viesenhäuser Hof” was excavated in several campaigns
between 1931 and 1993. It reflects a long period of habitation starting from the Early
Neolithic to the Iron Age. The Early Neolithic is represented by a large number of well-
preserved burials from the LBK, separated into two areas including burials from the early
phases of the LBK (area-2; 5500-5300 BCE) and the middle and later phases (area-1). Three
individuals from area-2 are newly reported in this study (a fourth LBK sample from the
Viesenhäuser Hof site was previously sequenced to high coverage and is co-analyzed with the
newly sequenced LBK in this study2). Based on morphology, individual
x LBK1976/I0022 (grave II-106, area-2, 5500-5300 BCE)
is a female who died at an estimated age of 50-65 years. The skeleton derives from a grave
that was excavated among 90 others from area-2 of the cemetery. The woman was buried in
the characteristic way of the LBK, lying in flexed position on the right side, but lacks any
grave goods. The burial was oriented from SW-NE with the skull facing NW. Most of the
body parts were represented and in anatomical order.
The second well-preserved skeleton
x LBK1992/I0025 (grave II-113, area-2, 5500-5300 BCE)
is a male who died at an estimated age of 50-70 years. The skeleton was found prone in the
grave with the position of the left leg indicated a flexed position. The burial was oriented
from SE-NW with the skull facing south. The grave only contained a single pottery sherd
found in the filling of the grave.
The third skeleton is a female
x LBK2155/I0026 (grave II-125, area-2, 5500-5300 BCE)
who died at 35-50 years. This individual was found prone with flexed legs imitating a seated
position. The grave was oriented from W-E with the skull facing north. The only associated
finding was a pottery fragment from the filling.

Epicardial Neolithic in Spain: Els Trocs

The Pyrenean cave site of Els Trocs (close to San Feliu de Veri, Bisaurri (Huesca), Spain) is
located in the Upper Ribagorza of Aragón at 1561 meters above sea level, inside a conical
hill, which dominates the surrounding ‘Selvaplana’ plain, traditionally used for pasture and
cultivation. The entrance to the cave (UTM coordinates: x-298.198, y-4.702.955, z-1561) is
oriented to the south and hidden by abundant vegetation and huge stone boulders, which
partially close the entrance (currently 2.3 m high and 1.8 m wide). After a steep access ramp,
a chamber of 15 m of length and 6 m of maximum width opens up, in which the temperature
remains stable at 6 and 8°C throughout the year. The excavations carried out inside the cave
have uncovered a stratigraphic sequence that in combination with a series of radiocarbon
dates characterize the following four phases of human occupation17:
TROCS I: the first occupation of the cave, dated to 5300-5000 BCE.
TROCS II: the Middle Neolithic horizon, dated around 4500 BCE.
TROCS III: intensive use spanning >1000 years between 4000-2900 BCE.
TROCS IV: upper layer of the cave from historical times (100 BCE) to today.

Human bones were discovered from all four phases, but especially from Trocs I during which
they appear to be associated with a marked and complex ritual behavior, which involved

26
manipulation of skeletal elements and the deposition of abundant faunal remains (mainly
young lambs) that had been consumed.

The first occupation phase of the cave (Trocs I) is related to the so-called Epicardial tradition.
However, this attribution is currently under re-consideration18 since both the incised-
impressed-grooved decorative techniques and the Cardial pottery types are assumed to be two
variants of the same archaeological culture responsible of the initial Neolithization of Iberia.
The samples successfully analyzed for this study are
x Troc1/I0409 (5311-5218 cal BCE, MAMS 16159)
x Troc3/I0410 (5178-5066 cal BCE, MAMS 16161)
x Troc4/I0411 (5177-5068 cal BCE, MAMS 16162)
x Troc5/I0412 (5310-5206 cal BCE, MAMS 16164)
x Troc7/I0413 (5303-5204 cal BCE, MAMS 16166)
The genetic data indicates that Troc4 was a close relative of Troc3. We excluded the data
from Troc4 from many genome-wide analyses since Troc3 was higher coverage.

Middle Neolithic
Baalberge, Salzmuende, and Bernburg in Germany: Esperstedt
Esperstedt is located in the Merseburg-Querfurt district, Saxony-Anhalt, and is situated in the
centre of the Querfurter Platte, a plain that is primarily loess.

The single grave of individual

x ESP30/I0807 (feature 6220; 3887-3797 calBCE, Er7784)
discovered in 2004 during investigations in advance to major roadworks could be assigned to
the Baalberge group.

Site 4b is located east of the Weida near Esperstedt and forms part of large-scale excavations
that were initiated in 2005 in the context of major infrastructural roadworks in Saxony-
Anhalt, Germany. Site 4b revealed mostly finds from the Middle Neolithic Salzmünde and
Bernburg cultures, viewed as regional forms of the Funnel Beaker tradition (or
Trichterbecherkultur TRB), which defines the Middle Neolithic in central Europe. The
genetic results from individual
x ESP24/I0172 (3360-3086 calBCE, Erl8699)
were retrieved from skeletal remains found in a settlement pit that was attributed to the
Bernburg culture, but had no associated grave goods. Radiocarbon dating from the >60 year-
old male individual matched both the temporally overlapping phases of the Salzmünde (3400-
3025 BCE) and Bernburg (3100-2650 BCE) cultures19.

Baalberge in Germany: Quedlinburg

The site Quedlinburg, Harzkreis, is situated in the fertile foothills of the northern Harz, a
region characterized by rich loess soils. The specific site Lehofsberg (Reference site 9) was
excavated during major roadworks for highway B6n, at which archaeological excavations
revealed a small graveyard with a dozen graves (without trapezoidal enclosure) from the
Baalberge culture and a total 20 burials spread over a length of 200 meters. The individuals
x QLB15/I0559 (feature 21033, 3645-3537 cal BCE, MAMS 22818)
x QLB18/I0560 (feature 21039, 3640-3510 calBCE, Er7856)

27
that were sampled for this study are likely to be burials of commoners as more elaborate
grave architectural elements such as cists are missing.

Middle Neolithic in Spain: La Mina

The site La Mina (Alcubilla de las Peñas, Soria, Spain) includes a classic passage grave with
a corridor longer than five meters oriented south-southeast. The site suffered from a
systematic process of closure and dismantling of its stone structure with all orthostats of the
chamber and the passage being removed. However, the burial chamber, from which the
remains of approximately 20 buried individuals have been recovered, remained intact. At the
time of the burials, the appearance of the tomb was modified by increasing the diameter and
height of the original mound, placing a menhir higher than three meters possibly on top of the
monument, and building a wall of orthostats parallel to the old burial surrounding the entire
perimeter wall. As a result, the original collective grave had been transformed into a
ceremonial center, and given its unique topographic location the mound became a territorial
landmark. A radiocarbon date from a human bone from the ossuary (3780-3700 calBCE, Beta
316132) places the tomb at the beginning of Megalithism in the inner Iberian Peninsula, at the
early 4th millennium.

The samples of La Mina originate from two seasons. The samples of Mina1-10 are from
season 2010 and Mina 11-18 from 2012. In both seasons, samples were taken directly at the
excavation with gloves and facemask and immediately stored under cool conditions. In 2010,
samples were taken by geneticist Sarah Lang and in 2011 by archaeologist Manuel Rojo-
Guerra. The samples that produced 390k data and that are included in our analysis are:
x Mina3/I0405 (3900-3600 BCE)
x Mina5/I0406 (3900-3600 BCE)
x Mina6/I0407 (3900-3600 BCE)
x Mina18/I0408 (3900-3600 BCE)

Late Copper/Early Bronze Age steppe chronology

The Yamnaya horizon was the first cultural complex that spread across all of the Pontic-
Caspian-Ural steppes, beginning about 3300 BCE. It is divided into at least six regional
groups and at least two chronological stages (up to four phases have been suggested), and the
earliest ceramics included at least two distinct types (originating on the lower Don and lower
Volga), both of which were decorated with cord impressions and usually were round-
bottomed. The unifying traits were: a mobile pastoral economy that operated from newly
invented wheeled vehicles using horseback-mounted herders; the introduction of new metal
types (shafthole axes, tanged daggers, and spiral hair-rings or Lockenringe), new metal
techniques, and copper mining on a significant scale, perhaps inspired by the Maikop culture;
the burial of a select few adult males (who made up 75-80% of the graves under kurgans in
the Samara region) under kurgans that usually covered 1 to 3 graves (although in the Kuban
and NW Pontic steppes a wider range of people was buried under kurgans); the burial of
wheels or whole wagons above or beside elite graves; deposition of animal sacrifices (usually
sheep-goat) in about 15% of graves; and erection of grave-associated stone stelae, in the
Ukrainian steppes. A few stratified settlements are found in the Ukrainian steppes west of the
Don but none are known east of the Don in the Volga-Ural region. The Mikhailovka
settlement on the Dnieper River is the standard stratigraphic guide to Yamnaya
chronology20,21. Pre-Yamnaya level 1 began about 3800 BCE, early Yamnaya level 2 began

28
about 3500 BCE, and late Yamnaya level 3 began about 2800 BCE. At Mikhailovka the
Yamnaya phase (early level 2) began about 3500-3300 BCE22, but the majority of early
Yamnaya kurgans began later, closer to 3300 BCE, from the Volga to the Dnieper, so 3300
BCE is a safer date for the general beginning of the Yamnaya era23. EBA Yamnaya evolved
into the MBA Poltavka culture in the Volga-Ural steppes and into the MBA Catacomb culture
in the Dnieper-Don-Caucasus steppes beginning about 2800 BCE. In the Dniester-Bug
steppes, graves of the late Yamnaya style continued as late as 2200 BCE.

Yamnaya in Russia: Ekaterinovka

The site Ekaterinovka is located in the Southern Steppe on the left bank of the Volga 128 km
southwest of the city of Samara. The >45 years old mature male individual included in this
study is
x SVP3/I0231 (Ekaterinovka, grave 1, 2910-2875 calBCE, Beta 392487).

Yamnaya in Russia: Lopatino I

A large cemetery of 39 kurgans was located on a low terrace beside the Sok River, Samara
oblast, Russia (N53°38’24”/E50°39’18”). Eight kurgans were excavated in different years by
various teams. Five were constructed in the Yamnaya period and three were added by the
MBA Poltavka culture. We included three individuals from this site:
x SVP5/I0357 (kurgan 35, central grave 1, 3090-2910 calBCE, Beta 39248)
was from a grave that contained the remains of two Yamnaya individuals, including
an adult woman and a child, an unusual pair, because 75-80% of individuals in
Yamnaya kurgans in the Samara region were adult males.
x SVP38/I0429 (kurgan 31, central grave 1, 3339-2917 calBCE, AA47804)
was from a grave of an adult male 35-45 years old, 175 cm tall, supine, with a
triangular flint projectile point beside him.
x SVP52/I0439 (kurgan 1, central grave 1, 3305-2925 calBCE, Beta 392491)
was from an adult male 25-35 years old, 178.5 cm tall.

Yamnaya in Russia: Ishkinovka I

Ishkinovka (or Ishkinino) is a kurgan cemetery located 30 km north of Novotroitsk on a right-
bank tributary of the Ural River, among the easternmost Yamnaya sites, 570 km southeast of
Lopatino I. The region contains copper ores exploited by miners who were active throughout
the Bronze Age beginning in the EBA, which might explain the Yamnaya occupation of this
eastern region, 250 km east of the Yamnaya mines at the Kargaly copper ore field24. The male
individual included in this study is
x SVP10/I0370 (Kurgan 3, grave 7, 3300-2700 BCE)

Yamnaya in Russia: Luzhki I

This unique cemetery, 1-2 km distant from Lopatino I, might represent the undocumented
majority of the Yamnaya population that was not buried under kurgans. Individual,
x SVP50/I0438 (3021-2635 calBCE, AA47807)
an adult male 25-35 years old was found in a cemetery with no obvious kurgan, surrounded
by the graves of 5 children. He was buried face down with his hands behind his back. He was
crippled by a blow with a heavy blunt instrument to his right hip that crushed his femur just
below the trochanter, with no healing evident, and his skull was gouged pre-mortem in six

29
places with a serrated tool. His wounds and burial position suggest that he was tortured and
executed.

Yamnaya in Russia: Kurmanaevka III

This cemetery of three Yamnaya kurgans was erected on the first terrace overlooking the
floodplain of the Buzuluk River, a left-bank tributary of the upper Samara River, in western
Orenburg oblast, Russia, 170 km southeast of Lopatino I.
x SVP54/I0441 (Kurgan 3, 3010-2622 calBCE, AA47805)
The sample is taken from the femur of an adult woman 35-45 years old whose bones
were heavily stained with red ochre. Her grave is a peripheral grave (burial 2), and is
located above the central grave of a male (burial 1).

Yamnaya in Russia: Lopatino II

A cemetery of three kurgans was located on the first terrace above the floodplain 2 km
northeast of the village of Lopatino on the Sok River in Samara oblast, Russia. One kurgan
(3) was assigned to the EBA Yamnaya culture, the second (kurgan 1) to the MBA Poltavka
culture, and the third (kurgan 2) to the final MBA II Potapovka culture, related culturally to
Sintashta east of the Urals.
x SVP57/I0443 (Grave 1, kurgan 3, 3300-2700 BCE)
contained a male aged 15-17, apparently killed by an unhealed blunt-force trauma to
his right parietal by a hammer-like weapon. Unusually for Yamnaya kurgans, none of
the three graves under kurgan 3 were located under the center of the mound, but
grave 1 was farthest from the center, near the northern margin.

Figure S3.2. Skeleton SVP58 from grave 1 at Kutuluk kurgan cemetery I25,26.

Yamnaya in Russia: Kutuluk

Kutuluk kurgan cemetery I, located 60 km east of the city of Samara, contained:
x SVP58/I0444 (central grave 1, kurgan 4, 3335-2881 calBCE, AA12570)

30
The remains are of male aged 25-35 years (Fig. S3.2), estimated height 176 cm, with no
obvious injury or disease, and buried with the largest metal object found in a Yamnaya grave
anywhere26. The object was a blunt mace 48 cm long, 767 g in weight, cast/annealed and
made of pure copper, like most Yamnaya metal objects.

Late Neolithic
Since most of our samples were collected in present-day Germany, we follow the
German/north-central European chronology for the Late Neolithic. Here, the Late Neolithic
horizon in central Europe is defined as 2725–2200 BCE, even though earliest signs of the
Corded Ware culture can be found around 2800 BCE, whereas remains of Bell Beaker
cultures prevail until 2050 BCE. The Bell Beaker horizon is evident in the archaeological
record from 2500 BCE onward in the southern Mittelelbe-Saale (MES), when it starts to
move into settlement areas previously occupied by Corded Ware people, the latter of which
have cultural affinities to archaeological groups further east such as the Globular Amphora
and the Yamnaya culture in the North Pontic and Russian steppes. The settlement density
increases during later Bell Beaker phases (2300-2050 BCE), when the Corded Ware is
superseded by Bell Beaker elements, often at the same sites. This Late Neolithic Bell Beaker
phenomenon is of interest, since the earliest archaeological evidence has been described in the
Tagus region of Western Iberia around 2800-2700 BCE and it has been traced
archaeologically over large parts of Western Europe but is also found as far as Hungary,
Ireland, and southern Scandinavia, and in smaller enclaves in North Africa27. Here, earlier
Neolithic cultures were overlain by discernable Bell Beaker elements, which became visible
in rich grave goods (including the eponymous bell-shaped ceramic beakers). During the
transition to the Bronze Age, Early Bronze Age cultural elements of the Únětice culture
appear contemporaneously to late elements of the Bell Beaker culture, again sometimes also
at the same site. A number of sites, such as Eulau and Quedlinburg show the presence of both
Corded Ware and Bell Beaker cultures, and later of Únětice culture, and highlight the cultural
diversity and dynamic at the time.

Corded Ware in Germany: Esperstedt

The site Esperstedt forms part of large-scale excavations initiated in 2005 in the context of
major infrastructural roadworks in Saxony-Anhalt, Germany, to build motorway A38.
Individuals from Esperstedt reference site 4 could be unambiguously assigned to the Corded
Ware, both by accompanying pottery and characteristic orientation of the burials28. Males
were usually buried in a right-hand side flexed position with head in the west and facing
south, while females were buried on their left-hand side with their head in the east. We
studied four individuals from this site:
x ESP11/I0104 (feature 6216, 2473-2348 calBCE, MAMS 21487)
x ESP16/I0103 (feature 6236, 2566-2477 calBCE, MAMS 21488)
x ESP22/I0049 (feature 6233.1, 2454-2291 calBCE, MAMS 21489)
x ESP 26/I0106 (feature 6216, 2454-2291 calBCE, MAMS 21490)

31
Figure S3.3. Feature 10049 from Rothenschirmbach, Germany, highlighting the
complex grave architecture at the site (Photo: H. Arnold, LDA Sachsen-Anhalt,
Germany).

Bell Beaker in Germany: Rothenschirmbach

The site Rothenschirmbach in the district Merseburg-Querfurt is also located on the loess
plain of the ‘Querfurter Platte’. The site features a densely occupied cemetery from the Bell
Beaker period including 14 burials. Females of the Bell Beaker culture were usually buried on
their right side with the head in the south and facing east, while males were buried with their
heads in the north, facing west. Key features of this cemetery are a high proportion of
children's burials and the large number of rich and embellished graves with partially complex
architecture (cists, menhirs, e.g. Fig. S3.3), and the rare finding of gold in a man’s grave (the
oldest gold find in central Germany)29. We sampled three individuals from this site:
Individual
x ROT3/I0060 (feature 10011, 2294-2206 calBCE, MAMS 22819)
is a 4-6 year old child (infans I) buried in a stone cist.
x ROT4/I0111 (feature 10142, 2414-2333 calBCE, Er8712)
featured remnants of a strong wooden cist covered by a 400kg menhir stone of 1.5 x
0.55m in size.
x ROT6/I0108 (feature 10044, 2497-2436 calBCE, Er8710)
is an adult buried in a open stone cist.

Bell Beaker and Late Neolithic in Germany: Benzingerode-Heimburg

x BZH6/I0059 (grave 2, feature 1287, 2286-2153 calBCE, MAMS 21486)
was attributed to the Bell Beaker period by excavator Tanja Autze as the pit overlaps
with a neighbouring grave harboring BBC features. The orientation matches BBC
customs, and the radiocarbon date supports assignment to this period.

32
 x BZH4/I0058 (grave 7, feature 4607, 2283-2146 calBCE, MAMS 21491)
had no grave goods, and the skeleton was oriented SE-NW, rightside-flexed position,
facing north.
x BZH12/I0171 (grave 3, feature 2350, 2204-2136 calBCE, KIA27952)
was buried in SE-NW orientation, in a right-sided flexed position, facing west. While
the burial customs do not match standard Bell Beaker or the Corded Ware rites, the
radiocarbon date supports a Late Neolithic assignment.
Interestingly, radiocarbon dates of all three individuals fall in the overlapping period of the
late Corded Ware, Bell Beaker and early Únětice period in this region30.

Bell Beaker and Late Neolithic in Germany: Quedlinburg VII and XIIa
Reference site XII-West is located northeast of Quedlinburg, Saxony-Anhalt, and forms part
of a small graveyard with mixed grave goods from Late Neolithic cultures31. The grave
x QUEXII6/I0112 (feature 6256, 2340-2190 calBCE, Er7038)
was archaeologically assigned to Bell Beaker based on the orientation, however grave
goods include both a Bell Beaker 'Füsschenschale' and a 'Corded Ware' cup with
ornaments characteristic for the single grave culture. The individual was an outlier
from other Bell Beaker samples in PCA, and so we marked it as such for genome-
wide analysis.
x QUEXII4/I0113 (feature 6255.1, 2290-2130 cal BCE (Er7283)
was also a female whose single grave good, a small cup, was uninformative. The
orientation of her body is typical for the Bell beaker period.

Another group of six graves was discovered at Quedlinburg reference site VII and attributed
to the Bell beaker culture based on form and orientation of the burials 32. We included a >50
year old male individual buried in an extreme flexed position.
x QLB28/I0806 (feature 19617, 2296-2206 calBCE, MAMS 22820)

Other Late Neolithic samples in Germany: Karsdorf

We also included a sample from an unusual burial at the Karsdorf site in Germany
x KAR22/I0550 (feature 00191, date unknown)
The original archaeological attribution to the Baalberge culture is ambiguous and while the
grave appears prehistoric, the determination was entirely based its position inside the eastern
gate of a burial complex from the Baalberge culture. The mitochondrial genome of KAR22
could be determined as T1a1, matching previous HVS-I based results1. By analyzing genomic
data KAR22 clusters with Late Neolithic Corded Ware individuals (Figure 2a), while the
admixture proportions in Figure 3 reveal a complete absence of an Early Neolithic component
(even less than other Corded Ware individuals). Given the large proportion of Yamnaya-
related ancestry, a broader group assignment to the Late Neolithic appears likely.
Radiocarbon dating results are pending.

Other Late Neolithic samples in Germany: Alberstedt

The site Alberstedt in Merseburg-Querfurt, Saxony-Anhalt, on the loess-bearing Querfurter
Platte, is located on a hilltop. This promontory site was used by the Corded Ware and Bell
Beaker people as a burial ground. The single grave of individual
x ALB3/I0118 (feature 7144.2, 2459-2345 calBCE, MAMS 21492)

33
was uncovered within a strip of 20m width in preparation for major roadworks. The grave
was initially ascribed to the Bell Beaker culture but is rather an unusual burial complex
strewn with cattle bones as well as a few sherds of Corded Ware-like pottery in the back
filling28. The radiocarbon date falls in line with both the Bell Beaker and Corded Ware
occupation phases of this region. Given the ambiguous archaeological classification we have
decided to use the location and date to classify this sample (Late Neolithic Alberstedt). The
intermediate position of this sample on the PCA plot (Figure 2a) between unambiguously
assigned Corded Ware (Esperstedt) and Bell Beaker (Rothenschirmbach) individuals, and
>50% Yamnaya ancestry are consistent with an individual who has mixed Corded Ware and
Bell Beaker ancestry.

Early Bronze Age

Únětice in Germany: Esperstedt
The reference site 4e at Esperstedt in Merseburg-Querfurt, Saxony-Anhalt/Germany includes
a small graveyard of 10 burials that could be assigned to the Early Bronze Age Únětice
culture. All individuals are buried on the their right-hand side in flexed position, which is
typical for this time period, with the head in the south and facing west. The graves at
Esperstedt are slightly tilted towards SE, facing NE. Individuals
x ESP2/I0114 (feature 3340.1, 2131-1979 calBCE, MAMS 21493) and
x ESP29/I0117 (feature 3332/3333, 2199-2064 calBCE, MAMS 21496)
form a small group and appear to be genetically closely related.
x ESP4/I0116 (feature 3322/3323, 2118-1961 calBCE, MAMS 21495)
was also buried as part of a small group of three graves nearby, while individual
x ESP3/I0115 (feature 1559.1, 1931-1780 calBCE, MAMS 21494),
a female adult from a more recent phase at reference site 4b, formed part of a double
inhumation with a ~10 year-old child at approximately 100 meters distance from the older
graveyard33.

Únětice in Germany: Quedlinburg VIII

The site in Quedlinburg, located in the northern Harz region, in the foreland of the Harz
mountains, harbors several graves from the Early Bronze Age. Three adjacent grave groups
were found next to the stream Sülze and could be assigned to the Únětice culture. From these,
we sampled individual
x QUEVIII6/I0164 (feature 3580, 2012-1919 calBCE, MAMS 21497)
whose otherwise ordinary grave was sealed with a stone cover.

Únětice in Germany: Eulau

The site Eulau, Burgenlandkreis, is located in the valley of the Saale river. The promontory
that forms "Eulau" is a loess above-gravel formation and was intensively populated during the
Early Bronze Age. Individual
x EUL41A/I0803 (feature 882, 2115-1966 calBCE, MAMS 22822)
was buried on the eastern edge of the promontory in a small burial ground of less than 10
burials attributed to the the Early Bronze Age. Individual
x EUL57B/I0804 (feature 1911.5, 2131-1982 calBCE, MAMS 22821)
was buried 200 meters distance away on the western edge of the promontory in another burial
ground with approximately 20 features. Feature 1911 is an unusual grave context, including
skeletal remains from three juvenile males in the bottom, and a cist in which a child and a

34
neonate had been laid down. The samples taken here are from one of the three male
individuals. The graves have been dated to 2200-1550 BCE based on archaeological context.

Bronze Age in Germany: Halberstadt-Sonntagsfeld, please see also N.B. below

x I0047/HAL16 (grave 19, feature 613.1, 2022-1937 calBCE, MAMS 21481)
features a skeleton buried in right-handed flexed position, with head in the west, facing south.
The grave contained one bone artifact, but no pottery. This individual was originally
attributed to the LBK based on the presence of an LBK settlement with associated burials
nearby15, but direct radiocarbon dating newly generated for this study revealed a younger date
overlapping with the Bronze Age Únětice culture.
x I0099/HAL36C (grave 40, feature 1114, 1113-1021 calBCE, MAMS 21484)
was buried in right-handed flexed position, head SSW, facing SE. Two decorated LBK pots
and two undecorated globular pots were found above the grave but it was not clear whether
they were part of the burial or the back filling. Thus, the skeleton was also originally thought
to be part of the LBK burial series found at the same site, but subsequent radiocarbon dating
performed for this study indicated a much younger date, placing this individual within the
Late Bronze Age Urnfield culture of the Mittelelbe-Saale region.

Nota bene: To our knowledge, these results are the first published case in which ancient DNA
data have been used to identify outlier individuals: individuals who are genetically distinct
from others at the same site that have been classified as being from the same archaeological
culture (in this case, the Early Neolithic LBK). The fact that this genetic outlier status is
consistent with the recent radiocarbon dates and the position of the individuals at the
periphery of the site – not directly associated with one of the grave groups that accompany
each of the LBK houses (Figure S3.4) – indicates that the genetic analysis is likely to be
accurately identifying individuals that are outliers. These results suggest that in the future,
genetic analysis may be useful for classifying burials when grave goods and other context are
missing or unclear.

Figure S3.4. Map of the Halberstadt settlement with groups of graves (red rectangles)
associated to LBK houses (purple). The graves dating to the Bronze Age (features 1114 and
0613) are clearly marked as outliers with 20-30m distances to the LBK graves (red
rectangles) (source: LDA Sachsen-Anhalt, Germany).

35
References

1 Brandt, G. et al. Ancient DNA reveals key stages in the formation of central
European mitochondrial genetic diversity. Science 342, 257-261, (2013).
2 Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations
for present-day Europeans. Nature 513, 409-413, (2014).
3 Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry
of Native Americans. Nature 505, 87-91, (2014).
4 Der Sarkissian, C. et al. Mitochondrial Genome Sequencing in Mesolithic
North East Europe Unearths a New Sub-Clade within the Broadly Distributed
Human Haplogroup C1. PLoS ONE 9, e87612, (2014).
5 Der Sarkissian, C. et al. Ancient DNA Reveals Prehistoric Gene-Flow From
Siberia in the Complex Human Population History of North East Europe.
PLoS Genetics 9, e1003296 (2013).
6 Vasiliev, I. B. & Ovchinnikova, N. V. in Samara: Integratsiya (eds A.A.
Vybornov, Y.I. Kolev, & A.E. Mamanova) pp. 216-277 (2000).
7 Bánffy, E. The Late Starcevo and the Earliest Linear Pottery Groups in
Western Transdanubia. Documenta Praehistorica 27, 173-185 (2000).
8 Price, T. D. Europe's first farmers. (Cambridge University Press, 2000).
9 Whittle, A. W. R. & Cummings, V. Going over: the mesolithic-neolithic
transition in North-West Europe. (Oxford University Press, 2007).
10 Bánffy, E., Marton, T. & Osztás, A. in Neolithization of the Carpathian Basin:
Northernmost distribution of the Starčevo/Körös Culture (eds J. K. Kozlowski
& P. Raczky) pp. 37-51 (2010).
11 Nagy, E. Az M6 – M60 gyorsforgalmi út Baranya megyei szakasza. Magyar
Múzeumok 2007, 21-24 (2007).
12 Schwarz, R. in Kataloge zur Dauerausstellung im Landesmuseum für
Vorgeschichte Halle, 2. Neolithikum (ed H. Meller) (in press).
13 Bocquet-Appel, J. P., Naji, S., Vander Linden, M. & Kozlowski, J. K.
Detection of diffusion and contact zones of early farming in Europe from the
space-time distribution of 14C dates. Journal of Archaeological Science 36,
807-820 (2009).
14 Shennan, S. & Edinborough, K. Prehistoric population history: from the late
glacial to the late neolithic in central and northern Europe. Journal of
Archaeological Science 34, 1339-1345 (2007).
15 Fritsch, B., Claßen, E., Müller, U. & Dresely, V. Die linienbandkeramischen
Gräberfelder von Derenburg „Meerenstieg II“ und Halberstadt
„Sonntagsfeld“, Lkr. Harz. Jahresschrift für mitteldeutsche Vorgeschichte 92,
25-229 (2011).
16 Meyer, C., Kürbis, O. & Alt, K. W. Das Massengrab von Wiederstedt, Ldkr.
Mansfelder Land. Auswertung und Gedanken zur Deutung im Kontext der
Linienbandkeramik. Jahresschrift für mitteldeutsche Vorgeschichte 88, 31-66
(2004).
17 Rojo Guerra, M. et al. Pastores trashumantes del Neolítico Antiguo en un
entorno de alta montaña: secuencia crono-cultural de la Cova de Els Trocs
(San Feliú de Veri, Huesca). BSAA arqueología LXXIX, 9-55 (2013).
18 Rojo Guerra, M. A. in Iberia. Protohistory of the far west of Europe: From
Neolithic to Roman conquest (ed M. Almagro Gorbea) pp. 43-69 (Universidad
de Burgos, Fundación Atapuerca, 2014).

36
19 Faron-Bartels, R., Bartels, R. & Döhle, H.-J. in Archäologie auf der
Überholspur. Ausgrabungen an der A38 Vol. 5 Archäologie in Sachsen-
Anhalt (eds V. Dresely & H. Meller) pp. 21-25 (Landesamt für Archäologie
Sachsen-Anhalt, Halle (Saale), 2006).
20 Lagodovskaya, E. F., Shaposhnikova, O. G. & Makarevich, M. L. Osnovnye
itogi issledovaniya Mikhailovskogo poseleniya. Kratkie Soobshcheniya
Institut Arkheologii 9, 21-28 (1959).
21 Lagodovskaya, E. F., Shaposhnikova, O. G. & Makarevich, M. L.
Mikhailivkske Poselennya. (1962).
22 Kotova, N. & Spitsyna, L. A. Radiocarbon chronology of the middle layer of
the Mikhailivka settlement. Baltic-Pontic Studies 12, 121-131 (2003).
23 Telegin, D. Y., Pustalov, S. Z. & Kovalyukh, N. N. Relative and absolute
chronology of Yamnaya and Catacomb monuments: the issue of co-existence.
Baltic-Pontic Studies 12, 132-184 (2003).
24 Chernykh, E. N. in Metallurgy in Ancient Eastern Eurasia from the Urals to
the Yellow River, Chinese Studies 31 (ed K.M. Linduff) pp. 223-279 (Edwin
Mellen Press, 2004).
25 Kuznetsov, P. F. Epokhi rannego I srednogo bronzovykh vekov Samarskogo
Povolzh’ya (Early and Middle Bronze Age in the Samara region of the Volga.
in 40 Years of Middle Volga archaeological expeditions (40 Let
Srednevolzhskoi arkheologicheskoi Ekspeditsii). Kraevedcheskie zapiski
(Historical Notes) 40, pp. 34-49 (2010).
26 Kuznetsov, P. An Indo-Iranian symbol of power in the earliest steppe
Kurgans. Journal of Indo-European Studies 33, 325-338 (2005).
27 Cunliffe, B. Facing the Ocean: The Atlantic and its People 8000BC - 1500
AD. (Cambridge University Press, 2001).
28 Leinthaler, B., Bogen, C. & Döhle, H.-J. in Archäologie auf der Überholspur.
Ausgrabungen an der A38 Vol. 5 Archäologie in Sachsen-Anhalt, Sonderband
(eds V. Dresely & H. Meller) pp. 59-82 (Landesamt für Archäologie Sachsen-
Anhalt, Halle (Saale), 2006).
29 Müller, U. in Archäologie auf der Überholspur. Ausgrabungen an der A38
Vol. 5 Archäologie in Sachsen-Anhalt, Sonderband (eds V. Dresely & H.
Meller) pp. 98-107 (Landesamt für Archäologie Sachsen-Anhalt, Halle
(Saale), 2006).
30 Autze, T. in Quer-Schnitt. Ausgrabungen an der B 6n. Benzingerode-
Heimburg Vol. 2 Archäologie in Sachsen-Anhalt (eds V. Dresely & H. Meller)
pp. 39-51 (Landesamt für Archäologie Sachsen-Anhalt, Halle (Saale), 2005).
31 Peters, E. in Archäologie XXL. Archäologie an der B 6n im Landkreis
Quedlinburg Vol. 4 (eds V. Dresely & H. Meller) pp. 112-115 (Landesamt für
Archäologie Sachsen-Anhalt, Halle (Saale), 2006).
32 Dalidowski, X. in Archäologie XXL. Archäologie an der B 6n im Landkreis
Quedlinburg Vol. 4 (eds V. Dresely & H. Meller) pp. 116-120 (Landesamt für
Archäologie Sachsen-Anhalt, Halle (Saale), 2006).
33 Faron-Bartels, R. & Bartels, R. in Archäologie auf der Überholspur.
Ausgrabungen an der A38 Vol. 5 Archäologie in Sachsen-Anhalt (eds V.
Dresely & H. Meller) pp. 108-123 (Landesamt für Archäologie Sachsen-
Anhalt, Halle (Saale), 2006).

37
Supplementary Information 4
Sex determination and Y chromosome analysis

Iosif Lazaridis*, Swapan Mallick, Nick Patterson, and David Reich

* To whom correspondence should be addressed ([email protected])

Sex determination
We determined the sex of the 69 individuals newly analyzed in this study by examining the number of
reads overlapping targets on 390k capture reagent on the X chromosome (n=1,829) and Y
chromosome (n=2,258) SNPs targeted by the 390k capture reagent.

We applied the read mapping and filtering procedure described in the Online Methods section. Table
S4.1 gives the number of non-duplicated reads mapping to the X and Y chromosome targets for each
individual.

The ratio of Y/(Y+X) reads1 (Fig. S4.1) shows a bimodal distribution, with 35 individuals having a
ratio that is 0.0011 ± 0.0012 (1 standard deviation) who we interpret as female, and 34 individuals
having a ratio of 0.508 ± 0.025 who we interpret as male. The remainder of this note focuses on Y
chromosome haplogroups for the 34 individuals determined to be male.

Fig. S4.1: Ratio of Y/(Y+X) alignments. Samples are sorted on the value of this ratio, demonstrating
a discontinuity between females (where the ratio is ~0) and males (where it is ~0.5).

0.7

0.6

0.5
Ratio of Y/(X+Y) reads

0.4

0.3

0.2

0.1

0
0 10 20 30 40 50 60 70
69 individuals

Y chromosome haplogroup determination.

The 390k capture reagent targeted all SNPs present in the Y-DNA SNP index of the International
Society of Genetic Genealogy (ISOGG) version 8.22 as of April 22, 2013
(http://isogg.org/tree/ISOGG_YDNA_SNP_Index.html). At each SNP, we represent the individual
using the majority allele (breaking ties randomly) for all non-duplicated reads overlapping the SNP,
requiring MAPQ≥30, base quality≥30, and trimming 2 bases at the ends of reads.

38
Table S4.1: Number of reads aligning to the X and Y chromosome targets

ID Xreads Yreads Sex

I0011 1091 3 XX
I0012 1151 1225 XY
I0013 898 977 XY
I0014 1581 3 XX
I0015 1173 1252 XY
I0016 1057 1145 XY
I0017 1621 1725 XY
I0022 831 0 XX
I0025 1606 3 XX
I0026 1662 1 XX
I0046 2403 2 XX
I0047 1826 1 XX
I0048 398 401 XY
I0049 1083 0 XX
I0054 3584 5 XX
I0056 410 374 XY
I0057 307 1 XX
I0058 1384 1 XX
I0059 1506 1 XX
I0060 257 0 XX
I0061 2477 2490 XY
I0099 1620 1673 XY
I0100 1748 1 XX
I0103 5231 4 XX
I0104 3069 3081 XY
I0106 354 0 XX
I0108 1738 0 XX
I0111 1071 2 XX
I0112 1737 2 XX
I0113 1050 2 XX
I0114 752 773 XY
I0115 665 2 XX
I0116 1301 1321 XY
I0117 1933 1 XX
I0118 3449 3 XX
I0124 688 680 XY
I0164 1722 1 XX
I0171 347 0 XX
I0172 5264 5621 XY
I0174 282 282 XY
I0176 143 0 XX
I0231 4596 4752 XY
I0357 1011 6 XX
I0370 627 670 XY
I0405 353 359 XY
I0406 1470 1472 XY
I0407 1054 1 XX
I0408 4245 5 XX
I0409 831 1 XX
I0410 1172 1176 XY
I0411 62 69 XY
I0412 4164 4272 XY
I0413 1483 1 XX
I0429 776 763 XY
I0438 701 758 XY
I0439 293 332 XY
I0441 276 0 XX
I0443 2869 2900 XY
I0444 590 572 XY
I0550 277 0 XX
I0559 214 168 XY
I0560 692 1 XX
I0659 642 614 XY
I0795 111 121 XY
I0803 714 0 XX
I0804 50 72 XY
I0806 243 284 XY
I0807 105 84 XY
I0821 155 175 XY

We performed Y-haplogroup determination by examining the state of SNPs present in ISOGG version
9.129 (accessed Dec 08, 2014); we used this later version—even though it includes many more SNPs
than were present in version 8.22 used during the design of the 390k capture reagent—in order to

39
obtain up-to-date Y-haplogroup nomenclature. We determined Y chromosome haplogroups by
identifying the most derived Y chromosome SNP in each individual.

.Table S4.2: Y-Haplogroup assignments for 34 ancient European males.

Mean coverage
ID on SNP targets Haplogroup Population
(genome-wide)
I0559 0.20 R*? Baalberge_MN
I0807 0.10 F* Baalberge_MN
I0806 0.31 R1b1a2a1a2 Bell_Beaker_LN
I0104 6.21 R1a1a1 Corded_Ware_LN
I0172 58.03 I2a1b1a Esperstedt_MN
I0099 8.90 R1a1a1b1a2 Halberstadt_LBA
I0061 8.01 R1a1 Karelia_HG
I0048 0.45 G2a2a LBK_EN
I0056 0.53 G2a2a LBK_EN
I0659 0.82 G2a2a1 LBK_EN
I0795 0.15 T1a LBK_EN
I0821 0.18 G2a2a1 LBK_EN
I0012 2.76 I2c2 Motala_HG
I0013 1.89 I2a1b Motala_HG
I0015 3.41 I2a1 Motala_HG
I0016 2.35 I2a1 Motala_HG
I0017 8.79 I2a1b Motala_HG
I0124 0.96 R1b1 Samara_HG
I0410 3.29 R1b1 Spain_EN
I0412 28.17 I2a1b1 Spain_EN
I0411 0.11 F* Spain_EN_relative_of_I0410
I0405 0.56 I2a1a1/H2? Spain_MN
I0406 9.28 I2a2a1 Spain_MN
I0174 0.36 H2 Starcevo_EN
I0116 3.86 I2c2 Unetice_EBA
I0804 0.07 I2 Unetice_EBA
I0114 1.07 I2a2 Unetice_EBA_relative_of_I0117
I0231 10.02 R1b1a2a2 Yamnaya
I0370 0.97 R1b1a2a2 Yamnaya
I0429 1.06 R1b1a2a2 Yamnaya
I0438 1.00 R1b1a2a2 Yamnaya
I0439 0.33 R1b1a Yamnaya
I0443 9.10 R1b1a2a Yamnaya
I0444 0.79 R1b1a2a2 Yamnaya

We discuss the specific results for individual samples below.

I0559 (Baalberge_MN)
An assignment to haplogroup R is possible based on P224:17285993C→T. This may represent
ancient DNA damage, so the assignment should be viewed with caution. The sample can be assigned

40
to the upstream haplogroup P1 (the newly defined parent node of haplogroups Q and R) based on
P230:17470112G→A. Downstream haplogroups that could be excluded were R1a1a
(M515:14054623T→A), and R1b1a2a1a (L151:16492547C→T), so it is R*(xR1a1a, R1b1a2a1a).

I0807 (Baalberge_MN)
The assignment to haplogroup F* is based on a single mutation P316:16839641A→T. We could
exclude haplogroups G (Page94:2846401C→T, PF2956:14993358A→G, PF3134:15275200C→G),
H1a2a (Z14683:17431284A→C), H1b1 (Z14006:7786084G→T), H1b2a (Z14385:22458616G→C),
H3a1 (Z13118:17493503T→C, Z13181:23121729G→A), H3a2 (Z12794:22191274A→C), I1a3
(S243:14401486C→T), I2a1b (M423:19096091G→A), J1a2b3a (L818:19136821A→G), J2a1h2
(L25:19136822T→C), J2b2a1a1 (Z631:2819161G→A), L (M11:21730647A→G), NO
(P194:16202980C→G) and P1 (P237:8334875A→G, P240:14598808T→C, P282:18028661A→G).

I0806 (Bell_Beaker_LN)
The individual was assigned to haplogroup R1b1a2a1a2 based on mutation P312:22157311C→A.
Two Bell Beaker individuals from Kromsdorf, Germany were previously determined2 to belong to
haplogroup R1b.

The individual also has upstream mutations for R1 (P236:17782178C→G), R1b1

(L278:18914441C→T), R1b1a2 (F1794:14522828G→A), and R1b1a2a1 (L51:8502236G→A). Its
haplotype is ancestral for R1b1a2a1a2a1a1a (S1217:7193830C→G, Z262:16320197C→T),
R1b1a2a1a2c1a (DF49:22735599G→A), R1b1a2a1a2c1a1 (DF23:17774409G→A), R1b1a2a1a2c1f1
(L554:15022777A→G), R1b1a2a1a2c1f2 (S868:19033817T→C), R1b1a2a1a2c1i
(CTS6581:16992602T→C) and R1b1a2a1a2c1l1a1 (CTS2457.2:14313081C→T).

I0104 (Corded_Ware_LN)
This individual was assigned to haplogroup R1a1a1 based on mutation M417:8533735G→A, and also
had six upstream mutations placing it in haplogroup R1a1a (M515:14054623T→A,
M198:15030752C→T, L168:16202177A→G, M512:16315153C→T, M514:19375294C→T, and
L449:22966756C→T) and four mutations placing it in haplogroup R1a1 (L145:14138745C→A,
L62:17891241A→G and L63:18162834T→C, L146:23473201T→A).

We could further exclude lineages R1a1a1a (CTS7083:17275047T→G) and R1a1a1b

(S441:7683058G→A, S224:8245045C→T), so I0104 could be more precisely described as
R1a1a1*(x R1a1a1a, R1a1a1b).

Three related individuals belonging to haplogroup R1a were previously described from Corded Ware
individuals from Eulau, Germany3. The R-M417 haplogroup has an estimated TMRCA of ~5,800
years ago4, which indicates that I0104 lived about 1.5ky after the foundation of this lineage from
which the vast majority of modern R1a-related chromosomes from across Eurasia are descended.
More than 96% of modern European R-M417 Y-chromosomes belong4 to lineage R1a1a1b1a-Z282
which can be excluded for individual I0104, both indirectly, based on its ancestral state for the
upstream mutation defining haplogroup R1a1a1b, and directly, as the individual was ancestral for the
Z282 polymorphism itself (15588401T→C). Thus, I0104 was related to the modern European set of
R1a Y-chromosomes but did not belong to the more dominant group within this set.

41
I0172 (Esperstedt_MN)
This individual was assigned to haplogroup I2a1b1a based on mutation L1498:18668472C→T, and
could also be assigned to upstream haplogroups I2a1b1 (L161.1:22513718C→T) and I2a1b
(CTS1293:7317227G→A, CTS1802:14074218A→T, L178:15574052G→A,
CTS8239:17893806A→G, M423:19096091G→A, CTS11030:22905944G→C).

Thus, I0172 belonged to a more derived clade than the ~8,000-year old Loschbour male5 from
Luxembourg, and may represent a hunter-gatherer Y-chromosomal lineage that was incorporated in
the population of Middle Neolithic farmers from Germany. Haplogroup I2a lineages were also
detected in Swedish hunter-gatherers5,6 from 7-5 thousand years ago, an early Hungarian individual
(~5,700 years cal BC) with a “hunter-gatherer” autosomal makeup that belonged to an early farmer
community7, as well as later ~5,000 year old individuals from Treilles, France8, while haplogroup I
lineages were observed in two early Neolithic farmers from Hungary belonging to the early Neolithic
Trans-Danubian Linear Pottery (LBKT) and Starcevo cultures9. It thus appears that there was gene
flow from male hunter-gatherers into the Early and Middle Neolithic farmers across Europe.

I0099 (Halberstadt_LBA)
This single individual from the Late Bronze Age could be assigned to haplogroup R1a1a1b1a2 based
on S204:16474793G→A, and to upstream R1a1a1b1a (S198:15588401T→C), and R1a1a1b1
(PF6217:21976303T→A), and R1a1a1b (S224:8245045C→T, S441:7683058G→A). It is thus more
derived than the earlier Corded Ware I0104 individual and belongs firmly within the present-day
European variation of R1a Y-chromosomes.

I0061 (Karelia_HG)
In contrast to I0104 and I0099, the hunter-gatherer from Karelia could only be assigned to haplogroup
R1a1 (M459:6906074A→G, Page65.2:2657176C→T) and the upstream haplogroup R1a
(L145:14138745C→A, L62:17891241A→G, L63:18162834T→C, L146:23473201T→A). It was
ancestral for the downstream clade R1a1a (M515:14054623T→A, M198:15030752C→T,
M512:16315153C→T, M514:19375294C→T, L449:22966756C→T). Thus, it can be designated as
belonging to haplogroup R1a1*(xR1a1a) and it occupied a basal position to the vast majority of
modern Eurasian R1a-related Y-chromosomes4, although more basal (R1a-M420*) Y-chromosomes
have been detected in Iran and eastern Turkey4. Overall, our detection of haplogroup R1a1 in a
northwest Russian hunter-gatherer establishes the early presence of this lineage in eastern Europe, and
is consistent with a later migration from eastern Europe into central Europe which contributed such
haplogroups to the Corded Ware population.

I0048 (LBK_EN)
This individual could be assigned to haplogroup G2a2a (PF3185:22894488C→T) and to upstream
haplogroup G2a (L31:14028148C→A). It was ancestral for haplogroup G2a2a1b
(L91:21645555G→C), so it could be designated G2a2a*(xG2a2a1b).

Haplogroup G2a has been found in early Neolithic farmers from Germany10, Hungary9, the Tyrolean
Iceman11, ~5,000 year old farmers from France8, and ~7,000 year old ones from Spain12. It it thus a
link between Early Neolithic farmers of central Europe and the Mediterranean, as its presence13 in
modern Sardinians, a population with known links to the early European farmers5,11,14,15 also suggests.

42
I0056 (LBK_EN)
This individual also belonged to haplogroup G2a2a (PF3147:7738069G→A, PF3175:18962113C→T,
PF3181:21808944C→A), but not to G2a2a1a (M286:22741799G→A) or G2a2a1b1
(FGC5668:22467833A→G), so it could be designated G2a2a*(xG2a2a1a, G2a2a1b1).

I0659 (LBK_EN)
This individual also belonged to haplogroup G2a2a1 (PF3170:18090604G→A), and to upstream
haplogroup G2a2a (PF3151:9785736A→G, PF3161:15702713A→C, PF3175:18962113C→T,
PF3184:22576860C→T, PF3185:22894488C→T), but not to downstream haplogroup G2a2a1
(PF3177:21327198C→T). Thus, this individual carried the derived state for one of the SNPs defining
haplogroup G2a2a1 (PF3170), and the ancestral state for another (PF3177), suggesting that the first of
these mutations occurred before the second.

I0795 (LBK_EN)
This individual belonged to haplogroup T1a (PF5604:7890461C→T, M70:21893881A→C). This is
the first instance of this haplogroup in an ancient individual that we are aware of and strengthens the
case for the early Neolithic origin of this lineage in modern Europeans16, rather than a more recent
introduction from the Near East where it is more abundant today.

I0821 (LBK_EN)
This individual belonged to haplogroup G2a2a1 (PF3155:14006343T→C) and also to upstream
haplogroup G2a2a (PF3166:16735582T→G), and G2a2 (CTS4367:15615340C→G). We could
exclude downstream haplogroup G2a2a1b (L91:21645555G→C), so it could be designated
G2a2a1*(xG2a2a1b).

I0012 (Motala_HG)
This is the Motala2 individual whose shotgun data was previously analyzed5. It belonged to
haplogroup I2c2 (PF3827:22444389T→A), with the upstream haplogroup I2c (L597:18887888T→A)
also supported. Our higher coverage capture data defines its haplogroup more precisely than the I
haplogroup previously reported5.

I0013 (Motala_HG)
This is the Motala3 individual whose shotgun data was previous analyzed5. It belonged to haplogroup
I2a1b (M423:19096091G→A), consistent with the previous analysis. Haplogroup I2a1b1
(L161.1:22513718C→T) could be excluded, and it could thus be designated I2a1b*(xI2a1b1). The
analysis of the shotgun data5 could reject M359.2 (previously listed as I2a1b1, but currently
designated as under “Investigation” by ISOGG), and I2a1b2-L621 (previously listed as I2a1b3).

I0015 (Motala_HG)
This is the Motala6 individual whose shotgun data was previous analyzed5. It belonged to haplogroup
I2a1 (P37.2:14491684T→C) and the upstream haplogroup I2a (L460:7879415A→C). Its
phylogenetic position could not be determined in the previous analysis of the shotgun data. We also
find that it was ancestral for I2a1a (L159.1:15810964T→G, M26:21865821G→A and
L158:23496560G→A), I2a1b (M423:19096091G→A), I2a1c (L233:14487362G→A), and I2a1e
(L1294:2887401T→C), so it could be designated I2a1*(xI2a1a, I2a1b, I2a1c, I2a1e).

I0016 (Motala_HG)
This is the Motala9 individual whose shotgun data was previous analyzed5. It belonged to haplogroup
I2a1a1a (L672:22228628T→A) and the upstream haplogroups I2a1 (P37.2:14491684T→C) and I2a
(L460:7879415A→C), so it could be better resolved than in the analysis of the shotgun data which

43
could only designated it as I*(xI1). It was also ancestral for I2a1a1a1a1b (Z118:20834727A→G), and
could be designated I2a1a1a*(xI2a1a1a1a1b).

I0017 (Motala_HG)
This is the Motala12 individual whose shotgun data was previous analyzed5. It could be assigned to
haplogroup I2a1b2a1 (L147.2:6753258T→C) and also the upstream haplogroup I2a1b
(L178:15574052G→A, M423:19096091G→A). However, in the shotgun data haplogroup I2a1b2-
L621 (previously known as I2a1b3) could be excluded based on mutation L621:1876008G→A, which
is inconsistent with this individual carrying the derived state for haplogroup I2a1b2a1. We do not
have a call for L621 in the capture data, so we are certain only of this individual’s assignment to
haplogroup I2a1b.

To summarize the data from the five Motala males, we could assign 4 of 5 males to haplogroup I2a1
and its subclades and one (Motala2/I0012) to haplogroup I2c2.

I0124 (Samara_HG)
The hunter-gatherer from Samara belonged to haplogroup R1b1 (L278:18914441C→T), with
upstream haplogroup R1b (M343:2887824C→A) also supported. However, he was ancestral for both
the downstream haplogroup R1b1a1 (M478:23444054T→C) and R1b1a2 (M269:22739367T→C) and
could be designated as R1b1*(xR1b1a1, R1b1a2). Thus, this individual was basal to most west
Eurasian R1b individuals which belong to the R-M269 lineage as well as to the related R-M73/M478
lineage that has a predominantly non-European distribution17. The occurrence of chromosomes basal
to the most prevalent lineages within haplogroups R1a and R1b in eastern European hunter-gatherers,
together with the finding of basal haplogroup R* in the ~24,000-year old Mal’ta (MA1) boy18
suggests the possibility that some of the differentiation of lineages within haplogroup R occurred in
north Eurasia, although we note that we do not have ancient DNA data from more southern regions of
Eurasia. Irrespective of the more ancient origins of this group of lineages, the occurrence of basal
forms of R1a and R1b in eastern European hunter-gatherers provide a geographically plausible source
for these lineages in later Europeans where both lineages are prevalent4,17,19.

I0410 (Spain_EN)
We determined that this individual belonged to haplogroup R1b1 (M415:9170545C→A), with
upstream haplogroup R1b (M343:2887824C→A) also supported. However, the individual was
ancestral for R1b1a1 (M478:23444054T→C), R1b1a2 (PF6399:2668456C→T, L265:8149348A→G,
L150.1:10008791C→T and M269:22739367T→C), R1b1c2 (V35:6812012T→A), and R1b1c3
(V69:18099054C→T), and could thus be designated R1b1*(xR1b1a1, R1b1a2, R1b1c2, R1b1c3).

The occurrence of a basal form of haplogroup R1b1 in both western Europe and R1b1a in eastern
Europe (I0124 hunter-gatherer from Samara) complicates the interpretation of the origin of this
lineage. We are not aware of any other western European R1b lineages reported in the literature
before the Bell Beaker period (ref. 2 and this study). It is possible that either (i) the Early Neolithic
Spanish individual was a descendant of a Neolithic migrant from the Near East that introduced this
lineage to western Europe, or (ii) there was a very sparse distribution of haplogroup R1b in European
hunter-gatherers and early farmers, so the lack of its detection in the published literature may reflect
its occurrence at very low frequency.

The occurrence of a basal form of R1b1 in western Europe logically raises the possibility that present-
day western Europeans (who belong predominantly to haplogroup R1b1a2-M269) may trace their
origin to early Neolithic farmers of western Europe. However, we think this is not likely given the

44
existence of R1b1a2-M269 not only in western Europe but also in the Near East; such a distribution
implies migrations of M269 males from western Europe to the Near East which do not seem
archaeologically plausible. We prefer the explanation that R-M269 originated in the eastern end of its
distribution, given its first appearance in the Yamnaya males (below) and in the Near East17.

I0412 (Spain_EN)
We determined that this individual belonged to haplogroup I2a1b1 (L161.1:22513718C→T), with
upstream haplogroup I2a1b also supported (CTS1293:7317227G→A, L178:15574052G→A,
M423:19096091G→A). Haplogroup I-L161.1 has not been studied in representative samples of
modern Europeans to our knowledge. A project devoted to this haplogroup in the genetic genealogy
community suggests a relatively high (but not exclusive) occurrence in the present-day British Isles
(https://www.familytreedna.com/public/I2a-L161/; administered by Robert Gabel (Ulrich); accessed
Dec. 09, 2014). This may be a hunter-gatherer lineage that was absorbed by early farmers of western
Europe, as its present-day distribution and discovery in an early Neolithic Iberian suggest.

I0411 (Spain_EN_relative_of_I0410)
This individual is not included in the Spain_EN sample as we that determined it was a relative to
individual I0410 and we retained I0410 because of the latter’s better quality. We could only assign it
to haplogroup F based on mutation P135:21618856C→T. Of the known subclades of F, it was found
to be ancestral for haplogroup G (F1551:9448354A→G), I1 (M450:7548915G→A), I2a
(S247:15224591G→A), J (CTS26:2675457A→T, YSC0000228:22172960G→T), L1b2
(M274:22737801C→T), T (PF5607:8459278G→A, CTS5268:16174116C→T,
CTS7749:17644174C→T), O2b (M176:2655180G→A), Q1a2a (L475:18146921G→A), Q1b1
(FGC1861:21365952G→A), R1a1a (L449:22966756C→T) and R1b1c2 (V35:6812012T→A).

I0405 (Spain_MN)
This individual was assigned to haplogroup I2a1a1 (L672:22228628T→A). We note, however, that
haplogroup H2 is also supported (L279:6932824G→T, L285:21869856C→T). Given the occurrence
of haplogroup I2a chromosomes in many individuals from prehistoric Europe (this study and ref.5-9),
assignment to I2a1a1 seems plausible. Haplogroup I-L672 is nested within haplogroup I-M26, which
is rare in Europe today except in Sardinians (40.9%) and other populations from Southwestern
Europe20. Haplogroup H2 is detected in an early Neolithic Starcevo individual (I0174, below), so we
cannot determine this individual’s haplogroup with certainty.

I0406 (Spain_MN)
This individual was assigned to haplogroup I2a2a1 (CTS9183:18732197A→G) with upstream
haplogroup I2a2a also supported (L368:6931594C→T, L34:7716262A→C, P221:8353707C→A,
P223:16699334C→G, P222:18888200C→G, M223:21717307G→A and P220:24475669G→T). We
could exclude downstream haplogroups I2a2a1a1a (L1195:18865320G→A), I2a2a1b1
(L702:7629205C→T), I2a2a1b2a (L801:21763755A→C), and I2a2a1b2b (L147.3:6753258T→C), so
this individual could be designated I2a2a1*(x I2a2a1a1a, I2a2a1b1, I2a2a1b2a, I2a2a1b2b).

Haplogroup I2a2a1 is nested within haplogroup I2a2a-M223, previously designated I1c, which occurs
at low frequency throughout Europe20, and represents another European hunter-gatherer lineage in the
Middle Neolithic farmers of Spain.

I0174 (Starcevo_EN)
This individual was assigned to haplogroup H2 (L281:8353840T→G). Upstream haplogroup F was
also supported (P142:7218079G→A, P145:8424089G→A, P138:14199284T→C,

45
P316:16839641A→T, P14:17398598C→T, P159:18097251C→A). An individual bearing mutation
P96 which also defines haplogroup H2 was found in the Netherlands21; while haplogroup H is rare in
present-day Europeans, its discovery in I0174 suggests that it was present in Neolithic Europe.

I0116 (Unetice_EBA)
This individual was assigned to haplogroup I2c2 (PF3827:22444389T→A) and upstream haplogroups
I2c (L597:18887888T→A), I2 (M438:16638804A→G) were also supported.

I0804 (Unetice_EBA)
This individual was assigned to haplogroup I2 (L68:18700150C→T) with upstream haplogroup F also
supported (P158:17493513C→T).

I0114 (Unetice_EBA_relative_of_I0117)
This individual was initially assigned to haplogroup I2a2a (L368:6931594C→T) with upstream
haplogroups I2a2 (L181:19077754G→T, P218:17493630T→G, P217:7628484C→T) and I2
(M438:16638804A→G) also supported. However, the sample was ancestral for other mutations
defining haplogroup I2a2a (L34:7716262A→C, P223:16699334C→G, M223:21717307G→A), so it
is possible that either it represents a branch of the Y-chromosome phylogeny that possessed the L368
but not the L34, P223, and M223 mutations, or that the derived L368C→T represents ancient DNA
damage. We thus assign it only to haplogroup I2a2.

I0231 (Yamnaya)
This individuals was assigned to haplogroup R1b1a2a2 (CTS1078:7186135G→C,
Z2105:15747432C→A) with upstream haplogroups R1b1a2a (L23:6753511G→A), and R1b1a2
(PF6399:2668456C→T, L265:8149348A→G, PF6434:8411202A→G, L150.1:10008791C→T,
PF6482:18381735A→G, M269:22739367T→C) also supported. It was ancestral for R1b1a2a2a
(L584:28731917C→T), so it could be designated R1b1a2a2*(xR1b1a2a2a).

Notably, the individual did not belong to haplogroup R1b1a2a1-M412 (8502236G→A), and it has
been observed that R-L23*(xM412) chromosomes “often exceed 10% frequency in the Caucasus,
Turkey and some SE Europe and Circum-Uralic populations” but “they typically display frequencies
≤5% in Western Europe (except for an instance of 27% in Switzerland's Upper Rhone Valley) in
contrast to the prominent spread of derived M412 varieties in West Europe (Figure 1f).” (ref. 17). A
study of modern Armenians22 reports a frequency of ~28% of L23 in modern Armenians while noting
that “The derived M412 allele, which is found in nearly all haplogroup R1b1b1*-L23 chromosomes in
Europe, is absent in the sampled Armenians, which also exhibit a scarcity of haplotype sharing with
Europeans, suggesting a limited role for Armenians in the introduction of R1b into Europe.”
Moreover, results from the Armenian DNA Project
(https://www.familytreedna.com/public/ArmeniaDNAProject/default.aspx?section=ysnp;
administered by Hovann Simonian, Mark Arslan, and Peter Hrechdakian; accessed Dec 09, 2014)
indicate the presence of the Z2103 derived state in many modern Armenians.

It is not possible to determine whether the appearance of R-Z2103 in the Yamnaya individual is due to
(i) gene flow from the south to the steppe and related to the autosomal signal of “dilution” of Eastern
European hunter-gatherers, or (ii) gene flow from the steppe to the south. Modern Armenians have a
signal of admixture from the Yamnaya, as when we test f3-statistics of the form f3(Armenian;
Yamnaya, X) we find the lowest Z-score for f3(Armenian; Yamnaya, BedouinB) = -0.00296 (Z=-7.1).
However, the lowest Z-score of statistics of the form f3(Armenian; X, Y) involves the (X, Y) =
(LBK_EN, Sindhi) pair (value -0.00575, Z=-15.3), so the signal of admixture from the Yamnaya is

46
not the strongest one for Armenians. Moreover, as shown in SI 7, the Yamnaya have a negative f3-
statistic with (X, Y) = (Karelia_HG, Armenian). A negative statistic for both Armenians and Yamnaya
with each other as a reference population may suggest that a third (unsampled) population admixed
into both the Yamnaya and to Armenians. The question of directionality can only be furthered
elucidated by the study of additional ancient samples from the Caucasus, Near East and the steppe.

I0370 (Yamnaya)
This individual was assigned to haplogroup R1b1a2a2 (CTS1078/Z2103:7186135G→C), with
upstream haplogroups R1b1a2 (M269:22739367T→C, L150.1:10008791C→T), R1b1a
(L320:4357591C→T) also supported.

I0429 (Yamnaya)
This individual was assigned to haplogroup R1b1a2a2 (Z2105:15747432C→A) and to the upstream
haplogroups R1b1a2a (L23:6753511G→A) and R1b1a2 (L150.1:10008791C→T,
M269:22739367T→C). It was ancestral for R1b1a2a2a (L584:28731917C→T) and so could be
designated R1b1a2a2*(x R1b1a2a2a).

I0438 (Yamnaya)
This individual could also be assigned to haplogroup R1b1a2a2 (Z2105:15747432C→A). It could
also be assigned to the upstream haplogroups R1b1a2a (L23:6753511G→A), R1b1a
(L320:4357591C→T). It was ancestral for R1b1a2a2a (L584:28731917C→T), and R1b1a2a2c
(CTS7822:17684699A→T), so it could be designated R1b1a2a2*(xR1b1a2a2a, R1b1a2a2c).

I0439 (Yamnaya)
This individual could be assigned to haplogroup R1b1a (P297:18656508G→C), with upstream
haplogroup R1 (M173:15026424A→C, M306:22750583C→A) also supported. It was ancestral for
haplogroup R1b1a2a1 (L51:8502236G→A) and so could be designated R1b1a*(xR1b1a2a1).

I0443 (Yamnaya)
This individual could only be assigned to haplogroup R1b1a2a (L49.1:2842212T→A,
L23:6753511G→A). It could also be assigned to the upstream haplogroups R1b1a2
(PF6399:2668456C→T, L150.1:10008791C→T, L1353:19179540G→A, PF6509:22190371A→G,
M269:22739367T→C, CTS12478:28590278G→A). The individual was ancestral for haplogroup
R1b1a2a1 (L51/M412:8502236G→A) and, unlike I0231, I0370 and I0438 also for R1b1a2a2
(Z2105:15747432C→A). Thus, it could be designated as R1b1a2a*(xR1b1a2a1, R1b1a2a2).

I0444 (Yamnaya)
The individual could be assigned to haplogroup R1b1a2a2 (CTS1078/Z2103:7186135G→C) and also
to the upstream haplogroups R1b1a2 (L150.1:10008791C→T) and R1b1 (M415:9170545C→A).

Summarizing the results from the Yamnaya males, all seven belonged to haplogroup R1b1a. Six of
these could be further assigned to haplogroup R1b1a2a, and five of these to haplogroup R1b1a2a2.
The uniformity of R1b Y-chromosomes in this sample suggests a patrilineal organization of the
Yamnaya, or at least of the people who were given expensive Kurgan burials. We cannot exclude the
presence of other haplogroups in the general population, or in other individuals located elsewhere in
the expansive Yamnaya horizon23. We also emphasize the absence of M412 (the dominant lineage
within haplogroup R-M269 in Europe) in this sample, as well as the absence of the R1a haplogroup
which was detected in the Corded Ware and Late Bronze Age Halberstadt individual from central
Europe. A survey of other European steppe groups may reveal the more immediate patrilineal kin of
the major founding lineages of modern European R1a and R1b chromosomes.

47
Discussion
In Table S4.3 we summarize results from previouslys of 61 ancient European and 25 ancient
Siberian/Central Asian Y-chromosomes >1,000BCE. In combination with our own results (Table
S4.2), this summary makes it clear that only a single R1b Y-chromosome has been found in 70
ancient Europeans outside Russia (1.4%) before the Late Neolithic period. In contrast, all 9 ancient
individuals (100%) from Russia belonged to haplogroups R1a and R1b, and 18/23 (78%) Bronze Age
individuals from Central Asia/Siberia belonged to haplogroup R1a. In Europe except Russia, both R1a
and R1b have been found in the Late Neolithic and Bronze Age periods for a combined frequency of
6/10 (60%). Present-day Europeans have high frequencies4,17 of haplogroups R1a and R1b.

Thus, it appears that before ~4,500 years ago, the frequency of R1a and R1b in Europe outside Russia
was very low, and it rose in the Late Neolithic/Bronze Age period. The young, star-like phylogenies
of these two haplogroups24 also suggest relatively recent expansions. The ubiquity of these
haplogroups in Russia, Siberia, and Central Asia suggest that their rise in Europe was likely to have
been due to a migration from the east, although more work is needed to trace these migrations and
also to correlate them with regions of the world that have not yet been studied with ancient DNA
(such as southern Europe, the Caucasus, the Near East, Iran, and Central and South Asia).
Nonetheless, the Y-chromosome results suggest the same east-to-west migration as our analysis of
autosomal DNA.

48
Table S4.3: Previous Y-chromosome studies on ancient Europeans, Siberians and Central
Asians >1,000BCE. We do not include the Motala individuals from ref.5 in this tabulation as they are
included in Table S4.2. We use haplogroup nomenclature from ISOGG version 9.129 and note the
terminal derived mutation assessed for each sample.

First author Year N Country Period Haplogroup

EUROPE
Seguin-Orlando25 2014 1 Russia Upper Paleolithic H/G C1- F3393
Haak3,+ 2008 1 Germany Late Neolithic R1a1-SRY10831.2
Haak10 2010 1 Germany Early Neolithic G2a2b-S126
2 Germany Early Neolithic F-M89
Lacan12,* 2011 5 Spain Early Neolithic G2a-P15
1 Spain Early Neolithic E1b1b1a1b1a-V13
Lacan8,** 20 France Middle Neolithic G2a-P15
2 France Middle Neolithic I2a1-P37.2
Lacan26,*** 2011 2 France Middle Neolithic I2a1
Lee2 2012 1 Germany Late Neolithic R1b1a2-M269
1 Germany Late Neolithic R1b-M343
Lazaridis5 2014 1 Luxembourg Mesolithic H/G I2a1b-L178
Olalde27 2014 1 Spain Early Neolithic C1a2-V20
Skoglund6 2014 1 Sweden Neolithic H/G I2a1-P37.2
Gamba7 2014 2 Hungary Early Neolithic I2a-L460
2 Hungary Early Neolithic C1a2-V20/V184
1 Hungary Late Bronze Age J2a1-S57
Szécsényi-Nagy9 2014 5 Hungary Early Neolithic G2a-P15
4 Hungary Early Neolithic G2a2b-S126
3 Hungary Early Neolithic F-M89
1 Hungary Early Neolithic I2a1-P37.2
1 Hungary Early Neolithic I-M170
1 Hungary Early Neolithic I1-M253
Keller11 2012 1 Italy Middle Neolithic G2a2a1b-L91
SIBERIA & CENTRAL ASIA
Fu 2014 1 Russia Upper Paleolithic H/G K
Raghavan18 2014 1 Russia Upper Paleolithic H/G R
28
Keyser 2009 7 Russia Bronze Age R1a1a-M17
1 Russia Bronze Age C
Hollard29 2014 4 Mongolia Bronze Age R1a1a1b2-Z93
3 Mongolia Bronze Age Q1a2a1-L54
1 Mongolia Bronze Age C-M130
30
Li 2010 7 China Bronze Age R1a1a-M198
+
Three individuals belonged to this haplogroup but the individuals were related to each other.
*
One of these 5 individuals was inferred to belong to haplogroup G2a based on its Y-STR profile.
**
Three of these 20 individuals were inferred to belong to haplogroup G2a based on Y-STR profile.
***
These individuals from the dolmen of la Pierre Fritte were inferred to belong to haplogroup I2a1
based on Y-STR profile.

49
References

1. Skoglund, P., Storå, J., Götherström, A. & Jakobsson, M. Accurate sex identification of
ancient human remains using DNA shotgun sequencing. Journal of Archaeological Science
40, 4477-4482, (2013).
2. Lee, E. J. et al. Emerging genetic patterns of the european neolithic: Perspectives from a late
neolithic bell beaker burial site in Germany. American Journal of Physical Anthropology 148,
571-579, (2012).
3. Haak, W. et al. Ancient DNA, Strontium isotopes, and osteological analyses shed light on
social and kinship organization of the Later Stone Age. Proceedings of the National Academy
of Sciences, (2008).
4. Underhill, P. A. et al. The phylogenetic and geographic structure of Y-chromosome
haplogroup R1a. Eur. J. Hum. Genet., (2014).
5. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).
6. Skoglund, P. et al. Genomic Diversity and Admixture Differs for Stone-Age Scandinavian
Foragers and Farmers. Science 344, 747-750, (2014).
7. Gamba, C. et al. Genome flux and stasis in a five millennium transect of European prehistory.
Nat. Commun. 5, 5257 (2014).
8. Lacan, M. et al. Ancient DNA reveals male diffusion through the Neolithic Mediterranean
route. Proceedings of the National Academy of Sciences, (2011).
9. Szécsényi-Nagy, A. et al. Tracing the genetic origin of Europe's first farmers reveals insights
into their social organization. bioRxiv, DOI: 10.1101/008664, (2014).
10. Haak, W. et al. Ancient DNA from European Early Neolithic Farmers Reveals Their Near
Eastern Affinities. PLoS Biol 8, e1000536, (2010).
11. Keller, A. et al. New insights into the Tyrolean Iceman's origin and phenotype as inferred by
whole-genome sequencing. Nat Commun 3, 698, (2012).
12. Lacan, M. et al. Ancient DNA suggests the leading role played by men in the Neolithic
dissemination. Proceedings of the National Academy of Sciences 108, 18255-18259, (2011).
13. Francalacci, P. et al. Low-Pass DNA Sequencing of 1200 Sardinians Reconstructs European
Y-Chromosome Phylogeny. Science 341, 565-569, (2013).
14. Sikora, M. et al. Population Genomic Analysis of Ancient and Modern Genomes Yields New
Insights into the Genetic Ancestry of the Tyrolean Iceman and the Genetic Structure of
Europe. PLoS Genet. 10, e1004353, (2014).
15. Skoglund, P. et al. Origins and Genetic Legacy of Neolithic Farmers and Hunter-Gatherers in
Europe. Science 336, 466-469, (2012).
16. Mendez, F. L. et al. Increased resolution of Y chromosome haplogroup T defines
relationships among populations of the Near East, Europe, and Africa. Hum. Biol. 83, 39-53,
(2011).
17. Myres, N. M. et al. A major Y-chromosome haplogroup R1b Holocene era founder effect in
Central and Western Europe. Eur J Hum Genet 19, 95-101, (2011).
18. Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of Native
Americans. Nature 505, 87-91, (2014).
19. Semino, O. et al. The Genetic Legacy of Paleolithic Homo sapiens sapiens in Extant
Europeans: A Y Chromosome Perspective. Science 290, 1155-1159, (2000).
20. Rootsi, S. et al. Phylogeography of Y-Chromosome Haplogroup I Reveals Distinct Domains
of Prehistoric Gene Flow in Europe. The American Journal of Human Genetics 75, 128-137,
(2004).
21. Karafet, T. M. et al. New binary polymorphisms reshape and increase resolution of the human
Y chromosomal haplogroup tree. Genome Res 18, 830-838, (2008).
22. Herrera, K. J. et al. Neolithic patrilineal signals indicate that the Armenian plateau was
repopulated by agriculturalists. Eur. J. Hum. Genet. 20, 313-320, (2012).
23. Anthony, D. W. The Horse, the Wheel, and Language: How Bronze-Age Riders from the
Eurasian Steppes Shaped the Modern World. (Princeton University Press, 2007).

50
24. Hallast, P. et al. The Y-Chromosome Tree Bursts into Leaf: 13,000 High-Confidence SNPs
Covering the Majority of Known Clades. Mol. Biol. Evol., DOI: 10.1093/molbev/msu1327,
(2014).
25. Seguin-Orlando, A. et al. Genomic structure in Europeans dating back at least 36,200 years.
Science 346, 1113-1118, (2014).
26. Lacan, M. La néolithisation du bassin méditerranéen : apports de l'ADN ancien Doctorat
thesis, Université Toulouse III Paul Sabatier, (2011).
27. Olalde, I. et al. Derived immune and ancestral pigmentation alleles in a 7,000-year-old
Mesolithic European. Nature 507, 225-228, (2014).
28. Keyser, C. et al. Ancient DNA provides new insights into the history of south Siberian
Kurgan people. Hum. Genet. 126, 395-410, (2009).
29. Hollard, C. et al. Strong genetic admixture in the Altai at the Middle Bronze Age revealed by
uniparental and ancestry informative markers. Forensic Science International: Genetics 12,
199-207, (2014).
30. Li, C. et al. Evidence that a West-East admixed population lived in the Tarim Basin as early
as the early Bronze Age. BMC Biol. 8, 15, (2010).

51
Supplementary Information 5
Principal Component Analysis

Iosif Lazaridis* and David Reich

* To whom correspondence should be addressed ([email protected])

In Fig. 2a we show the PCA analysis that was conducted using smartpca from the
EIGENSOFT1 software package on 777 present-day West Eurasians2 and 92 ancient
individuals at 66,225 transversion SNPs.

We excluded transition SNPs in this analysis as non-UDG-treated ancient DNA libraries from
the literature had high rates of damage causing apparent C→T and G→A substitutions, which
potentially leads to misleading results, for example by causing ancient DNA samples to
appear as extreme outliers, dominating one the first two principal components.

To avoid visual clutter in Fig. 2a, we represented modern individuals as grey dots, and used
colored and labeled symbols to represent the ancient individuals.

In Figure S5.1 we show the complement of Fig. 2a by color coding modern individuals and
representing ancient ones as grey dots.

In Fig. S5.2 we show a PCA over all 354,212 SNPs, with ancient individuals projected (using
the lsqproject: YES option) onto the variation of the modern ones. Topologically, this is very
similar to Fig. 2a. An interesting aspect of Fig. S5.2 is that while most clusters are preserved,
there is a spurious clustering of MA1 with the Yamnaya. However, MA1 and the Yamnaya
are clearly not a clade, as, for example, the statistic f4(MA1, Yamnaya; Karitiana, Chimp) is
significantly negative both on all SNPs (Z=-6.3), or limiting to transversions (Z=-3.9).

We chose to show Fig. 2a in the main paper as limiting to transversions did not seem to
impede our ability to distinguish among populations, did not produce any spurious visual
clustering that contradicted the analysis of f-statistics (SI7) and allowed us to treat ancient and
modern individuals in the same way.

52
Figure S5.1: Complement of PCA of Fig. 2a, showing ancient individuals as grey dots
and color coding modern ones.

53
Figure S5.2: PCA analysis with ancient individuals projected onto the variation of the
present-day ones.

References

1. Reich, D., Price, A. L. & Patterson, N. Principal component analysis of genetic data. Nat.
Genet. 40, 491-492, (2008).
2. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).

54
Supplementary Information 6
ADMIXTURE analysis

Iosif Lazaridis* and David Reich

* To whom correspondence should be addressed ([email protected])

We carried out admixture analysis using ADMIXTURE1 (v. 1.23), including 2,437 individuals (2,345
modern humans2 and 92 ancient ones. This includes Ust’-Ishim3, Kostenki144, MA15, LaBrana16, the
Iceman7, Loschbour and Stuttgart2, Neolithic, Copper, Bronze, and Iron Age individuals from
Hungary8, a farmer and hunter-gatherers from Sweden9, and 67 new individuals from the present
study. We excluded individuals I0114 and I0411 that appear to be related to individuals I0117 and
I0410 respectively (Online methods), as in our experience, unsupervised ADMIXTURE clustering
analysis often infers spurious clusters when pairs of relatives are included in the analysis dataset.

The initial set of 354,212 SNPs was pruned for linkage disequilibrium in PLINK10 (v. 1.07) using
parameters --indep-pairwise 200 25 0.5, resulting in a pruned set of 230,538 SNPs used for analysis.
(LD-pruning is recommended by the authors of ADMIXTURE, and we chose a relatively high r 2=0.5
in order to retain a large number of SNPs, as more aggressive pruning would reduce the SNPs
available for analysis for the individuals with most missing data.) ADMIXTURE was run with cross-
validation (--cv) and with 100 different seeds (chosen at random from the set 1…1000). The number
of ancestral populations (K) was varied between 2 and 20.

Figure S6.1: Log likelihood of best of 100 replicates as a function of K

In Extended Data Fig. 1 we show the admixture proportions for 236 populations, including 33 ancient
ones. The 33 ancient populations are placed at the beginning of the figure and individuals are given a
higher line width for added clarity. The populations are clustered (using the function hclust in R)

55
using their mixture proportions across all K values, with the goal of visually clustering populations
that have similar mixture proportions. We visualize in Extended Data Fig. 1 the runs (for each value
of K) that had the highest loglikelihood. In Fig. S6.1 we show the change in log likelihood with K,
showing a steady increase with increasing K.

In Fig. S6.2 we show a box-and-whiskers plot of the cross-validation error for K=2 to K=20. This
seems to plateau as K increases. The lowest median CV error is attained for K=16.

Figure S6.2: Cross-validation error for 100 replicates as a function of K

Fig. S6.3 shows a detail of Extended Data Fig. 1 showing only the ancient populations. We observe
that none of the populations with at least two individuals have any major outliers, thus justifying our
decision to pool individuals from each archaeologically defined population for the main analyses of
our paper. The relative homogeneity of the populations also suggests that the number of SNPs
available for analysis, even after LD-pruning, is adequate to correctly cluster individuals.

The Early/Middle Neolithic European populations belong almost entirely to the “orange” ancestral
population from K=2 to K=8, while hunter-gatherers show a relationship to eastern non-Africans from
K=3 to K=8, consistent with sharing more genetic drift with these populations due to their lack of
“Basal Eurasian” ancestry2. From K=4 to K=6, the hunter-gatherers and late Neolithic/Bronze Age
(LN/BA) groups possess some of the “pink” component that is dominant in Native Americans; this
may reflect either the presence of west Eurasian-related “Ancient North Eurasian” ancestry in Native
Americans5 or of the same type of ancestry in European hunter-gatherers. An interesting pattern
occurs at K=8, with all the late LN/BA groups from central Europe and the Yamnaya having some of
the “light green” component that is lacking in earlier European farmers and hunter-gatherers; this
component is found at high frequencies in South Asian populations and its co-occurrence in late

56
Neolithic/Bronze Age Europeans (but not earlier ones) and South Asians might reflect a degree of
common ancestry associated with late Neolithic migratory movements (e.g., the ~5,800-year old
TMRCA of Y-chromosome haplogroup R1a-M417 suggests some gene flow affecting both Europe
and South Asia in this time frame11, although this date is subject to uncertainty due to poor estimates
of the human mutation rate.)

At K=9 a European hunter-gatherer ancestral population (“dark blue”) appears; this was not present in
an earlier analysis of the Human Origins modern populations and a much smaller number of ancient
individuals2. The inclusion of a large number of ancient hunter-gatherers has probably caused such an
ancestral population to appear in this analysis. European farmers now appear to be mixture of a Near
Eastern (orange) and European hunter-gatherer (dark blue) ancestral populations, with an increase in
the hunter-gatherer ancestry during the Middle Neolithic (reflecting the “resurgence” of such ancestry
shown in PCA, Fig. 2a) and also during the Late Neolithic. Note, also, the persistent presence of the
“light green” component that ties LN/EBA groups to South Asia between K=9 and K=15. A similar
(darker green) component also distinguishes LN/EBA groups from earlier ones at K=16; this
component appears to be highly represented in groups from South Asia, the Near East, and the
Caucasus. The existence of this component may correspond to the evidence for “dilution” of EHG
ancestry in the Yamnaya (SI7), showing them to have evenly split ancestry between the “dark blue”
hunter-gatherer and “dark green” component; the analysis of SI9 also suggests an even split between
an EHG and a Near Eastern component in the ancestry of the Yamnaya. The “dark green” component
seems to have been carried from a Yamnaya-related population to the Corded Ware and other Late
Neolithic and Bronze Age populations of central Europe. A useful topic for future work is to study the
relationship of LN/BA populations to contemporary South Asians, Caucasian and Near Eastern
populations and to see if this affinity (in contrast to earlier Europeans) may be related to the dispersal
of Indo-European languages.

57
Figure S6.3: Ancient populations ADMIXTURE results. For the full set of populations, see Extended Data Fig. 1.

58
References

1. Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry in

unrelated individuals. Genome Res. 19, 1655-1664, (2009).
2. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).
3. Fu, Q. et al. Genome sequence of a 45,000-year-old modern human from western Siberia.
Nature 514, 445-449, (2014).
4. Seguin-Orlando, A. et al. Genomic structure in Europeans dating back at least 36,200 years.
Science 346, 1113-1118, (2014).
5. Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of Native
Americans. Nature 505, 87-91, (2014).
6. Olalde, I. et al. Derived immune and ancestral pigmentation alleles in a 7,000-year-old
Mesolithic European. Nature 507, 225-228, (2014).
7. Keller, A. et al. New insights into the Tyrolean Iceman's origin and phenotype as inferred by
whole-genome sequencing. Nat. Commun. 3, 698, (2012).
8. Gamba, C. et al. Genome flux and stasis in a five millennium transect of European prehistory.
Nat. Commun. 5, 5257 (2014).
9. Skoglund, P. et al. Genomic Diversity and Admixture Differs for Stone-Age Scandinavian
Foragers and Farmers. Science 344, 747-750, (2014).
10. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based
linkage analyses. Am. J. Hum. Genet. 81, 559-575, (2007).
11. Underhill, P. A. et al. The phylogenetic and geographic structure of Y-chromosome
haplogroup R1a. Eur. J. Hum. Genet., (2014).

59
Supplementary Information 7
Population relationships and admixture using f-statistics

Iosif Lazaridis*, Nick Patterson, and David Reich

* To whom correspondence should be addressed ([email protected])

We carried out analysis of population relationships with f-statistics1,2 to confirm formally the
qualitative patterns observed via Principal Components Analysis3,4 (PCA) (SI5) and
ADMIXTURE5 analysis (SI6). We use italics to refer to a variable (that is iterated over
different populations, e.g. MN is Spain_MN, Esperstedt_MN, or Baalberge_MN) and non-
italics to refer to a concrete population (e.g., MN is a meta-population consisting of all
Middle Neolithic individuals). Significance of statistics is assessed using a block jackknife6
with a 5cM block size, using the ADMIXTOOLS suite of programs1.

Figure S7.1: “Outgroup” statistic f3(Dinka; X, Y). Panel (a) shows that population Y2 shares
more genetic drift (indicated in blue) with X than Y1 does. In panel (b), an admixed population
X shares genetic drift with Y1 and Y2 that is weighted by the admixture proportions α1 and α2.
(a) (b)

Outgroup f3-statistics
We first examined “outgroup” f3-statistics7 of the form f3(Dinka; X, Y) (Table S7.1). These
measure the amount of genetic drift shared by two populations X and Y after their divergence
from the Dinka, an African outgroup population that is phylogenetically close to the Out-of-
Africa bottleneck8 and has no detectible archaic Eurasian admixture9 (results with other
African outgroups such as Yoruba or Ju_hoan_North (San) are qualitatively similar). Fig.
S7.1 shows this statistic. We note that in Fig. S7.1b an admixed population X may share more
genetic drift with Y2 than with Y1 even if it has less ancestry related to Y2 than to Y1; this is due
to the fact that Y2 may have experienced more genetic drift than Y1 since their separation from
Dinka. Thus, f3(Dinka; X, Y2)> f3(Dinka; X, Y1) does not imply that α2>a1. It does, however,
imply that α2>0 (that X shares at least some genetic drift with Y2 that is not shared with Y1).

These observations are consistent with the groupings of Fig. 2a (PCA plot) but show only the
population that shares most genetic drift with each population X. In Extended Data Fig. 2 we
show the pairwise heat matrix of this statistic for all populations under study. All Holocene
ancient European group together at the exclusion of Upper Paleolithic Eurasians MA17,
Kostenki1410, and Ust_Ishim11, suggesting that they share common genetic drift as a result of

60
at least some shared history. Two major groupings of Holocene ancient Europeans involve
Early/Middle Neolithic farmers on the bottom-left quadrant, and hunter-gatherers, Yamnaya,
and Late Neolithic/Bronze Age central Europeans on the top-right quadrant. The strong
relationship between European hunter-gatherers is clearly seen in the middle of the plot, with
finer WHG (Loschbour, LaBrana1, and HungaryGamba_HG) and EHG (Karelia_HG and
Samara_HG) groupings also evident. MA1 clearly shares less genetic drift with Europeans
than Europeans do with each other. However, MA1 shares more genetic drift with EHG than
Europeans shared with EHG. Early and Middle Neolithic farmers group with each other, with
EN and MN sub-groupings also evident; the Middle Neolithic group also includes Copper and
Bronze Age individuals from Hungary12. Importantly, the Late Neolithic/Bronze Age samples
group with the hunter-gatherers and Yamnaya, and not the Early/Middle Neolithic ones, as
might be expected if they were a direct continuation of the EN/MN population of Europe.

Table S7.1: Populations maximizing f3(Dinka; X, Y) for each ancient population X.

X Y f3(Dinka; X, Y) std. error
Alberstedt_LN Esperstedt_MN 0.167986 0.002435
Baalberge_MN LBKT_EN 0.173694 0.006135
Bell_Beaker_LN HungaryGamba_HG 0.163192 0.001971
BenzigerodeHeimburg_LN Motala_HG 0.165990 0.001716
Corded_Ware_LN Karsdorf_LN 0.166848 0.003064
Esperstedt_MN Spain_MN 0.175997 0.001956
Halberstadt_LBA Esperstedt_MN 0.166007 0.002412
HungaryGamba_BA HungaryGamba_HG 0.165637 0.002290
HungaryGamba_CA Starcevo_EN 0.169771 0.003655
HungaryGamba_EN Starcevo_EN 0.171040 0.002473
HungaryGamba_HG Loschbour 0.212255 0.002640
HungaryGamba_IA SwedenSkoglund_NHG 0.156723 0.002325
Iceman Starcevo_EN 0.170761 0.003149
Karelia_HG Samara_HG 0.193494 0.002698
Karsdorf_LN SwedenSkoglund_MHG 0.170276 0.011495
Kostenki14 Loschbour 0.149636 0.002404
LaBrana1 Loschbour 0.212582 0.002587
LBK_EN Esperstedt_MN 0.175495 0.001809
LBKT_EN Starcevo_EN 0.182378 0.008451
Loschbour LaBrana1 0.212582 0.002587
MA1 Samara_HG 0.173080 0.002898
Motala_HG Loschbour 0.194857 0.002037
Samara_HG Karelia_HG 0.193494 0.002698
Spain_EN Spain_MN 0.174572 0.001652
Spain_MN Esperstedt_MN 0.175997 0.001956
Starcevo_EN LBKT_EN 0.182378 0.008451
Stuttgart LBK_EN 0.168289 0.001642
SwedenSkoglund_MHG HungaryGamba_HG 0.181851 0.005697
SwedenSkoglund_MN Baalberge_MN 0.165744 0.002777
SwedenSkoglund_NHG Motala_HG 0.192765 0.001963
Unetice_EBA Motala_HG 0.165904 0.001468
Ust_Ishim Chane 0.125962 0.002174
Yamnaya Samara_HG 0.169258 0.001955

61
We also plot the value of the f3(Dinka; X, Y) for different populations X of particular interest
(Fig. S7.3-7). These outgroup f3-statistics strongly suggest that the ancestry of the Late
Neolithic/Bronze Age groups was not entirely drawn from the local Early/Middle Neolithic
population. In particular, the Corded Ware (Fig. S7.7a) shares more genetic drift with both the
hunter-gatherers from Russia, the hunter-gatherers from Sweden, and the Yamnaya group
from Russia than with any earlier (MN or EN) farmer groups from Europe. This strongly
suggests that this group, which occupies the northernmost position in the PCA plot (Fig. 2a),
does indeed possess some ancestry that is not rooted in its local central European context.

Figure S7.3 [Western European hunter-gatherers]: Populations that share most genetic
drift with a population X according to the statistic f3(Dinka; X, Y). The estimated value
and ±3 standard errors are indicated. Three hunter-gatherers from Spain (LaBrana1),
Luxembourg (Loschbour), and Hungary (HungaryGamba_HG, or KO1 of ref.12) share more
genetic drift with each other.

(a) Loschbour (b) LaBrana1

(c) HungaryGamba_HG

62
Figure S7.4 [Ancient North Eurasians and Eastern European hunter-gatherers]:
Populations that share most genetic drift with a population X according to the statistic
f3(Dinka; X, Y). The estimated value and ±3 standard errors are indicated. The MA1
Upper Paleolithic shares more genetic drift with the two eastern European hunter-gatherers
Karelia_HG and Samara_HG; these two EHG samples share more drift with each other.

(a) MA1 (b) Karelia_HG

(c) Samara_HG

63
Figure S7.5 [Early Neolithic Europeans]: Populations that share most genetic drift with a
population X according to the statistic f3(Dinka; X, Y). The estimated value and ±3 standard
errors are indicated. Early Neolithic Europeans share more genetic drift either with other EN
populations, or the MN populations that succeed them.

(a) LBK_EN (b) LBKT_EN

(c) Starcevo_EN (d) Spain_EN

64
Figure S7.6 [Middle Neolithic Europeans]: Populations that share most genetic drift with a
population X according to the statistic f3(Dinka; X, Y). The estimated value and ±3 standard
errors are indicated. Middle Neolithic Europeans share more genetic drift either with other MN
populations, or the EN populations that precede them.

(a) Esperstedt_MN (b) HungaryGamba_CA

(c) Spain_MN (d) SwedenSkoglund_MN

65
Figure S7.7 [Late Neolithic / Bronze Age Europeans]: Populations that share most genetic drift
with a population X according to f3(Dinka; X, Y). The estimated value and ±3 standard errors
are indicated. These populations share most drift with hunter-gatherers, other LN/BA populations,
Yamnaya, or Esperstedt_MN, the youngest sample from the Middle Neolithic of Germany.

(a) Corded_Ware_LN (b) Bell_Beaker_LN

(c) Unetice_EBA (d) Halberstadt_LBA

66
Western European hunter-gatherers (WHG) in Luxembourg, Spain, and Hungary
In a previous study13, Western European hunter-gatherers (WHG) were defined by the
Loschbour Mesolithic hunter-gatherer from Luxembourg, while noting that the LaBrana1
Mesolithic hunter-gatherer from Spain was closely related to it, a result that is again shown in
Fig. S7.3. An unknown hunter-gatherer (UHG) was proposed to have contributed WHG-
related ancestry into the early European farmers of Germany, thus postulating the existence of
WHG-related populations somewhere in the route from the Near East to central Europe. The
recent publication of KO1, an early Neolithic individual from Hungary12 (named
HungaryGamba_HG here, Fig. S7.3p) showed that such an individual with hunter-gatherer
ancestry did in fact live in central Europe during the early Neolithic period.

To confirm that this was indeed related to Loschbour, we used the statistic f4(Loschbour,
HungaryGamba_HG; Test, Chimp), iterating Test over all other ancient or modern
populations. All but one of these statistics were not significant (|Z|<3), confirming the
impression from PCA (Fig. 2a) and outgroup f3-statistics (above), that HungaryGamba_HG is
related to WHG. The significant statistic was f4(Loschbour, HungaryGamba_HG; LaBrana1,
Chimp) = 0.00429 (Z=5.6) which suggests some structure within the WHG population, as
LaBrana1 shares more alleles with Loschbour than with HungaryGamba_HG. This statistic
continues to be significant when limiting to transversions (Z=3.9).

Eastern European hunter-gatherers (EHG) in Russia

We studied individuals from Russia to identify potential sources of “Ancient North Eurasian”
ancestry in later Europeans, as the best representative of that ancestry, the Mal’ta (MA1)
specimen from Upper Paleolithic Siberia, is both geographically and temporally remote from
Neolithic and present-day Europeans. The two hunter-gatherers, Samara_HG and Karelia_HG
are from far-apart locations in European Russia: Samara_HG is on the Volga, on the steppe,
and Karelia_HG in Lake Onega, far to the north. Despite this fact, these two individuals
appear to be mutually share the most drift, and MA1 shares more drift with them than with
any other populations (Fig. S7. 4).

We confirmed these observations directly by taking the statistic f4(Karelia_HG, Samara_HG;

Test, Chimp) which is not significant for all modern/ancient Test populations (|Z|<1.9),
indicating that these two hunter-gatherers did in fact form a clade with respect to all other
humans. Moreover, the statistic f4(Karelia_HG, Test; MA1, Chimp) is significantly positive
(|Z|>4.7 on all SNPs, and |Z|>3.4 on transversions) for all Test populations except three
(Test=Samara_HG, |Z|=0.3, consistent with Samara_HG and Karelia_HG forming a clade,
and Test=Karsdorf_LN or SwedenSkoglund_MHG where <43,000 sites are available for co-
analysis in all three ancient populations involved in the statistic). We conclude that MA1
shares more alleles with Karelia_HG than with other ancient/modern human populations; it
was previously shown that MA1 shares more alleles with Native Americans7 than any other
modern humans, but our results show that it shares even more alleles with EHG than with
Native Americans, e.g., f4(Karelia_HG, Karitiana; MA1, Chimp) = 0.00388 (Z=5.3).

However, MA1 and EHG do not form a clade with respect to other humans, as, e.g.,
f4(Karelia_HG, MA1; Loschbour, Chimp) = 0.00625 (Z=7.4). In SI8 we show that MA1,
EHG, and WHG cannot be related to each other by a simple tree and at least one of them must
be admixed. While the direction of gene flow cannot be resolved uniquely, the above statistic
clearly shows that there is some common genetic drift shared by all “European hunter-

67
gatherers” (both WHG and EHG) at the exclusion of MA1. This is consistent with the PCA
(Fig. 2a) which shows all European hunter-gatherers occupying the left end of PC1.

Scandinavian Mesolithic and Neolithic hunter-gatherers

Co-analyzing our Motala_HG sample of Mesolithic hunter-gatherers with the StoraForvar11
(SwedenSkoglund_MHG) and later Pitted Ware (PWC) Neolithic hunter-gatherers from
Sweden8 (SwedenSkogldun_NHG), it is evident that 8-5 thousand years ago, the hunter-
gatherers of Sweden were fairly similar in the context of West Eurasian populations (PCA,
Fig. 2a). Indeed, when we study the statistic f4(Motala_HG, SwedenSkoglund_MHG or
SwedenSkoglund_NHG; Test, Chimp), its Z-score is always Z>-2.3 for all other ancient or
modern populations, suggesting that the hunter-gatherers of ref. 8 have no other discernible
ancestry than the Motala_HG of Mesolithic individuals from Sweden. The
SwedenSkoglund_NHG sample is <3,000 cal BCE old, but seems (within the limits of our
resolution) not to have acquired any admixture from other populations that was not already
present in Motala_HG ~2 thousand years earlier.

It was previously shown that both Motala_HG13 and SwedenSkoglund_NHG8 could be

modeled as mixtures of WHG and MA1. We estimate f4(Motala_HG, Loschbour;
Karelia_HG, Chimp) = 0.0041 (Z=7.6), and f4(SwedenSkoglund_NHG, Loschbour;
Karelia_HG, Chimp) = 0.00198 (Z= 2.9), which suggests that Eastern European hunter-
gatherers share more alleles with Scandinavian hunter-gatherers (SHG) than with the WHG.
In SI9 we show that SHG can be modeled as mixtures of EHG and WHG.

Table S7.2: Relationship of Holocene hunter-gatherers to Upper Paleolithic ones.

f4(HG1, HG2;
HG1 HG2 Upper Paleolithic Upper Paleolithic, Chimp) Z
Karelia_HG Motala_HG Kostenki14 -0.00075 -1.3
Loschbour Motala_HG Kostenki14 0.00158 2.9
Loschbour Karelia_HG Kostenki14 0.00247 3.4
Karelia_HG Ust_Ishim Kostenki14 0.00661 7.7
Motala_HG Ust_Ishim Kostenki14 0.00731 10.0
Loschbour Ust_Ishim Kostenki14 0.00897 11.0
Karelia_HG Kostenki14 Ust_Ishim -0.00044 -0.6
Motala_HG Kostenki14 Ust_Ishim -0.00026 -0.4
Karelia_HG Motala_HG Ust_Ishim -0.00016 -0.3
Loschbour Kostenki14 Ust_Ishim 0.00037 0.5
Loschbour Karelia_HG Ust_Ishim 0.00064 0.9
Loschbour Motala_HG Ust_Ishim 0.00052 1.1

Ust Ishim and Kostenki14

To test whether the three main groups of Holocene hunter-gatherers (WHG, EHG, SHG)
differed in their relationship to Upper Paleolithic Eurasians like Kostenki1410 and Ust Ishim11,
we tested statistics of the form f4(HG1, HG2; Upper Paleolithic, Chimp), which we report in
Table S7.2. Holocene hunter-gatherers (and Kostenki14) appear to be equally related to
Ust_Ishim, consistent with this individual being an outgroup to later populations11.
Kostenki14 shares more alleles with Holocene European hunter-gatherers than with
Ust_Ishim, consistent with it being an early European10. There is marginal signal (Z=3.4),
confirmed in transversions only (Z=3.0) that Kostenki14 shares more alleles with Loschbour

68
than with Karelia_HG. The signal also exists when replacing Karelia_HG with Samara_HG,
although weaker (Z=2.7 on all SNPs and Z=2.5 on transversions only), given the lower
number of SNPs available for analysis for that sample. This statistic casts doubt on the long-
term continuity of the hunter-gatherers of eastern Europe, which must be further investigated
with additional hunter-gatherers from different times in that region.

Early Neolithic Europeans were descended from a common ancestral population

We study statistics of the form f4(LBK_EN, Test; Other EN, Chimp), varying Other EN to be
any Early Neolithic population from the set: Starcevo_EN, LBKT_EN, Spain_EN,
HungaryGamba_EN, Stuttgart, and Test to be any modern or ancient (but not Early Neolithic)
population. We use LBK_EN as a baseline for early farmers, as this is the population with the
highest sample size (N=12).

We find only two significantly negative statistics, f4(LBK_EN, Spain_MN; Spain_EN,

Chimp) = -0.00133 (Z=-4.2), and f4(LBK_EN, Esperstedt_MN; Spain_EN, Chimp) = -
0.00173 (Z=-3.6). The first can be interpreted as Spain_MN having some drift in common
with Spain_EN (but not LBK_EN), indicating continuity of Neolithic populations in Spain.
The second is surprising, as it indicates some gene flow from Spanish related populations into
Germany. With these two exceptions, all other statistics are not significant (|Z|<3) or
significantly positive (Z>3), suggesting that all Early Neolithic Europeans had ancestry in
common with the LBK_EN and are thus descended from a common ancestral population.

We also studied statistics of the form f4(EN1, EN2; Test, Chimp) which test whether a pair of
Early Neolithic populations (EN1, EN2) form a clade with respect to a Test population, varying
Test to be any other ancient (but not Early Neolithic) or modern population. We list
significant statistics (|Z|>3) in Table S7.3. We note that in our capture data (which have been
identically processed which is likely to reduce systematic biases due to differences in error
processes), the only significant statistic is f4(LBK_EN, Spain_EN; Loschbour, Chimp) < 0.

Variable WHG-related ancestry of Early and Middle Neolithic Europeans

Early European farmers had WHG-related ancestry13 and it was observed that a Funnel
Beaker culture farmer from Sweden (SwedenSkoglund_MN in our analysis) had more hunter-
gatherer ancestry than the approximately contemporaneous (~5,000 years ago) Tyrolean
Iceman8. We were able to assess the extent of hunter-gatherer admixture in early European
farmers with the statistic f4(LBK_EN, European Farmer; Loschbour, Chimp), which
measures whether Loschbour (the best sample from the WHG population) shares more alleles
with LBK_EN or with another European Farmer population. We chose LBK_EN as the
comparative baseline as it is the population with the most (N=12) individuals. This statistic is
plotted in Fig. S7. 8, showing a striking pattern: Loschbour shares more alleles with all
European farmer groups younger than 6,000 BCE, from Germany, Spain, Sweden, Italy, and
Hungary, than with the early Neolithic LBK_EN sample from Germany. Thus, the early
Neolithic population of central Europe in Germany and Hungary (LBK_EN as well as
HungaryGamba_EN, Stuttgart, LBKT_EN, and Starcevo_EN) had less hunter-gatherer
ancestry than the people that followed them across Europe. We can document the resurgence
at a local level for the countries of Germany, Hungary and Spain, in which we have Neolithic
samples from different periods (Table S7.4).

69
Table S7.3: Significant statistics of the form f4(EN1, EN2; Test, Chimp)
EN1 EN2 Test f4(EN1, EN2; Test, Chimp) Z
LBK_EN HungaryGamba_EN Alberstedt_LN 0.00170 5.4
LBK_EN HungaryGamba_EN Algerian 0.00071 3.0
LBK_EN HungaryGamba_EN Bell_Beaker_LN 0.00121 4.4
LBK_EN HungaryGamba_EN BenzigerodeHeimburg_LN 0.00143 4.6
LBK_EN HungaryGamba_EN Canary_Islanders 0.00081 3.1
LBK_EN HungaryGamba_EN Corded_Ware_LN 0.00164 5.9
LBK_EN HungaryGamba_EN Halberstadt_LBA 0.00212 6.3
LBK_EN HungaryGamba_EN Karelia_HG 0.00102 3.1
LBK_EN HungaryGamba_EN Mbuti 0.00063 3.5
LBK_EN HungaryGamba_EN Motala_HG 0.00117 4.3
LBK_EN HungaryGamba_EN Tunisian 0.00075 3.3
LBK_EN HungaryGamba_EN Tuscan 0.00073 3.0
LBK_EN HungaryGamba_EN Unetice_EBA 0.00138 5.3
LBK_EN Spain_EN Loschbour -0.00108 -3.0
LBK_EN Stuttgart Alberstedt_LN 0.00177 3.8
LBK_EN Stuttgart BenzigerodeHeimburg_LN 0.00174 4.3
LBK_EN Stuttgart Halberstadt_LBA 0.00175 3.8
LBK_EN Stuttgart Karelia_HG 0.00145 3.1
LBK_EN Stuttgart Unetice_EBA 0.00130 3.8
Spain_EN HungaryGamba_EN Alberstedt_LN 0.00214 5.3
Spain_EN HungaryGamba_EN Algerian 0.00090 3.2
Spain_EN HungaryGamba_EN Basque 0.00093 3.3
Spain_EN HungaryGamba_EN Bell_Beaker_LN 0.00142 4.3
Spain_EN HungaryGamba_EN BenzigerodeHeimburg_LN 0.00159 4.3
Spain_EN HungaryGamba_EN Canary_Islanders 0.00108 3.4
Spain_EN HungaryGamba_EN Corded_Ware_LN 0.00163 4.9
Spain_EN HungaryGamba_EN Halberstadt_LBA 0.00178 4.6
Spain_EN HungaryGamba_EN Karelia_HG 0.00134 3.4
Spain_EN HungaryGamba_EN LaBrana1 0.00134 3.3
Spain_EN HungaryGamba_EN Motala_HG 0.00152 4.6
Spain_EN HungaryGamba_EN Saharawi 0.00093 3.2
Spain_EN HungaryGamba_EN Sardinian 0.00109 3.8
Spain_EN HungaryGamba_EN Spanish 0.00090 3.2
Spain_EN HungaryGamba_EN Unetice_EBA 0.00149 4.8
Spain_EN Stuttgart Alberstedt_LN 0.00224 4.4
Spain_EN Stuttgart BenzigerodeHeimburg_LN 0.00190 4.2
Spain_EN Stuttgart Karelia_HG 0.00173 3.4
Spain_EN Stuttgart LaBrana1 0.00203 3.9
Spain_EN Stuttgart Motala_HG 0.00143 3.4
Spain_EN Stuttgart Unetice_EBA 0.00142 3.6

Table S7.4: Resurgence of WHG ancestry in three European countries during the
Middle Neolithic. (The sample from Hungary is from the Copper Age, but shows the same
pattern of resurgence as the earlier Middle Neolithic samples from Germany and Spain).
EN MN f4(EN, MN; Loschbour, Chimp) Z
LBK_EN Baalberge_MN -0.00376 -6.3
LBK_EN Esperstedt_MN -0.00343 -6.1
Spain_EN Spain_MN -0.00275 -6.4
HungaryGamba_EN HungaryGamba_CA -0.00206 -3.3

70
Figure S7.8: Western European hunter-gatherers like Loschbour shared more alleles
with Middle Neolithic European farmers, documenting variable amounts of hunter-
gatherer ancestry in Neolithic Europe. The estimated value of the statistic f4(LBK_EN,
European Farmer; Loschbour, Chimp) and ±3 standard errors are shown.

We also see (Fig. S7.8) that Loschbour shares more alleles (|Z|>4) with both the Tyrolean
Iceman and SwedenSkoglund_MN, two samples from ~5,000 years ago, than with LBK_EN.
We cannot speak of “WHG resurgence” in these two cases, as there are no earlier farmer
samples from these two Sweden and Italy. Loschbour does not (|Z|<1.4) share more alleles
with SwedenSkoglund_MN than it does with Baalberge_MN and Esperstedt_MN, Middle
Neolithic samples from Germany, and neither does Motala_HG (|Z|<1.8). Thus, there is no
statistical evidence for the occurrence of additional hunter-gatherer ancestry in farmers from
Scandinavia above that which existed in contemporaneous farmers of central Europe. As for
the Tyrolean Iceman, it is necessary to collect additional earlier samples from southern
Europe to see if its elevated (relative to LBK_EN) affinity to WHG is due to a resurgence of
such ancestry, or reflects higher WHG ancestry in that region since the early Neolithic. We do
see some evidence that even early Neolithic Europeans varied in their WHG ancestry (Fig.
S7.8) as f4(LBK_EN, Spain_EN; Loschbour, Chimp) = -0.00108 (Z=-3.0), a marginal signal
which, however, persists when limiting to transversions (Z=-3.5), which suggests that while
early farmers in Germany and Spain were closely related, they may have had different levels
of WHG-related admixture.

Source of WHG resurgence

The fact that the resurgence of hunter-gatherer ancestry occurred in several European
countries raises the question of its source. It is possible that pockets of hunter-gatherers14
existed in Europe long after the arrival of first farmers; their later admixture with farmer
communities would account for the observed phenomenon. Alternatively, there may have

71
existed larger areas of Europe where populations with substantial hunter-gatherer ancestry
persisted and gene flow from these refuge areas increases the amount of this ancestry in the
European farmers. We tested statistics of the form f4(EN, MN; WHG1, WHG2) for the (EN,
MN) population pairs of Table S7.4 and the three WHG samples. However, all of these were
not significant (|Z|<2.4), and thus, at present we cannot identify the source of WHG-related
resurgence in any of the available genomes of WHG individuals. However, we can
confidently exclude either the SHG or the EHG as sources of this phenomenon, as f4(EN,
MN; WHG, EHG) = -0.00108 (Z=-6.2), and f4(EN, MN; WHG, SHG) = -0.00057 (Z=-4.5),
which shows that the WHG are a better source of this ancestry in Middle Neolithic Europeans
than the hunter-gatherers of Sweden or Russia.

Dilution of EHG ancestry in the Yamnaya

The Yamnaya pastoralists from the Samara district are not descended only from the EHG that
precede them in eastern Europe. To show this, we estimate the value of the statistic
f4(Karelia_HG, Yamnaya; Test, Chimp) for different modern/ancient Test populations. In
Table S7.5, we show the populations that maximize/minimize this statistic. European hunter-
gatherers, MA1, and Native Americans share more alleles with Karelia_HG than with
Yamnaya, and Near Eastern populations share more alleles with Yamnaya than Karelia_HG.

Table S7.5: Dilution of EHG in the Yamnaya: f4(Karelia_HG, Yamnaya; Test, Chimp)
Test Statistic Z Test Statistic Z
Iraqi_Jew -0.00176 -4.2 Samara_HG 0.00594 8.9
Armenian -0.00161 -3.9 Motala_HG 0.00532 11.3
Abkhasian -0.00157 -3.7 SwedenSkoglund_NHG 0.00476 8.2
Georgian -0.00151 -3.6 MA1 0.00436 6.7
Yemenite_Jew -0.00149 -3.6 SwedenSkoglund_MHG 0.00414 2.9
Saudi -0.00137 -3.3 Mixe 0.00401 8.3
Iranian_Jew -0.00131 -3.2 Wayuu 0.00401 7.3
Druze -0.00131 -3.2 Piapoco 0.00396 7.8
Tunisian_Jew -0.00130 -3.1 Cabecar 0.00390 7.6
Lebanese -0.00125 -3.1 Aymara 0.00388 8.1

It seems that an unknown population, related to present-day Near Eastern populations, entered
eastern Europe at some time between the EHG and the formation of the Yamnaya.
Nonetheless, the EHG shared more alleles with the Yamnaya than with nearly all Early,
Middle, Late Neolithic and Bronze Age populations from Europe outside Russia (Table S7.6).

Arrival of Eastern European hunter-gatherer ancestry during the Late Neolithic

Both PCA (Fig. 2a) and ADMIXTURE analysis (Fig. 2b) show that the shift that occurred in
Europe between the Middle and Late Neolithic was qualitatively different from the WHG
resurgence. We can show this formally by studying the statistic f4(LBK_EN, Test; Loschbour,
Karelia_HG) which tests whether allele frequency differences between the early LBK_EN
farmers and a Test population are correlated to differences between eastern and western
European hunter-gatherers. Fig. S7.9 shows the value of this statistic for different MN/LN/BA
populations. For Middle Neolithic Europeans (bottom half), the statistic is negative,
corresponding to the WHG-related resurgence, but for Late Neolithic/Bronze Age ones (and
maximally for the Yamnaya steppe population) the statistic is positive, demonstrating the
arrival of eastern populations related to the Yamnaya and to the EHG during this period.

72
Table S7.6: Karelia_HG shares more alleles with Yamnaya than with most EN/MN/BA
Europeans outside Russia
Test f4(Test, Yamnaya, Karelia_HG, Chimp) Z
LBKT_EN -0.00811 -5.5
Stuttgart -0.00704 -13.0
HungaryGamba_EN -0.00669 -16.0
Starcevo_EN -0.00639 -7.0
HungaryGamba_CA -0.00607 -9.0
LBK_EN -0.00565 -14.7
Spain_EN -0.00538 -12.1
Baalberge_MN -0.00470 -7.1
Iceman -0.00466 -8.0
SwedenSkoglund_MN -0.00440 -6.7
Spain_MN -0.00361 -8.1
HungaryGamba_BA -0.00328 -6.4
Esperstedt_MN -0.00282 -4.3
Halberstadt_LBA -0.00186 -3.2
Bell_Beaker_LN -0.00181 -4.7
BenzigerodeHeimburg_LN -0.00176 -3.6
Unetice_EBA -0.00161 -4.5
Alberstedt_LN -0.00152 -2.6
Karsdorf_LN -0.00060 -0.6
Corded_Ware_LN -0.00001 0.0

Figure S7.9: The statistic f4(LBK_EN, Test; Loschbour, Karelia_HG) demonstrates the
“eastern” shift of Late Neolithic/Bronze Age Europeans. The estimated value of the
statistic and ±3 standard errors are shown.

We can also show that the change between the Middle and Late Neolithic is related to eastern
Europe by the statistics f4(MN, LNBA; EHG, WHG) = -0.0032 (Z=-16.4) , and f4(MN,
LNBA; EHG, SHG) = -0.00189 (Z=-11.2), which show that allele frequency changes of
populations between these two periods show an affinity to the EHG in the LNBA.

73
The Corded Ware are intermediate between the Yamnaya and the Middle Neolithic
To determine that the Corded Ware (the earliest Late Neolithic population from Germany,
and the one which maximally expresses the “eastern” shift in Fig. S7. 9) differ from Middle
Neolithic Europeans and the Yamnaya that temporally precede them in central and eastern
Europe respectively, we study the statistics f4(Corded_Ware, Esperstedt_MN, Test, Chimp),
and f4(Corded_Ware, Yamnaya; Test, Chimp) (Fig. S7.10). The first of these statistics shows
that in relation to the Middle Neolithic population of Germany, the Corded Ware share more
alleles with eastern populations, including EHG, MA1, Yamnaya, and Native Americans. The
second shows that in relation to the Yamnaya, the Corded Ware share more alleles with
Middle and Early Neolithic European populations. Thus, these statistics paint a similar picture
as the PCA analysis (Fig. 2a) which places the Corded Ware in an intermediate position
between the Yamnaya and Middle Neolithic Europeans.

Figure S7.10: The Corded Ware are intermediate between the Yamnaya and the Middle
Neolithic Europeans as assessed by the statistics f4(Corded_Ware, Esperstedt_MN, Test,
Chimp) (panel a), and f4(Corded_Ware, Yamnaya; Test, Chimp) (panel b). The
estimated value of the statistic and ±3 standard errors are shown.

(a) (b)

Differences between Late Neolithic/Bronze Age groups

While the Late Neolithic/Bronze Age groups appear to differ from the Middle Neolithic inhabitants of
Europe in the same way (having more ancestry related to the Eastern European hunter-gatherers,
MA1, and Yamnaya), they are not completely homogeneous. We studied differences between them
using the statistic f4(X, Y; Test, Chimp), allowing (X, Y) to be a pair of LN/BA populations, and Test to
be Loschbour, Karelia_HG, Yamnaya, and LBK_EN; by doing so, we wanted to examine if these
populations differed from each other with respect to WHG, EHG, Yamnaya, and EN populations of
Europe. The significant statistics (|Z|>3) are shown in Table S7.7 and show that the Corded_Ware_LN
had more ancestry related to eastern groups Karelia_HG and Yamnaya than any of the other LN/BA
Europeans. It appears that the “eastern” ancestry introduced into Central Europe via the Corded Ware
was later diluted in this region. This finding is inconsistent with a process of continuous migration
from the east during the LN/BA period, as that would imply that temporally later populations would
share more alleles with eastern populations as a result of having experienced more gene flow.

74
Table S7.7: Significant (|Z|>3) statistics of the form f4(X, Y; Test, Chimp) for (X, Y) a pair of
LN/BA populations, and Test either Loschbour, Karelia_HG, Yamnaya, or LBK_EN.
X Y Test Z X
Alberstedt_LN HungaryGamba_BA Yamnaya 0.00161 3.1
Bell_Beaker_LN HungaryGamba_BA Yamnaya 0.00121 3.0
Corded_Ware_LN Bell_Beaker_LN Karelia_HG 0.00182 4.4
Corded_Ware_LN Unetice_EBA Karelia_HG 0.00162 4.0
Corded_Ware_LN Halberstadt_LBA Karelia_HG 0.00187 3.2
Corded_Ware_LN HungaryGamba_BA Karelia_HG 0.00327 6.2
Corded_Ware_LN Bell_Beaker_LN Yamnaya 0.00162 5.1
Corded_Ware_LN BenzigerodeHeimburg_LN Yamnaya 0.00144 3.6
Corded_Ware_LN Unetice_EBA Yamnaya 0.00151 4.9
Corded_Ware_LN Halberstadt_LBA Yamnaya 0.00165 3.7
Corded_Ware_LN HungaryGamba_BA Yamnaya 0.00286 7.3
Unetice_EBA HungaryGamba_BA Karelia_HG 0.00161 3.2
Unetice_EBA HungaryGamba_BA Yamnaya 0.00129 3.4

Differences between the Corded Ware and present-day Europeans

We also studied differences between the Corded Ware and present-day Europeans using
statistics of the form f4(European, Corded_Ware_LN; Other, Chimp), with Other chosen from
the list: LBK_EN, Loschbour, Karelia_HG, Yamnaya. These statistics are plotted in Fig. S7.
11, and show that both the EHG and the Yamnaya share more alleles with the Corded Ware
than with any present-day European population. This is expected in the case of southern
Europeans (as the Corded Ware horizon was a central/northern European phenomenon, and
one might not expect present-day southern Europeans to form a clade with the Corded Ware
population), but we find that it is also true for all present-day northern Europeans as well.
This suggests that the ancestry introduced into Europe from the steppe during the Late
Neolithic was later diluted, a process that had already begun during the Late Neolithic period
itself (Table S7.7). This dilution may have involved the pre-existing farming population of
Europe, but in parts of Europe may have included populations with substantial hunter-
gatherer ancestry, as indicated by the fact that the statistic f4(European, Corded_Ware_LN;
Loschbour, Chimp) is significantly positive for some European populations such as
Lithuanians, Estonians, and Icelanders (Fig S7.11c).

Evidence of admixture with f3-statistics

The f4-statistics examined so far do not resolve the direction of gene flow unambiguously. For
example, negative f4(EN, MN; WHG, Chimp) could indicate gene flow into MN from WHG
or into WHG from MN. Likewise, negative f4(MN, LNBA; EHG, Chimp) may be interpreted as
either gene flow into LNBA from EHG or into EHG from LNBA. Our interpretation favoring
Middle Neolithic WHG “resurgence” and Late Neolithic EHG “arrival” is based in part on the
chronological order of the samples, as it seems unlikely that the MN and LN/BA populations
of Europe would represent lineages that admixed into the earlier WHG and EHG than vice
versa. However, we can also prove that our interpretation is correct by examining statistics of
the form f3(Test; X, Y) , using the inbreed: YES parameter of the qp3Pop program of
ADMIXTOOLS1. Post-admixture genetic drift may mask a negative signal of admixture1 and
application of f3-statistics on pseudo-haploid data must account for the pseudo-inbreeding
induced by the complete homozygosity of the Test population individuals in a manner
analogous to the real inbreeding present in some modern human populations 2. Significantly
negative statistics demonstrating the three main events proposed in our paper (WHG
resurgence during the MN, EHG dilution in the Yamnaya, and MN/Yamnaya admixture in the

75
formation of LN/BA populations) are listed in Table S7. 8. These statistics can only be
estimated when Test has at least 2 individuals.

Figure S7. 11: Present-day Europeans do not form a clade with the Corded Ware.
(a) f4(European, Corded_Ware_LN; Yamnaya, Chimp) (b) f4(European, Corded_Ware_LN; Karelia_HG, Chimp)

(c) f4(European, Corded_Ware_LN; Loschbour, Chimp) (d) f4(European, Corded_Ware_LN; LBK_EN, Chimp)

Conclusions
We can discern three different groups of hunter-gatherers who lived in Europe before the
arrival of the first farmers: western European hunter-gatherers (WHG) in Spain, Luxembourg,
and Hungary; eastern European hunter-gatherers (EHG) in Russia, and Scandinavian hunter-
gatherers (SHG) in Sweden. We can show that the early farmers of Europe descended from a
common ancestral population. However, by the Middle Neolithic period they had experienced
a resurgence of ancestry related to the WHG. In the Russian steppe, where farming did not get
established, the Yamnaya pastoralists emerged as a mixture of the EHG and a Near Eastern
population. Thus, while hunter-gatherer ancestry was increasing in western Europe, it was
decreasing in eastern Europe. The two regions of Europe came into contact during the Late
Neolithic period, when, beginning with the Corded Ware, migrants from the east, related to
the EHG or the Yamnaya, arrived in central Europe. This does not seem to have been a
process of slow infiltration of eastern populations over a prolonged period of time, as the

76
earliest population (Corded Ware) was more closely related to the eastern groups than all the
later ones. Other Late Neolithic and Bronze Age populations, as well as all present-day
Europeans, have less ancestry related to these eastern migrants than the Corded Ware
population did. In SI9 we quantify the magnitude of these population transformations by
estimating mixture proportions in the affected populations.

Table S7.8: Confirmation of main inferences using f3-statistics.

Test X Y f3(Test; X, Y) Z
Spain_MN Spain_EN Loschbour -0.01001 -5.2
Baalberge_MN LBK_EN Loschbour -0.01129 -2.8
Yamnaya Karelia_HG Armenian -0.00765 -6.3
Corded_Ware_LN Esperstedt_MN Yamnaya -0.00791 -4.0
Bell_Beaker_LN Esperstedt_MN Yamnaya -0.00815 -5.0
BenzigerodeHeimburg_LN Esperstedt_MN Yamnaya -0.00829 -3.2
Unetice_EBA Esperstedt_MN Yamnaya -0.01055 -7.2
Corded_Ware_LN Baalberge_MN Yamnaya -0.00394 -1.8
Bell_Beaker_LN Baalberge_MN Yamnaya -0.00689 -4.2
BenzigerodeHeimburg_LN Baalberge_MN Yamnaya -0.00747 -3.0
Unetice_EBA Baalberge_MN Yamnaya -0.01139 -8.0
HungaryGamba_BA Yamnaya HungaryGamba_CA -0.01088 -3.3

References
1. Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065-1093, (2012).
2. Reich, D., Thangaraj, K., Patterson, N., Price, A. L. & Singh, L. Reconstructing Indian
population history. Nature 461, 489-494, (2009).
3. Patterson, N., Price, A. L. & Reich, D. Population structure and eigenanalysis. PLoS Genet. 2,
e190, (2006).
4. Price, A. L. et al. Principal components analysis corrects for stratification in genome-wide
association studies. Nat. Genet. 38, 904-909, (2006).
5. Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry in
unrelated individuals. Genome Res. 19, 1655-1664, (2009).
6. Busing, F. T. A., Meijer, E. & Leeden, R. Delete-m Jackknife for Unequal m. Statistics and
Computing 9, 3-8, (1999).
7. Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of Native
Americans. Nature 505, 87-91, (2014).
8. Skoglund, P. et al. Genomic Diversity and Admixture Differs for Stone-Age Scandinavian
Foragers and Farmers. Science 344, 747-750, (2014).
9. Prufer, K. et al. The complete genome sequence of a Neanderthal from the Altai Mountains.
Nature 505, 43-49, (2014).
10. Seguin-Orlando, A. et al. Genomic structure in Europeans dating back at least 36,200 years.
Science 346, 1113-1118, (2014).
11. Fu, Q. et al. Genome sequence of a 45,000-year-old modern human from western Siberia.
Nature 514, 445-449, (2014).
12. Gamba, C. et al. Genome flux and stasis in a five millennium transect of European prehistory.
Nat. Commun. 5, 5257 (2014).
13. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).
14. Bollongino, R. et al. 2000 Years of Parallel Societies in Stone Age Central Europe. Science
342, 479-481, (2013).

77
Supplementary Information 8
Phylogenetic relationships of ancient Eurasians

Iosif Lazaridis*, Nick Patterson, and David Reich

* To whom correspondence should be addressed ([email protected])

Overview
In this section we model the phylogenetic relationships of Eurasian hunter-gatherers using the
ADMIXTUREGRAPH1,2 software that allows one to propose an admixture graph model of
history (a tree augmented with admixture edges), fit model parameters (genetic drift edges
and admixture proportions), and report the difference between estimated and modeled f-
statistics as the number of standard errors (Z-score) using a block jackknife1,3.

The goals of this note are two-fold:

(1) We show that the model we previously developed for the deep history of Eurasian
populations4 fits the new data reported in this study.
We do not simply add the newly available samples to the previous model and test if they
fit. Instead, we start afresh, exploring the space of possible models more fully by (i)
exhaustively iterating tree topologies for a set of populations given a specified number of
admixture events, and (ii) when adding a new population to an existing admixture graph,
iterating over all possible placements of this population as either a simple branch or a 2-
way mixture. By doing so, we mitigate against the subjectivity of specifying an admixture
graph “by hand”, or relying on an approach such as TreeMix5 that does not search the
space of models exhaustively and may overlook topologies that fit the data.

(2) We show that several very different models of population relationships can fit the data
once we add Eastern Hunter Gatherers to the modeling.
When we add Karelia_HG (representing EHG), we are able to obtain fits to the data when
we allow an additional admixture event. However, several very different models of
history can be made to fit and are equally parsimonious in the sense of specifying the
same number of admixture events. Because we do not want our inferences about mixture
to be biased by arbitrary choices about a phylogenetic model relating all the populations,
we developed new techniques, described in detail in SI9 and SI10, which are able to infer
mixture proportions without precisely modeling phylogenetic relationships.

Point #2 is the key motivation for this note. While we could in theory have tried alternative
model-fitting approaches like TreeMix5, we did not attempt to do so because these fit the
same underlying data as our ADMIXTUREGRAPH analysis. We know from this note that
several very different models equally parsimoniously fit the data. Thus, even if TreeMix
found additional models that fit, it would not change our conclusion that the whole exercise of
model building based on allele frequency correlation statistics does not produce a unique
solution. In other words, the main point of the model-building in this note is to motivate the
development of methods, shown in SI9 and SI10, that work robustly regardless of the true
underlying model.

78
Details
We begin by fitting the basic model4 to the new data, substituting the Stuttgart Early
European Farmer (EEF), which belonged to the Linearbandkeramik archaeological culture
with the LBK_EN group (n=12). This model is fitted successfully (Fig. S8.1), with no f-
statistics differing between estimated and modeled values by |Z|>3.

Figure S8.1: Basic model from ref.4 applied to data of this study.

In ref.4, the amount of “Basal Eurasian” admixture in the Stuttgart farmer was also estimated using
the f 4 -ratio f 4 (Stuttgart, Loschbour; Onge, MA1) / f 4 (Mbuti, MA1; Onge, Loschbour) as 44r10%.
When we estimate this parameter using LBK_EN instead of Stuttgart and obtain an estimate of
40.0r9.3%. Some basal Eurasian admixture into LBK_EN is necessary, as shown by the fact that
when we force 0% “Basal Eurasian” admixture into LBK_EN, the model does not fit, with the worst
discrepancy being for the statistic f 3 (LBK_EN; Onge, MA1) whose fitted value is lower than the
estimated one by Z=11.3. A model without Basal Eurasian admixture into LBK_EN seriously
underestimates the shared genetic drift between MA1 and Onge; by introducing “Basal Eurasian”
ancestry into LBK_EN the fact that eastern non-Africans share more alleles with European hunter-
gatherers and MA1 is taken into account4.

Fitting Ust-Ishim
Existence of “Basal Eurasian” admixture into Europeans was further supported by a study of the Ust’-
Ishim6 Upper Paleolithic Siberian (~45 thousand years ago). Ust’-Ishim was inferred to occupy a
basal position to eastern non-African and European hunter-gatherer populations but not to modern
Europeans, which was interpreted as due to admixture from a population occupying an even more
basal position to Europeans, as proposed in ref. 4. We attempted to fit Ust-Ishim anywhere on the
graph of Fig. S8.1, but all phylogenetic placements failed (|Z|>3.8) except the placement of Ust-Ishim
RQWKHQRQB$IULFDQĺ;HGJHZKLFKLVVKDUHGE\DOO(XUDVLDQVRWKHUWKDQ%DVDO(XUDVLDQV)LJS8.2).

79
Figure S8.2: Ust-Ishim is basal to all Eurasians except “Basal Eurasians” in the sense of ref.4

Thus, the split between “Basal” and other Eurasians must have taken place no later than
~45,000 years ago (the age of the Ust-Ishim individual).

Fitting Kostenki14
A study of the Kostenki14 Upper Paleolithic individual from Russia7 (K14) proposed that this
individual already had Basal Eurasian ancestry ~36,000 years ago, almost 30,000 years before
the appearance of this type of ancestry in Europe with the early farmers4. This suggestion was
based on three lines of evidence (Fig. 2 of ref. 7):
1. D(Mbuti, East Asia; Mesolithic European Hunter Gatherer, K14) < 0. We confirm this
asymmetry in our data, as D(Mbuti, Han; Loschbour, Kostenki14) = -0.0177 (Z=-3.7).
2. D(Mbuti, East Asia; Early European Farmer, K14) = 0. We confirm this symmetry in our
data as D(Mbuti, Han; LBK_EN, Kostenki14) = 0.0034 (Z=0.8).
3. Presence of “Middle Eastern” admixture in ADMIXTURE analysis of this individual. We
also confirm this point in our ADMIXTURE analysis (SI6). However, we note that
interpretation of this finding is not straightforward, as old samples such as Ust_Ishim and
Kostenki14 may predate the differentiation of populations, and the fact that they are seen
as admixed from multiple sources in ADMIXTURE analysis (SI6), does not indicate that
they are mixtures of these much later populations.

Arguments #1, 2 do not suffice to draw the conclusion that Kostenki14 had the same Basal
Eurasian ancestry as Neolithic European farmers. To illustrate this point, we use the model of
Fig. S8.2 as a baseline. First, we attempt to fit Kostenki14 into this model as either a simple
branch or a 2-way mixture.

80
No phylogenetic placement of Kostenki14 as a simple branch fits the data (|Z|>4.6). However,
we could fit Kostenki14 as a 2-way mixture (Fig. S8.3-5). All successful models agree that
Kostenki14 shares genetic drift with Loschbour and LBK_EN to the exclusion of non-
European populations and is “European” in this sense. However, the model of Fig. S8.3
suggests it has Basal Eurasian ancestry that split off before the split of this type of ancestry in
LBK_EN, while those of Fig. S8.4-5 are equivalent and suggest that its Basal Eurasian
ancestry is the same as that found in LBK_EN and there is a trifurcation of the Basal Eurasian
ancestry for the three lineages {LBK_EN, Kostenki14, main Eurasian ancestry node “A”}.

The hypothesis of Basal Eurasian ancestry in Kostenki14 needs to be further tested, as the
negative D(Mbuti, Han; Loschbour, Kostenki14) statistic could also reflect gene flow
EHWZHHQ+DQļ/RVFKERXUWKLVLVa priori plausible, as these populations are much younger
than Kostenki14 and may share intra-Eurasian genetic drift that Kostenki14 lacks because of
its age. The possibility of later gene flow between Europeans and eastern non-Africans must
be further tested with additional ancient samples from Upper Paleolithic Europe and Asia.

Figure S8.3 Kostenki14 as a 2-way mixture with basal Eurasian admixture before that of
the early farmers

81
Figure S8.4 Kostenki14 as a 2-way mixture with basal Eurasian admixture like that of
the first farmers (however, topologically equivalent to Fig. S8.5 due to 0 drift in
QRQB$IULFDQĺ&EUDQFK

82
Figure S8.5 Kostenki14 as a 2-way mixture with basal Eurasian admixture after that of the first
farmers (but topologically equivalent to Fig. S8GXHWRGULIWLQQRQB$IULFDQĺ&EUDQFK

83
Figure S8.6: Karelia_HG as 2-way mixture of MA1 and EHG (the model fails
marginally with |Z|=3.1)

Eastern European hunter-gatherers (EHG) as a WHG-ANE mixture

A population that had not been previously studied is the Eastern European hunter-gatherers
(EHG) from European Russia, so we attempt to fit Karelia_HG (the best sample from the
EHG) to the model of Fig. S8.1. No models could accommodate Karelia_HG as a simple (not
admixed) population (|Z|>9), or a 2-way mixture (|Z|>3). However, three models failed only
marginally (3.1<|Z|<3.7) and we show one of them in Fig. S8.6. The model of Fig. S8.6
derives the Western Eurasian-related ancestry in Karelia_HG from a node D related to the
population that admixed into the LBK_EN. The other two models (not shown) differ only
slightly, by deriving the admixture into Karelia_HG either from the Loschbour branch, or a
node that is basal to the Western_Eurasian node. Given the number of f-statistics computed, a
largest |Z|-score of 3.1-3.7 does not provide very strong rejection of the model

84
All three of these 2-way mixture models arrive at a similar inference of 38-40% ANE and 60-
62% WHG ancestry in Karelia-HG, which is higher than the ~20% ANE ancestry inferred for
Motala12 in a previous study4, consistent with the fact that EHG are the population sharing
more alleles with “Ancient North Eurasians” (SI 7) than any other. This is also consistent
with a previous hypothesis4 of a “Hunter” population with 60-80% WHG/(WHG+ANE)
ancestry that combined with early European farmers (EEF) to produce (in different
proportions) present-day Europeans. The estimate 60% (Fig. S8.6) is at the low end of the
inferred 60-80% range proposed in this previous study4. However, the present study reveals
the Early European Farmers experienced a “resurgence” of WHG-related ancestry during the
Middle Neolithic (SI 9), and thus, the WHG-related ancestry in present-day Europeans could
come both from that event (that is, a Middle Neolithic population with a WHG/(WHG+ANE)
ratio of 100%), and an eastern population bearing EHG-related ancestry (with a
WHG/(WHG+ANE) ratio of 60%). The sum of these two processes may explain the 60-80%
range of this ratio of the previous study4.

A different detail of Fig. S8.6 is that the ANE ancestry in Karelia_HG is derived from the
branch of “Ancient_North_Eurasian” that goes into the Karitiana Native Americans, rather
than the MA1 branch. This is plausible, given the fact that modern humans arrived in the
Americas after the time of MA1 (~24,000 years ago), giving added opportunity for more
genetic drift to accumulate in the “Ancient North Eurasian” population; this could be shared
by the EHG hunter-gatherers of European Russia from the Holocene period.

Eastern European hunter-gatherers as an unadmixed population

Fig. S8.6 shows Karelia_HG as a 2-way mixture of ANE and WHG, but this may be due to
the fact that the model of Fig. S8.1 was developed before the EHG were sampled. We do not
want to reify ANE and WHG simply on the basis of their having been described earlier in the
literature, so we also explored models in which EHG are not admixed. This was driven, in
part, by our observation that the statistic f 4 (Karelia_HG, MA1; Karitiana, Chimp) = -0.00014
(Z=-0.2) which appears to be show a perfect symmetry in the relationship of (Karelia_HG,
MA1) with Native Americans. Fig. S8.6 accounts for this “symmetry” between MA1 and
Karelia_HG with respect to Native Americans by proposing a balancing of two processes:
first, Karelia_HG share more alleles with Native Americans due to sharing additional
common genetic drift (and thus Native Americans should share more alleles with Karelia_HG
than with MA1), but, second, Karelia_HG has WHG-related ancestry which dilutes this
affinity to Native Americans. However, this requires that the additional common drift shared
by Karelia_HG and Native Americans to be nearly perfectly balanced with the dilution due to
WHG ancestry, resulting in the observed symmetry, which is not very parsimonious.

To study the relationship between ANE, EHG, and WHG without necessarily accepting the
model of Fig. S8.1, we attempted to fit MA1, Karelia_HG, and Loschbour as a simple tree
using Mbuti as an outgroup. There are 3 possible arrangements of this kind (depending on
whether MA1, Karelia_HG, or Loschbour is chosen as an outgroup to the two others), but
none fit successfully (|Z|>6), suggesting that at least one of the 3 populations is admixed. We
have already explored the possibility that MA1 and Loschbour are unadmixed (Fig. S8.1)
which leads to the inference that EHG are admixed (Fig. S8.6), so we only examine the two
remaining possibilities: (i) that MA1 is admixed and Loschbour, Karelia_HG are unadmixed,
and (ii) that Loschbour is admixed and MA1, Karelia_HG are unadmixed. These two models
fit successfully, and are shown in Fig. S8.7.

85
Figure S8.7: Alternative phylogenies that fit the data in which (a) MA1 is admixed, or
(b) Loschbour is admixed.

(a) (b)

We next added new populations to the models of Fig. S8.7. First, we added the Onge8, as an isolated
eastern non-African population that could constrain these models (Fig. S8.8), and obtained good fits.

Next, we added the Karitiana Native Americans, to account for the different relationship between
Native Americans to west Eurasian groups, as they share more alleles with MA1 and EHG than with
Loschbour. The Karitiana cannot be added as an unadmixed population to the models of Fig. S8.5
(|Z|>8). They can, however, be modeled as a 2-way mixture, with successful (|Z|<3) models shown in
Fig. S8.9. Thus, for every model we analyzed, our results support ref. 9 that Native Americans are
anciently admixed.

86
Figure S8.8: Adding Onge as an eastern non-African constraint to the models of Fig. S8.4

In Fig. S8.9a (when MA1 is admixed), Native Americans shown to be a mixture of a population J
related to the Onge and a different population K related to MA1 and Karelia_HG. Note that this
model accounts for the symmetry in the relationship of Native Americans to MA1 and EHG in a
different way than that of Fig. S8.6. In Fig. S8.9a, Native American ancestry is more closely related to
MA1 than to EHG, however, it is MA1 that derives part of its ancestry from an early node C which
“dilutes” its affinity to Native Americans, while Karelia_HG shares the genetic drift on the path
FĺAĺB fully.

Fig. S8.9b and Fig. S8.9c present two different solutions (when Loschbour is admixed). Both (b) and
(c) agree with (a) in deriving ancestry of the Karitiana from both eastern non-Africans and
populations related to west Eurasians4,9,10. Unlike the models of Fig. S8.9a and Fig. S8.6, however,
they do not resolve the symmetric relationship of Native Americans to MA1 and Karelia_HG by
postulating that Native Americans share additional common genetic drift with MA1 or Karelia_HG,
but this ancestry is diluted in MA1 and Karelia_HG, resulting in the observed symmetry. Rather, Fig

87
S8.9b and Fig. S8.9c propose a trifurcation of Ancestral North Eurasians, whereby the ancestry that
goes into Native Americans is equally related to both MA1 and Karelia_HG, both of which are
unadmixed. (Note the 0 length edges AĺG and GĺA in the models of Fig. S8.6b and Fig. S8.6c,
reflecting this trifurcation).

Figure S8.9: Adding Karitiana as a constraint for ancient North Eurasian-related populations

(a) (b) (c)

We finally added LBK_EN to the population models of Fig. S8.9. This could only be fit as a 2-way
mixture that included a “Basal Eurasian” component, with the successful models shown in Fig. S8.10-
11. The remainder of the ancestry of LBK_EN is inferred to be a sister group of Loschbour (Figs.
S8.10, 11b, 12b) or a sister group of a component of Loschbour (Figs. S8.11a, 12a).

This exploration of possible phylogenies has revealed that seven models with three admixture events
are consistent (|Z|<3) or marginally inconsistent (3<|Z|<4) with the studied set of populations (Fig.
S8.6, 10, 11, 12). The models of Fig. S8.6 have an advantage in postulating that EHG are a mixture of
populations related to MA1 and Loschbour, which are actual individuals, while those of Figs S8.10,
11, 12 propose that MA1 or Loschbour are admixed populations, although at present there are no
actual individuals that represent some of the admixing populations. Nonetheless, all these models
agree on several points: the dual origin of Native Americans9, the existence of Basal Eurasian ancestry
in Early Neolithic Europeans4, the fact that MA1 shares more alleles with Karelia_HG than with
Loschbour, but MA1 and Karelia_HG are symmetrically related to Native Americans, and, finally, the
fact that the three group of Eurasian hunter-gatherers (EHG, WHG, and ANE) cannot be related to
each other by a simple tree, and at least one of them must be admixed.

Uncertainty about the modeling of ancient Europeans may be a consequence of two factors.

First, it is possible that these populations were related in complex ways with gene flow between them.
We show that at least one admixture event is required to jointly fit (EHG, WHG, and ANE). Models

88
with more admixture events or bidirectional gene flow are more complex and thus less parsimonious;
however, it is possible that they better capture the actual history of these populations.

Second, the sparse sampling of pre-Neolithic Europeans makes inferences about their relationships
problematic. With the exception of Holocene hunter-gatherers from Sweden where multiple
individuals are now known (ref. 4,10 and this study), a total of nine individuals of likely hunter-
gatherer ancestry have autosomal data that is amenable to analysis (Samara_HG and Karelia_HG in
this study, Loschbour4, LaBrana11, KO1 (HungaryGamba_HG)12, Kostenki147, MA19, Ust-Ishim6, and
Tianyuan13). Tens of thousands of years since the earliest settlement of Europe by anatomically
modern humans14 and until the Holocene remain to be filled with data points that will doubtless
inform our understanding of the distant past of European populations.

In the current paper, we do not address these issues, except to show, as we did above, the most salient
points shared by some successful models, and to reject others. Fortunately, for the much shorter
period since the Neolithic transition, there is an abundance of data from different parts of Europe and
different points in time, allowing us to reconstruct some key transitions without explicitly modeling
the deep phylogenetic relationships of Eurasian populations (SI9).

Figure S8.10: Adding LBK_EN to the Fig. S8.9a model (MA1 admixed; this weakly fails |Z|=3.2)

89
Figure S8.11: Adding LBK_EN to the models of Fig. S8.9b (Loschbour is admixed)

(a) (b)

90
Figure S8.12: Adding LBK_EN to the models of Fig. S8.9c (Loschbour is admixed)

(a) (b)

91
References

1. Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065-1093, (2012).
2. Reich, D., Thangaraj, K., Patterson, N., Price, A. L. & Singh, L. Reconstructing Indian
population history. Nature 461, 489-494, (2009).
3. Busing, F. T. A., Meijer, E. & Leeden, R. Delete-m Jackknife for Unequal m. Statistics and
Computing 9, 3-8, (1999).
4. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).
5. Pickrell, J. K. & Pritchard, J. K. Inference of population splits and mixtures from genome-
wide Allele frequency data. PLoS Genet. 8, e1002967, (2012).
6. Fu, Q. et al. Genome sequence of a 45,000-year-old modern human from western Siberia.
Nature 514, 445-449, (2014).
7. Seguin-Orlando, A. et al. Genomic structure in Europeans dating back at least 36,200 years.
Science 346, 1113-1118, (2014).
8. Thangaraj, K. et al. Reconstructing the origin of Andaman Islanders. Science 308, 996-996,
(2005).
9. Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of Native
Americans. Nature 505, 87-91, (2014).
10. Skoglund, P. et al. Genomic Diversity and Admixture Differs for Stone-Age Scandinavian
Foragers and Farmers. Science 344, 747-750, (2014).
11. Olalde, I. et al. Derived immune and ancestral pigmentation alleles in a 7,000-year-old
Mesolithic European. Nature 507, 225-228, (2014).
12. Gamba, C. et al. Genome flux and stasis in a five millennium transect of European prehistory.
Nat. Commun. 5, 5257 (2014).
13. Fu, Q. et al. DNA analysis of an early modern human from Tianyuan Cave, China. Proc.
Natl. Acad. Sci. USA 110, 2223–2227, (2013).
14. Higham, T. et al. The earliest evidence for anatomically modern humans in northwestern
Europe. Nature 479, 521-524, (2011).

92
Supplementary Information 9
Inference of admixture proportions without detailed phylogenetic modeling

Iosif Lazaridis*, Nick Patterson, and David Reich

* To whom correspondence should be addressed ([email protected])

In this section, we develop a method that can be used (i) to identify reference populations that may
have contributed ancestry to a Test population, and (ii) to estimate mixture proportions from these
reference populations for Test. The method uses the intuition that the reference populations are not
identically related to a panel of “focal” or “outgroup” populations, but share different amounts of
genetic drift with them as a result of their deep evolutionary history (which is, however, not explicitly
modeled). These “outgroups” must be devoid of recent gene flow with either the Test or the candidate
reference population, as such gene flow introduces additional common genetic drift. One way to
identify them is to pick a varied set of world populations that are (i) geographically remote from the
area under study, and (ii) do not show evidence of admixture from that area using an algorithm such
as ADMIXTURE1 that can identify recently admixed populations and individuals. In practice, we will
use the following set ࣩ of 15 previously identified2 outgroups for West Eurasia:

“World Foci 15” set of outgroups ࣩ: Ami, Biaka, Bougainville, Chukchi, Eskimo, Han,
Ju_hoan_North, Karitiana, Kharia, Mbuti, Onge, Papuan, She, Ulchi, Yoruba

Figure S9.1: If Test and Ref1 form a clade with respect to the outgroups, then statistics involving
either one of them and the outgroups are identical.

Admixture proportion estimation with non-negative least squares

If Test forms a clade with a reference population Ref1 since their divergence from a triple of outgroups
A, B, C (Fig. S9.1) we expect that f4(Test, A; B, C) ≈ f4(Ref1, A; B, C), f4(Test, B; A, C) ≈ f4(Ref1, B; A,
C), f4(Test, C; A, B) ≈ f4(Ref1, C; A, B). However, if Test also has ancestry related to another
population Ref2 then it will be the case that f4(Test, A; B, C) ≈ αf4(Ref1, A; B, C)+(1-α)f4(Ref2, A; B, C)
where α is the fraction of ancestry from Ref1 and 1-α from Ref2 (Fig. S9.2). More generally, if Test has
ancestry related to N different populations, then we can write:

݂ସ ሺܶ݁‫ݐݏ‬Ǣ ‫ܣ‬Ǣ ‫ܤ‬ǡ ‫ܥ‬ሻ ൎ σே

௜ୀଵ ߙ௜ ݂ସ ሺܴ݂݁௜ Ǣ ‫ܣ‬Ǣ ‫ܤ‬ǡ ‫ܥ‬ሻ (S9.1)
௡ିଵ
with σே௜ୀଵ ߙ௜ ൌ ͳ and ߙ௜ ൒ Ͳǡ ݅ ൌ ͳǡ ǥ ǡ ܰ. Given a set of n outgroups, there are ݊൫ ଶ ൯ arrangements
of the outgroups, as each of them can take the position of A in the above expression, and all pairs of
the remaining n-1 ones could take the position of B, C. For the set of 15 outgroups there are 1,365

93
different arrangements, and hence 1,365 equations of the above form. We can then write t≈RαT where
t is a column vector of 1,365 f4-statistics involving the Test population, R is a 1,365-by-N matrix of f4-
statistics (each column for one of the N references) and α is a column vector of N mixture proportions
(to be estimated). This problem can be solved for α by least squares, minimizing ԡ‫ ݐ‬െ ܴܽ୘ ԡଶଶ subject
to the constraint that α has non-negative elements and ԡܽԡଵ ൌ ͳ. We use the implementation of least
squares in the lsqlin function of Matlab (http://www.mathworks.com/help/optim/ug/lsqlin.html) which
estimates the vector ߙො of the mixture proportions and also reports resnorm = ԡ‫ ݐ‬െ ܴߙො ୘ ԡଶଶ , the squared
2-norm of the residuals.

Figure S9.2: Test is a mixture related to two reference populations Ref1 and Ref2. Three possible
topologies are shown: (a) Ref1 and Ref2 are symmetrically related to A, B, C. (b) Ref1 and Ref2 are
differentially related to one outgroup A, and symmetrically related to the other two B, C. (c) Ref1 and
Ref2 are related to two of the outgroups A, C and symmetrically related to the third B. In case (a) all
f4-statistics involving Test, Ref1, Ref2 are identical and cannot be used to estimate mixture proportions.
In cases (b, c) f4-statistics involving Test are a weighted average of those involving Ref1 and Ref2.

The advantage of this method is that the relationship of Ref1, Ref2, and A, B, C need not be explicitly
modeled. Each of them may have a history of admixture in its evolutionary past that is arbitrarily
complex, but for each portion of its ancestry one of the topological relationships of Fig. S9.2 will hold
and thus the f4-statistics involving Test will be the weighted average of the reference populations.

94
Figure S9.3: LBK_EN and Spain_EN appear to be symmetrically related to the outgroups, but
there is some variation between LBK_EN and Karelia_HG which may be used to infer
admixture estimates for populations having ancestry from sources related to them.

This method works if the reference populations are not all identically related to the outgroups, but
does not work if they are identically related to the outgroups. Concretely, if the outgroups are archaic
humans and primates, or alternatively Loschbour and LaBrana2, they are likely to be identically
related (from the perspective of shared genetic drift) to the references populations. Thus there will be
no leverage to discern whether test samples are more closely related to Ref1 or Ref2.

It is useful to plot f4(Ref1, A; B, C) and f4(Ref2, A; B, C) against each other (Fig. S9.3). Two closely
related populations (e.g., LBK_EN and Spain_EN) have a virtually identical relationship to the
outgroups, while two other populations (e.g., LBK_EN and Karelia_HG) have a visibly differentiated
relationship. This difference is not huge, as LBK_EN and Karelia_HG are West Eurasian populations
that are mostly symmetrically related to the outgroups. However, it is sufficient to infer ancestry
proportions for populations of mixed LBK_EN and Karelia_HG-related ancestry, as we will show.

We do not assume that particular reference populations contributed ancestry to a Test population. We
also do not assume that a Test population is admixed or that it is a mixture of a particular number N of
reference populations. Our approach is as follows: for each Test population the set of reference
populations ࣬ is a set of temporally preceding ancient populations. For the Middle Neolithic
populations these include:

“preMN”: Karelia_HG, Motala_HG, LBK_EN, Spain_EN, HungaryGamba_EN, Yamnaya,

Loschbour

For the Late Neolithic/Bronze Age populations they include:

“preLN/BA”: Karelia_HG, Motala_HG, LBK_EN, Spain_EN, Baalberge_MN, Esperstedt_MN,

Spain_MN, HungaryGamba_EN, Yamnaya, Loschbour

This list includes the best representative (higher number of SNPs) of WHG and EHG (Loschbour and
Karelia_HG), the best representative of early farmers from Spain, Germany, and Hungary (Spain_EN,

95
LBK_EN, and HungaryGamba_EN), and Middle Neolithic farmers (Spain_MN, Baalberge_MN,
Esperstedt_MN), and Yamnaya.

Given a Test population we vary N=1, 2, 3. If N=1 then we attempt to make Test a sister group of a
preceding population; N=2 a 2-way mixture, etc. We test all ൫ȁ࣬ȁ ே
൯ combinations of reference
populations, fitting α for each one and recording its resnorm. We record the residuals:

ܴ݁‫݈ܽݑ݀݅ݏ‬ሺܶ݁‫ݐݏ‬Ǣ ‫ܣ‬ǡ ‫ܤ‬ǡ ‫ܥ‬ሻ ൌ σே ො௜ ݂ସ ሺܴ݂݁௜ ǡ ‫ܣ‬Ǣ ‫ܤ‬ǡ ‫ܥ‬ሻ െ ݂ସ ሺܶ݁‫ݐݏ‬ǡ ‫ܣ‬Ǣ ‫ܤ‬ǡ ‫ܥ‬ሻ
௜ୀଵ ܽ (S9.2)

We also record the Z-score. If our estimate of a statistic is ݂ସ ሺܶ݁‫ݐݏ‬ǡ ‫ܣ‬Ǣ ‫ܤ‬ǡ ‫ܥ‬ሻ ൌ ݉ േ ݁ then, the Z-
ଵ
score is ܴ݁‫݈ܽݑ݀݅ݏ‬ሺܶ݁‫ݐݏ‬Ǣ ‫ܣ‬ǡ ‫ܤ‬ǡ ‫ܥ‬ሻ. The maximum |Z| provides an additional measure of how well the
௘
mixture model explains all the observations.

Middle Neolithic Europeans

We have seen (SI7) that Middle Neolithic Europeans are a mixture of Early Neolithic Europeans and
western European hunter-gatherers. Our analysis of the three MN populations seems to confirm this
(Figs. S9.4-6). For N=1, the best fitting population is an Early Neolithic one, however a marked
lowering of resnorm is seen for N=2 with the best model suggesting an EN+WHG mixture, and
virtually no change for N=3. This is consistent with our hypothesis of Middle Neolithic WHG-related
resurgence in Europe. It is important to note that the chosen set of outgroups have no power to
discriminate between different Early Neolithic populations (as was clear from Fig. S9.3). So it is not
surprising that Spain_EN+Loschbour and LBK_EN+Loschbour models provide a virtually identical
fit for the different MN populations. However, EN and WHG are clearly differentially related to the
outgroups, which makes it possible to correctly infer that MN is a mixture of EN+WHG.

Late Neolithic/Bronze Age Europeans

We have seen (SI7) that Late Neolithic Europeans in central Europe (Germany) are not a continuation
of the earlier farmers of Europe but harbor ancestry from an “eastern” population (EHG or Yamnaya).
Our method seems to confirm this (Figs. S9.7-10). For none of the LN/BA populations is ancestry
from an EN/MN farmer group the best N=1 model, and in all populations a marked improvement is
observed between N=1 and N=2, but not between N=2 and N=3. The N=2 models shown in panel (b)
of Figs. S9.7-10 always show the best fit when a “western” and “eastern” population is paired. For the
Corded_Ware_LN, the best models involve European farmers and Yamnaya in proportions of
approximately ~1/5 and ~4/5 or European farmers and Karelia_HG in proportions of ~2/3 and ~1/3.
Lower “eastern” contributions are inferred for the other LN/BA populations, consistent with the
results of SI7 that demonstrate that EHG and Yamnaya share more alleles with Corded_Ware_LN
than with other LN/BA populations. Once again, the method cannot distinguish between the closely
related populations such as LBK_EN and Spain_EN, however, the key point is that it arrives at
similar estimates of turnover regardless of which population is chosen as a reference.

96
Figure S9.4: Modeling Esperstedt_MN as a mixture of earlier populations. (a) We present in the
left column a histogram of raw f4-statistic residuals and on the right Z-scores for the best-fitting
(lowest resnorm) model at each N. In the bottom row, we show on the left how resnorm and the
maximum |Z| score change for different N. (b) resnorm of different N=2 models.

(a)

(b)

97
Figure S9.5: Modeling Baalberge_MN as a mixture of earlier populations. (a) We present in the
left column a histogram of raw f4-statistic residuals and on the right Z-scores for the best-fitting
(lowest resnorm) model at each N. In the bottom row, we show on the left how resnorm and the
maximum |Z| score change for different N. (b) resnorm of different N=2 models.

(a)

(b)

98
Figure S9.6: Modeling Spain_MN as a mixture of earlier populations. (a) We present in the left
column a histogram of raw f4-statistic residuals and on the right Z-scores for the best-fitting (lowest
resnorm) model at each N. In the bottom row, we show on the left how resnorm and the maximum |Z|
score change for different N. (b) resnorm of different N=2 models.

(a)

(b)

99
Figure S9.7: Modeling Corded_Ware_LN as a mixture of earlier populations. (a) We present in
the left column a histogram of raw f4-statistic residuals and on the right Z-scores for the best-fitting
(lowest resnorm) model at each N. In the bottom row, we show on the left how resnorm and the
maximum |Z| score change for different N. (b) resnorm of different N=2 models.

(a) (b)

Figure S9.8: Modeling Bell_Beaker_LN as a mixture of earlier populations. (a) We present in the
left column a histogram of raw f4-statistic residuals and on the right Z-scores for the best-fitting
(lowest resnorm) model at each N. In the bottom row, we show on the left how resnorm and the
maximum |Z| score change for different N. (b) resnorm of different N=2 models.

(a) (b)

100
Figure S9.9: Modeling BenzigerodeHeimburg_LN as a mixture of earlier populations. (a) We
present in the left column a histogram of raw f4-statistic residuals and on the right Z-scores for the
best-fitting (lowest resnorm) model at each N. In the bottom row, we show on the left how resnorm
and the maximum |Z| score change for different N. (b) resnorm of different N=2 models.

(a) (b)

Figure S9.10: Modeling Unetice_EBA as a mixture of pre-Late Neolithic populations. (a) We

present in the left column a histogram of raw f4-statistic residuals and on the right Z-scores for the
best-fitting (lowest resnorm) model at each N. In the bottom row, we show on the left how resnorm
and the maximum |Z| score change for different N. (b) resnorm of different N=2 models.

(a) (b)

101
Figure S9.11: U-shaped curve of resnorm and max|Z| for Corded_Ware_LN as a variable 2-way
mixture of LBK_EN and Karelia_HG or Yamnaya.

U-shaped curve for an admixed population

A useful way of visualizing that a Test population is indeed formed by admixture between two others
Ref1 and Ref2 is to show the variation in resnorm and max |Z| as we model Test as a variable mixture
of (x, 1-x) fraction of ancestry from Ref1 and Ref2, sliding x between 0 and 1. This provides a useful
confirmation about the existence of admixture. We do this for Corded_Ware_LN in Fig. S9.11 as a
mixture of LBK_EN and Yamnaya or Karelia_HG (similar curves exist for the other pairings of Fig.
S9.7b . The optimal proportion corresponds well to the 18.2% LBK_EN and 81.8% Yamnaya derived
with our least squares optimizer (Fig. S9.7b). The jagged appearance of the max |Z| plot is due to the
fact that it is dominated by the single worst behaving Z-score at each value of x, whereas the resnorm
is an overall measure of fit that takes account all 1,365 f4-statistics.

Regression-based estimation of admixture proportions

For N=2 it is also possible to detect the presence of admixture and estimate its parameters using
simple linear regression. Detection of admixture in a Test population can be done using the statistic
f3(Test; Ref1, Ref2) which we have used in SI7 to prove that admixture took place in the Middle and
Late Neolithic populations under study. This statistic makes use of the fact that allele frequencies in

102
Test are intermediate between those of Ref1 and Ref2 and thus allele frequency differences Test-Ref1
and Test-Ref2 are negatively correlated. We have seen (Fig. S9.2) that statistics of the form f4(Test, A;
B, C) are intermediate between f4(Ref1, A; B, C) and f4(Ref2, A; B, C) when Test is formed by
admixture between populations related to Ref1 and Ref2 and Ref1 and Ref2 are asymmetrically related
to the outgroups. If that is true, then the difference f4(Test, A; B, C)- f4(Ref1, A; B, C) will be anti-
correlated with f4(Test, A; B, C)- f4(Ref2, A; B, C). These differences are equal to f4(Test, Ref1; B, C)
and f4(Test, Ref2; B, C). We plot them in Fig. S9.12 for Test=Corded_Ware_LN, Ref1=Karelia_HG or
Yamnaya and Ref1=Esperstedt_MN or Baalberge_MN or LBK_EN showing that they do in fact show
a pattern of negative correlation with Early/Middle Neolithic populations from Germany that precede
them. Such patterns of negative correlation also exist when using the other Neolithic populations as
references for the Corded Ware, which are genetically very similar to those from Germany, and one
should bear in mind that the admixture detected here need not have taken place in Germany itself, but
might involve other Neolithic populations between Germany and eastern Europe.

The pattern observed in Fig. S9.12 is more “crisp” when using Karelia_HG as an eastern reference,
but this does not necessarily indicate that Karelia_HG is the admixing population. As the proportion
of the eastern population increases the line becomes more vertical (notice the reduced range of the
statistic on the x-axis when Ref1=Yamnaya in all the right-side plots of Fig. S9.12), and at the limit
(for a hypothetical population X with 100% descent from the eastern population), all variation in the
statistics of the horizontal axis would be noise, as the expected value on the horizontal axis would be
zero. Another interesting case is when the two references form a clade with each other relative to the
Test population. We show an example of this in Fig. S9.13 where we model Corded_Ware_LN as a
mixture of LBK_EN and Spain_EN. A line with positive slope ≈1 is observed, also seen in Fig. S9.3
above. A positive correlation indicates that Corded_Ware_LN differs from Spain_EN and LBK_EN
in the same manner, indicating that it is not a mixture of these two populations.

103
Figure S9.12: Evidence for admixture in the Corded_Ware_LN between a Neolithic central
European population from Germany (LBK_EN, Baalberge_MN, or Esperstedt_MN) and an
eastern European population (Karelia_HG or Yamnaya). A linear regression through (0,0) is
shown.

104
Figure S9.13: When the two references are a clade with respect to the Test (as Spain_EN and
LBK_EN are with respect to Corded_Ware_LN), a line with positive slope is observed.

Plotting the f4(Test, Ref1; B, C) and f4(Test, Ref2; B, C) statistics to detect admixture has an advantage
over the use of the f3(Test; Ref1, Ref2) statistic in that these statistics are not affected by post-admixture
drift in the admixed population, but rather rely on allele frequency correlations deep in the phylogeny.
An example of this power is furnished by the case of Motala_HG which does not have a negative f3-
statistic using Loschbour and Karelia_HG as reference populations. When we test statistics of the
form f3(Motala_HG; X, Y), iterating over all pairs (X, Y) of ancient or modern populations, we obtain
the lowest statistic f3(Motala_HG; Loschbour, Karelia_HG) = 0.00404, (Z=1.8), which is, however
positive. Additional evidence for admixture in Motala_HG is provided by the statistics f4(Motala_HG,
Loschbour, Karelia_HG, Chimp) = 0.0041 (Z=7.6) and f4(Karelia_HG, Motala_HG; Loschbour,
Chimp) = -0.00696 (Z=-11.5), which show that Karelia_HG shares more alleles with Motala_HG than
with Loschbour and Loschbour shares more alleles with Motala_HG than with Karelia, which is
expected if Motala_HG is a mixture of these two populations. When we plot f4(Motala_HG,
Loschbour; B, C) and f4(Motala_HG, Karelia_HG; B, C) statistics (Fig. S9.14a, b), we observe a clear
pattern of negative correlation, which shows that allele frequencies of Motala_HG are intermediate
between those of Loschbour and Karelia_HG once we remove the post-admixture genetic drift in
Motala_HG which is uncorrelated to allele frequencies between the outgroups B, C.

105
Figure S9.14: Motala_HG is a mixture of (a) Loschbour and Karelia_HG or (b) LaBrana1 and
Samara_HG. By plotting these statistics we recover a signal of admixture that is not shown by the
f3(Motala_HG; Loschbour, Karelia_HG) statistic due to post admixture genetic drift in Motala_HG
ancestors. However, Karelia_HG (c) and Samara_HG (d) do not appear to be a mixture of
Loschbour and MA1. The lack of a negative correlation does not disprove the presence of such
admixture, but it provides an additional motivation for considering the alternative models in SI8.

(a) (b)

(c) (d)

We also explored the possibility that the eastern European hunter-gatherers like Karelia_HG and
Samara_HG were admixed, as we observed that they are intermediate between Loschbour and MA1
as measured by the statistics f4(Karelia_HG, MA1; Loschbour, Chimp) = 0.00625 (Z=7.4) and
f4(Karelia_HG, Loschbour; MA1, Chimp) = 0.00601 (Z=7.4), and it can be modeled as a mixture of
these two elements (SI8). However, when we use the regression method, we obtain a slightly positive
correlation for Karelia_HG and a slightly negative one for Samara_HG (Fig. S9.14c, d), which do not
appear convincing signals of admixture as in the case of Motala_HG. Note, however, that the lack of a
negative correlation does not, in itself, indicate lack of admixture, as it is possible that allele
frequency differences between the reference populations (Loschbour and MA1) may not be
sufficiently correlated with those of the outgroups, which is a necessary condition for the preservation
of this signal of admixture.

106
It is also possible to extract admixture proportions (for the case N=2) from a regression like the one in
Fig. S9.12 and Fig. S9.14, and we indicate estimated mixture proportions in these two Figures.
Remember that f4(Test, A; B, C) ≈ αf4(Ref1, A; B, C)+(1-α)f4(Ref2, A; B, C). This can be re-written as
(α+1-α)f4(Test, A; B, C) ≈ αf4(Ref1, A; B, C)+(1-α)f4(Ref2, A; B, C) or α(f4(Test, A; B, C)- f4(Ref1, A; B,
C)) ≈ (1-α)(f4(Ref2, A; B, C)- f4(Test, A; B, C)), or, αf4(Test, Ref1; B, C)≈ (α-1)(f4(Test, Ref2; B, C)), or,
௔ ௔
f4(Test, Ref2; B, C) ≈ f4(Test, Ref1; B, C). The coefficient can be fitted using a large number of
௔ିଵ ௔ିଵ
outgroups (as in Figs. S9.12 and S9.14), yielding an estimate for α.

For a fixed pair of outgroups (B, C):

௙ర ሺ்௘௦௧ǡோ௘௙భ Ǣ஻ǡ஼ሻ ିଵ
ܽො ൌ ሺͳ െ ሻ (S9.3)
௙ర ሺ்௘௦௧ǡோ௘௙మ Ǣ஻ǡ஼ሻ

So, an estimate can be derived analytically (similar to an f4-ratio3,4). However, by using a larger
number of outgroups, the mixture proportion can be estimated from the regression equation which
makes use of the correlation between f4-statistics involving a large number of outgroups, and also
obviates the need of choosing particular outgroup populations when a large number of them are
available for analysis.

More generally, we view this admixture estimation procedure as a substantial generalization of f4-ratio
estimation3,4. Unlike f4-ratio estimation, the method can work even without a formal model of the
historical relationships of all the analyzed populations. All that is required is that the Test population
be related to the outgroups via ancestral populations that form clades with the Ref1 and Ref2
populations (the reference populations themselves may be complicated and poorly understood
mixtures of the outgroups).

Interpretation of negative correlation

We have so far treated the 15 world outgroups as a “magic set” whose members are unequally related
to ancient reference populations and can thus help detect admixture (and estimate its proportions) in a
Test population. To gain understanding on why this method works, we examined the particular
statistics plotted in Fig. S9.12 and noticed that the pattern of linear anti-correlation was most strongly
affected by Native American and North Asian populations. We re-plot the data points of Fig. S9.13b
in Fig. S9.15, color-coding five clusters that were inferred in unsupervised manner with a model-
based clustering algorithm (mclust5,6).

We examined the most extreme “orange” and “purple” clusters of Fig. S9.15 which contribute to the
inference of admixture, and list the relevant statistics in Table S9.1. The statistics involving Yamnaya
and LBK_EN have consistently opposing signs (thus Corded_Ware_LN is intermediate between these
populations), and involve comparisons with the Native American Karitiana and the Chukchi and
Eskimo populations from northeast Siberia. The inclusion of these populations (who have “Ancient
North Eurasian” ancestry2) can discriminate between LBK_EN and Yamnaya, as Yamnaya share most
drift with them, LBK_EN least, and Corded_Ware_LN intermediate.

107
Figure S9.15: Admixture in the Corded_Ware_LN between Esperstedt_MN and Yamnaya reference
populations can be detected due to the statistics color-coded “orange” “blue” “green” and “purple”, as
inferred by mclust5. The cluster of “red square” statistics has a mean close to (0,0) and consists of
outgroup population pairs to which the two references are approximately equally related. The line
passing through the means of the “purple” and “orange” clusters has intercept ~0. It is these statistics
that primarily contribute to admixture inference.

108
Table S9.1: Corded_Ware_LN is intermediate between Yamnaya and LBK_EN in sharing
alleles with “Siberian/Native American” populations
O2 O3 f4(Corded_Ware_LN, Yamnaya; O2, O3) Z f4(Corded_Ware_LN, LBK_EN; O2, O3) Z
Biaka Chukchi 0.00074 3.3 -0.00270 -11.7
Biaka Eskimo 0.00079 3.5 -0.00275 -11.4
Biaka Karitiana 0.00087 3.3 -0.00325 -11.8
Chukchi Ju_hoan_North -0.00049 -2.1 0.00305 12.7
Chukchi Mbuti -0.00069 -2.9 0.00281 11.4
Chukchi Yoruba -0.00073 -3.4 0.00273 12.5
Eskimo Ju_hoan_North -0.00055 -2.3 0.00311 12.3
Eskimo Mbuti -0.00075 -3.0 0.00286 11.2
Eskimo Yoruba -0.00078 -3.6 0.00278 12.2
Ju_hoan_North Karitiana 0.00063 2.2 -0.00361 -12.7
Karitiana Mbuti -0.00082 -2.9 0.00336 11.7
Karitiana Yoruba -0.00086 -3.3 0.00328 12.5

Robustness to outgroup removal

Estimated admixture proportions rely on the set of 15 outgroups, so a potential source of error is the
choice of these outgroups. If estimated admixture proportions are sensitive to this choice then this
would reduce our confidence in them, as they would be driven by (little understood) details in the
history of the outgroups in relation to West Eurasians, rather than the admixture processes in the
history of the tested groups. To investigate the resilience of our method to the outgroup populations,
we estimated mixture proportions by dropping each of the outgroups in turn. Results are summarized
in Table S9.2 and indicate that the method is not sensitive to the particular set of outgroups.

Figure S9.16: Adding EN or MN populations to the set of outgroups reduces the amount of
Esperstedt_MN related ancestry (“red”) and increases the amount of Karelia_HG related
ancestry (“blue”) inferred for LN/BA populations.

(a) Corded_Ware_LN (b) Bell_Beaker_LN

(a) BenzigerodeHeimburg_LN (b) Unetice_EBA

109
Table S9.2: Estimated admixture proportions are robust to the choice of outgroups. (These proportions are estimated using the non-negative least
squares method). The estimate for the best N=2 pair of reference populations when using all 15 outgroups is given in the top row; in the bottom row the
maximum difference between this estimate and the estimate when one of the 15 outgroups is dropped is given. This is always ≤4%, suggesting that inferences
are not driven by any single outgroup.

Baalberge_MN Esperstedt_MN Spain_MN Karsdorf_LN Corded_Ware_LN Bell_Beaker_LN Alberstedt_LN BenzigerodeHeimburg_LN Unetice_EBA Halberstadt_LBA

Karelia_HG

Karelia_HG

Karelia_HG

Karelia_HG
Motala_HG

Motala_HG
Spain_MN

Spain_MN

Spain_MN

Spain_MN

Spain_MN
Loschbour

Loschbour

Loschbour
Spain_EN

Spain_EN
LBK_EN

LBK_EN

LBK_EN
Yamnaya

WorldFoci15 0.825 0.175 0.338 0.662 0.788 0.212 0.816 0.184 0.209 0.791 0.248 0.752 0.250 0.750 0.289 0.711 0.563 0.437 0.297 0.703
Ami 0.825 0.175 0.333 0.667 0.790 0.210 0.814 0.186 0.208 0.792 0.247 0.753 0.250 0.750 0.289 0.711 0.563 0.437 0.297 0.703
Biaka 0.824 0.176 0.335 0.665 0.783 0.217 0.814 0.186 0.203 0.797 0.245 0.755 0.246 0.754 0.288 0.712 0.557 0.443 0.299 0.701
Bougainville 0.823 0.177 0.343 0.657 0.784 0.216 0.817 0.183 0.209 0.791 0.248 0.752 0.250 0.750 0.289 0.711 0.562 0.438 0.294 0.706
Chukchi 0.827 0.173 0.349 0.651 0.795 0.205 0.824 0.176 0.203 0.797 0.253 0.747 0.253 0.747 0.289 0.711 0.566 0.434 0.293 0.707
Eskimo 0.819 0.181 0.346 0.654 0.788 0.212 0.811 0.189 0.200 0.800 0.249 0.751 0.263 0.737 0.293 0.707 0.571 0.429 0.295 0.705
Han 0.825 0.175 0.335 0.665 0.791 0.209 0.817 0.183 0.208 0.792 0.248 0.752 0.250 0.750 0.289 0.711 0.563 0.437 0.297 0.703
Ju_hoan_North 0.827 0.173 0.298 0.702 0.788 0.212 0.815 0.185 0.224 0.776 0.241 0.759 0.247 0.753 0.284 0.716 0.552 0.448 0.294 0.706
Karitiana 0.814 0.186 0.374 0.626 0.752 0.248 0.829 0.171 0.218 0.782 0.251 0.749 0.234 0.766 0.287 0.713 0.562 0.438 0.301 0.699
Kharia 0.825 0.175 0.336 0.664 0.787 0.213 0.818 0.182 0.208 0.792 0.248 0.752 0.254 0.746 0.289 0.711 0.564 0.436 0.298 0.702
Mbuti 0.843 0.157 0.342 0.658 0.810 0.190 0.826 0.174 0.204 0.796 0.249 0.751 0.256 0.744 0.285 0.715 0.567 0.433 0.297 0.703
Onge 0.832 0.168 0.327 0.673 0.790 0.210 0.809 0.191 0.210 0.790 0.248 0.752 0.248 0.752 0.286 0.714 0.562 0.438 0.300 0.700
Papuan 0.828 0.172 0.341 0.659 0.790 0.210 0.821 0.179 0.213 0.787 0.245 0.755 0.248 0.752 0.293 0.707 0.562 0.438 0.294 0.706
She 0.825 0.175 0.331 0.669 0.791 0.209 0.817 0.183 0.208 0.792 0.247 0.753 0.250 0.750 0.289 0.711 0.563 0.437 0.298 0.702
Ulchi 0.818 0.182 0.340 0.660 0.792 0.208 0.815 0.185 0.211 0.789 0.247 0.753 0.249 0.751 0.287 0.713 0.563 0.437 0.297 0.703
Yoruba 0.825 0.175 0.345 0.655 0.788 0.212 0.796 0.204 0.202 0.798 0.249 0.751 0.250 0.750 0.289 0.711 0.564 0.436 0.306 0.694
Max. diff 0.018 0.018 0.040 0.040 0.036 0.036 0.020 0.020 0.016 0.016 0.006 0.006 0.016 0.016 0.004 0.004 0.011 0.011 0.009 0.009

110
Robustness to outgroup addition
We can also add more populations to the set of 15 outgroups and see whether our conclusions are
affected. The 15 outgroups are a “conservative” choice as they lack (both due to geography and as
empirically assessed with ADMIXTURE) close ties to Europe. However, this comes at the cost of
potentially having little or no relevance to the different relatedness of European populations. We have
seen that this is not the case, as f4-statistics relating them to ancient European populations make it
possible to recover the presence of admixture and to estimate its proportions. However, by attempting
to include outgroups that are genetically closer to the populations under study (and which thus share
more genetic drift), we can increase our ability to perform these tasks and gain new insight.

We add ancient populations to the set of outgroups and re-estimate admixture proportions using the
new set of 15+1 outgroups. An interesting pattern occurs when we attempt to fit Corded_Ware_LN as
a mixture of Esperstedt_MN (which is the youngest Middle Neolithic sample from Germany
predating the appearance of the Corded Ware) and Karelia_HG (Fig. S9.16). When any of the EN/MN
European populations is added to the set of outgroups, the fraction of Esperstedt_MN ancestry
inferred for Corded_Ware_LN is reduced. The same pattern is also observed for other LN/BA
populations. This hints that European farmer populations do not function as outgroups, violating in
some way our assumptions.

To better understand how European farmer populations influence admixture proportions when added
as outgroups, we plot statistics of the form f4(Corded_Ware_LN, Karelia_HG; O2, O3) and
f4(Corded_Ware_LN, Esperstedt_MN; O2, O3) in Fig. S9.17. In each panel of Fig. S9.17 we add a
different European farmer population (LBK_EN, Spain_EN, HungaryGamba_EN, and Iceman),
representing different locations in Europe. All present a similar pattern: statistics of the form
f4(Corded_Ware_LN, Esperstedt_MN; O2, European Farmer) do not fall on the same line as those
with the “World Foci 15” outgroups, but appear to be transposed by a constant positive value. This is
unexpected if Corded_Ware_LN shares alleles with European Farmer only via a population like
Esperstedt_MN. While the regression of “World Foci 15” outgroups passes clearly through the origin,
௔
in accordance with the modeling prediction that f4(Test, Esperstedt_MN; O2, O3) ≈ f4(Test,
௔ିଵ
Karelia_HG; O2, O3) derived above, the statistics f4(Test, Karelia_HG; O2, European Farmer) and
f4(Test, Esperstedt_MN; O2, European Farmer) clearly do not have this behavior.

Fig. S9.18 suggests an explanation for these findings. We empirically observe that f4(Test,
Karelia_HG; O2, European Farmer) = -(1-α)z and f4(Test, Esperstedt_MN; O2, European Farmer)
=x+αz. Thus:

௙ర ሺ்௘௦௧ǡ•’‡”•–‡†–̴Ǣைమ ǡா௨௥௢௣௘௔௡ி௔௥௠௘௥ሻ ௫ାఈ௭ ఈ ௫

ൌ ൌ ൅
௙ర ሺ்௘௦௧ǡƒ”‡Ž‹ƒ̴
Ǣைమǡா௨௥௢௣௘௔௡ி௔௥௠௘௥ሻ ିሺଵିఈሻ௭ ఈିଵ ିሺଵିఈሻ௭
(S9.4)

Multiplying by f4(Test, Karelia_HG; O2, European Farmer), and using the fact that f4(Test,
Karelia_HG; O2, European Farmer) = -(1-α)z:

ఈ
݂ସ ሺܶ݁‫ݐݏ‬ǡ•’‡”•–‡†–̴Ǣܱଶ ǡ ‫ݎ݁݉ݎܽܨ݊ܽ݁݌݋ݎݑܧ‬ሻ ൌ ݂ ሺܶ݁‫ݐݏ‬ǡƒ”‡Ž‹ƒ̴
Ǣܱଶ ǡ ‫ݎ݁݉ݎܽܨ݊ܽ݁݌݋ݎݑܧ‬ሻ ൅ ‫ݔ‬ (S9.5)
ఈିଵ ସ

111
This is the functional form observed in Fig. S9.17; the positive offset x is seen for all European farmer
populations when added as outgroups, and this suggests that the admixing population Q (Fig. S9.18)
split before the differentiation of European farmers from each other. Note also that the positive offset
would appear if at least some of the ancestry of the Test population came from Q (for which European
Farmer is not a “real outgroup”) and some of it came from a population closely related to
Esperstedt_MN (for which European Farmer is a real outgroup).

Figure S9.17: Adding early European farmers to the set of outgroups

(Esperstedt_MN+Karelia_HG mix). In blue are statistics involving only “World Foci 15”
outgroups. In red are statistics where O3 is the early European farmer population. The two fitted lines
differ by a positive vertical offset.

112
Figure S9.18: When a European Farmer population is treated as an outgroup, but in fact has
ancestry that is as closely related to the true admixing population Q as population
Esperstedt_MN that we use as a reference, the assumptions of our negative regression analysis
are violated. In this simplified model, the European Farmer population shares genetic drift x with
Esperstedt_MN after the split of the true admixing population Q, which results in an offset of the
scatterplot by an amount x for f4-statistics involving European Farmer. In truth, the history is more
complicated (elsewhere in this paper, we show that both European Farmer and Esperstedt_MN have
WHG-related ancestry in different proportions). However, we hypothesize that the patterns in Figure
S9.17 are driven by the fact that a substantial amount of the ancestry in European Farmer and
Esperstedt conforms to these relationships.

In Fig. S9.19 we show the same statistics as Fig. S9.17, but using Yamnaya as an eastern reference
population. The substantial offset seen in Fig. S9.17 is not evident when using Yamnaya. There is
evidence (SI7) that Yamnaya has diluted EHG (Karelia_HG-related) ancestry from a source related to
present-day Near Eastern populations. If some of the farmer-related ancestry in the Corded Ware was
not of local European origin but came with the Yamnaya, ultimately from the Near East, then this
ancestry would represent a farmer-related population that was basal to European farmers and thus
occupy a phylogenetic position similar to that of Q in Fig. S9.18.

113
Figure S9.19: Adding early European farmers to the set of outgroups
(Esperstedt_MN+Yamnaya mix). Unlike Fig. S9.17, the red statistics involving a European farmer
as O3 line up fairly well with the blue “World Foci 15” outgroups.

Ancient genomes from the Caucasus, the Near East, and Central Asia might reveal the existence of
Neolithic populations there that may be involved in the ancestry of ancient steppe populations and
central Europeans. The existence of such ancestry in the Yamnaya is visualized in Fig. S9.20, where
this population is plotted as a mixture of Karelia_HG and either Iraqi_Jew or Armenian, showing a
clear pattern of negative correlation.

114
Figure S9.20: The Yamnaya show a pattern of negative correlation when using Karelia_HG and
either Iraqi_Jew or Armenian as references. This is consistent with it having a component of
ancestry related to the Caucasus/Near East, suggesting that ancient DNA work in these regions may
reveal a good surrogate for this type of ancestry.

Figure S9.21: Adding different outgroups in position O3 of the statistics f4(Corded_Ware,

Eastern; O2, O3) vs. f4(Corded_Ware, Esperstedt_MN; O2, O3). (a) Eastern=Karelia_HG, (b)
Eastern=Yamnaya

(a) (b)

In Fig. S9.21 we plot statistics of the form f4(Corded_Ware, Eastern; O2, O3) vs. f4(Corded_Ware,
Esperstedt_MN; O2, O3), color coding different ancient European outgroups for O3. We plot a number
of different regression equations on this plot: the solid black line corresponds to the regression using
only the world outgroups (which is the same as in Fig. S9.12). The dashed black line is the regression
through (0,0) using all plotted data points, and the dotted black line is the regression using all data
points but with a non-zero intercept. The color-coded lines are regressions for the subsets of color-

115
coded points representing the different choices for O3. Fig. S9.21 visually explains why the addition
of early farmers as outgroups results in a change of ancestry estimates (Fig. S9.16): the inclusion of
early farmers in the set of outgroups results in a regression (dashed line) with a different slope to
account for the “blue” statistics involving European farmers in the top left quadrant of Fig. S9.16a).

Fig. S9.16a also shows that statistics involving European hunter-gatherers (Kostenki14, Loschbour,
and Motala_HG) are also transposed by a positive amount along the vertical axis corresponding to the
statistic f4(Corded_Ware_LN, Esperstedt_MN; O2, European hunter-gatherer). This reinforces the
idea of shared drift between Esperstedt_MN and European hunter-gatherer that is not shared by a
component in the ancestry of Corded_Ware_LN. Thus, Corded_Ware_LN has ancestry from a
component that is basal to both European farmers (who have the maximum positive offset) and
European hunter-gatherers. A slight negative offset is seen in Fig. S9.16b. for the statistics involving
European hunter-gatherers and world outgroups. This hints that the Yamnaya population, while
visibly a much better ancestral source for the Corded Ware, may in fact not be the exact admixing
population. This also agrees with the results of N=3 discussed below and presented in Extended Data
Fig. 3, which estimates that the Corded Ware can be modeled as 29.1% Esperstedt, 9.4% Samara_HG,
and 61.5% Yamnaya, which suggests that the population of eastern migrants had a slightly higher
proportion of EHG ancestry in its makeup than the Yamnaya sample from Samara. Such a conclusion
might also be drawn from the f4-statistic presented in SI7 (Table S7.6) that shows
f4(Corded_Ware_LN, Yamnaya; Karelia, Chimp) = -0.00001 (Z=0.0). If Corded_Ware_LN was a
simple mixture of a population related to our Yamnaya sample and of Neolithic Europeans, this
statistic should be negative. However, if Corded_Ware_LN is descended from a population that has a
higher proportion of EHG ancestry than the Yamnaya population, then the dilution of EHG ancestry
due to European Neolithic admixture (which would cause the statistic to be negative), would be
counterbalanced by its increase due to this higher EHG ancestry (which would cause it to be positive).
It is quite possible that the variable mixtures of EHG and farmer populations existed in the European
steppe, and our Yamnaya population represents only a point in a continuum of such mixtures.

Figure S9.22: Adding multiple ancient outgroups. The regression using

(Esperstedt_MN+Yamnaya) is virtually identical when using World Foci 15 + Ancient outgroups and
when using only World Foci 15 (blue dots). However, the two lines are noticeably different for the
(Esperstedt_MN+Karelia_HG) model.

116
Finally, we included all ancient populations predating the Corded Ware in the set of outgroups
(excluding the Samara_HG, Karelia_HG, Yamnaya, Esperstedt_MN, Baalberge_MN that were
possible reference populations for the Corded Ware):

“World Foci 15 + Ancients”: Ami, Biaka, Bougainville, Chukchi, Eskimo, Han, Ju_hoan_North,
Karitiana, Kharia, Mbuti, Onge, Papuan, She, Ulchi, Yoruba, Starcevo_EN, LBKT_EN, LBK_EN,
Spain_EN, HungaryGamba_EN, Stuttgart, Spain_MN, Loschbour, LaBrana1, HungaryGamba_HG,
Motala_HG, SwedenSkoglund_MHG, SwedenSkoglund_NHG, MA1, Kostenki14, Ust_Ishim

We estimate mixture proportions using the regression method in Fig. S9.22. The regression plot is
visibly similar when using only “World Foci 15” or “World Foci 15 +Ancients” when Yamnaya is the
eastern reference population, but visibly different when Karelia_HG is the eastern reference.

In Extended Data Fig. 3 we also estimate mixture proportions for the Corded_Ware_LN using the
non-negative least squares optimizer using “World Foci 15 + Ancients” as outgroups. The two worst
performing N=2 models (Extended Data Fig. 3), involve either descent from only the Middle
Neolithic populations or the Eastern European hunter-gatherers. The two best performing ones
involve descent from the Yamnaya and either of the Middle Neolithic populations of Germany in
proportions ~3/4 and ~1/4 respectively. The estimated mixture proportions are very similar to those
derived using a related method in SI10.

Proximate ancestral populations for present-day Europeans

We also attempted to estimate ancestry proportions for present-day European populations in terms of
the ancient European populations. We recognize that some European populations have East Eurasian,
Near Eastern, or African ancestry that may not be modeled by the ancient European populations
included in our study (from Iberia and Central/Northern/Eastern Europe). It is important to remember
that inference of particular ancestral populations does not imply direction of migration for present-day
European populations. For example, Sardinians often appear similar to ancient Neolithic farmers from
continental Europe2,7-9, but this does not imply migration from continental Europe into Sardinia.
Instead, the finding that Cardial Ware-related Neolithic Iberians were similar to those from central
Europe suggests that early farmers that followed a Mediterranean route into Europe may have also
settled Sardinia. Similarly, there are no ancient samples from southeastern Europe in our study, yet
groups from the Balkans can be modeled2 as a 3-way mixture of the same ancestral populations that
contributed to central, western, and northern Europe2. The fact that many European groups today can
be well-described by mixtures of populations represented by ancient individuals from a small number
of European locations suggests that the sampled ancient individuals are representative of widely
distributed populations.

In Fig. S9.23 we show the best-fitting (lowest resnorm) models for European populations using the
“World Foci 15” outgroups and reference populations by period: WHG, EHG, SHG, EN, MN,
LN/BA, Yamnaya. Varying N=1, 2, 3 we observe some interesting patterns. At N=1, present-day
Europeans from central/northern Europe prefer the LN/BA population as a reference, while those
from southern Europe prefer the EN/MN ones. This is not surprising given the fact that LN/BA
Europeans are temporally closer to the present and belonged to archaeological cultures (such as
Corded Ware or Bell Beaker) with wide geographical ranges. There is also substantial variation in the
goodness of fit for different populations, hinting that many of them are not well-modeled with N=1
(and in agreement with our findings in SI7 that indicate discontinuity between LN/BA populations
and present-day ones).

117
Figure S9.23: Identifying proximate ancestral populations for present-day Europeans. (a, b, c):
Best reference populations for present-day Europeans with N=1, 2, 3 admixing populations. (d):
Change in resnorm as N changes.

At N=2 most populations appear to be a mixture of two different elements which are EN/MN and
LN/BA. There are interesting exceptions such as the Lithuanians (who have LN/BA plus SHG,
consistent with our finding that they have more hunter-gatherer ancestry than Late Neolithic
Europeans, SI7). For most other northern European populations it seems that the best N=2 model
involves admixture between LN/BA and MN, consistent with our observations about the dilution of
the ancestry brought into Late Neolithic Europe in modern populations (SI7). Conversely, modern
southern Europeans (who generally fit worse than northern ones for N=1) appear to be a mixture of
EN/MN and LN/BA. This does not necessarily imply that LN/BA populations of central northern
Europe admixed into the populations of southern Europe as the LN/BA central Europeans may be the
best proxy in our dataset for the ANE-bearing populations that affected the entire European
continent2. Unlike central Europe where the appearance of ANE-related populations can be dated to
the Late Neolithic by the results of our paper, we do not know when such populations arrived

118
elsewhere in Europe. Did they arrive at the same time, as part of a general phenomenon affecting the
entire European continent? Alternatively, did they arrive later, perhaps related to migrations from
north to south that overlap the period of recorded history10,11?

N=3 seems to provide very little benefit in the fit of the mixture model, indicating that three proximate
sources are not needed for most European populations. This finding is not in contradiction with the
three source population model for Europeans.2 Instead, the finding shows that these source
populations contributed to many present-day European populations through the intermediary of as few
as two proximate populations (in different proportions), which we are able to model as clade with
samples for which we have ancient DNA data.

Fig. S9.24d shows visually how model fit improves as N increases: significant resnorm at N=1 is
greatly reduced at N=2 but there is virtually no change at N=3. Some populations, including
previously identified Jewish, Mediterranean, and Northeastern European outliers2, differ markedly
from most Europeans even at N=3, suggesting that they possess ancestry from additional sources not
included in the set of references.

Proportions of EN, WHG, and EHG/Yamnaya ancestry in ancient and modern Europeans
The previous analysis used MN and LN/BA Europeans as reference populations for modern ones, but
it is also useful to treat them as test populations and estimate their ancestral proportions together with
those of modern Europeans. Fig. S9.24 shows that when WHG admixture is added to EN, residuals
for most European populations are reduced, consistent with most Europeans not being descended
from EN farmers of Europe. Four populations indicate no change in residuals: Sardinians are the
population that is closest to early European farmers2,7-9,12 with an estimated ~90% descent from them
(ref. 2 and Fig. S9.23b), while Maltese, Ashkenazi Jews, and Sicilians may have Near Eastern
admixture not mediated via early European farmers2.

Fig. S9.25 shows that when EHG/Yamnaya admixture is added on top of the EN+WHG one, residuals
continue to improve for European populations. Three noticeable exceptions are Sardinians (who may
have little if any “Ancient North Eurasian” admixture2), MN Europeans (which is not surprising as
they precede the arrival of EHG/Yamnaya-related populations into central Europe), and Maltese (an
island European population that may also have been less affected by this type of ancestry). For all
other European populations (including the LN/BA and modern Europeans across the continent), the
addition of EHG/Yamnaya admixture visibly improves the residuals. This indicates that ancient
eastern European populations eventually contributed ancestry to most European populations. In
central Europe this occurred as early as the Late Neolithic. Future ancient DNA research is likely to
show whether the rest of Europe was also affected at that time or whether this type of ancestry spread
earlier or later to different parts of Europe.

Three northeastern European populations (Finns, northwestern Russians, and Mordovians) continue to
have noticeably higher residuals at N=3 in Fig. S9.25. We have previously determined that this is due
to more recent gene flow from Siberia2, perhaps related to the spread of Finno-Ugric languages.
Unfortunately, we do not have access to an ancient DNA sample that might be representative of this
population; clearly Eastern European hunter-gatherers like Karelia_HG who lived in northeastern
Europe are not good representatives of this population (as their effect is Europe-wide and not limited
to northeastern Europe and as they fail to fully bring northeastern European residuals in line with
those of other Europeans). We therefore tried using Nganasan as a new reference population. The
Nganasan are a north Siberian Samoyedic-speaking population with little evidence of West Eurasian

119
admixture in ADMIXTURE K=3 (SI6) analysis and with a high frequency13 (92.1%) of Y-haplogroup
N, which may be associated with the arrival of a separate stream of Siberian influence into parts of
Europe. Fig. S9.26 shows that an N=4 model with the addition of the Nganasan to either the
EN+WHG+EHG or the EN+WHG+Yamnaya model visibly improves the residuals for the
northeastern European outlier populations, but has little effect on most others.

Figure S9.24: EN+WHG admixture improves residuals across most European populations. (a)
resnorm for N=1 EN model. (b) resnorm for N=2 EN+WHG model. (c) change in resnorm between
N=1 and N=2.

120
Figure S9.25: EHG or Yamnaya admixture improves residuals across most European
populations. (a) resnorm for N=3 EN+WHG+EHG model. (b) change in resnorm between N=2 and
N=3 EN+WHG+EHG model. (c) resnorm for N=3 EN+WHG+Yamnaya model. (d) change in
resnorm between N=2 and N=3 EN+WHG+Yamnaya model.

121
Figure S9.26: Adding Nganasan as a 4th reference population improves residuals for
northeastern European populations. The effect is strongest for Finns, Russians, and Mordovians,
followed by Hungarians and Estonians. A slighter effect is seen in a few other populations, especially
from eastern Europe; low levels of recent East Eurasian admixture in eastern Europe have been
proposed using haplotype sharing10.

122
The three populations with the highest resnorm at N=4 in Fig. S9.26 are Maltese, Sicilians, and
Ashkenazi Jews. We thus added the BedouinB as a 5th reference population, as there is evidence that
these populations may have an excess of Near Eastern ancestry not mediated via the early European
farmers. In Fig. S9.27 we show that this addition visibly improves the residuals for the three outlier
European populations (and to a lesser extent the Spanish), consistent with this evidence. As the
number of reference populations increases it may become increasingly difficult to accurately estimate
ancestry proportions, especially when using modern reference populations (who may have their own
history of admixture) and a particular set of outgroups (that may not be sufficient to distinguish
between closely related reference populations).

In choosing to add Nganasan and BedouinB as 4th and 5th reference populations we used prior
knowledge of the additional mixtures that may have taken place in Europe. It is also possible to
identify the sources of such mixtures without a priori knowledge. To show that our choice of a
Siberian and Near Eastern population could have been discovered in an unsupervised manner, we
added all modern and ancient populations to the EN+WHG+EHG model and identified (for each
outlier population) the additional references that minimize its resnorm. These are shown in Table S9.3
and indeed the two groups of outliers fit better when either Near Eastern or contemporary north Asian
populations are added as a 4th ancestral population.

Summary
In this section we devised methods for detecting admixture and estimating its proportions without
detailed phylogenetic modeling. Our assumptions were that a Test population is formed as a mixture
of an arbitrary number of other reference populations that are differentially related to a set of
outgroup populations. This approach allowed us to study admixture in ancient Europe without using a
tree or admixture graph model. We were able to show that Middle Neolithic Europeans had
experienced <30% resurgence of western European hunter-gatherer ancestry while the Corded Ware
(the earliest of the Late Neolithic populations) had experienced >1/3 gene flow from a population
related to eastern European hunter-gatherers. Furthermore, we showed that the Yamnaya population
from the early Bronze Age in the Russian steppe may have been related to the population that
admixed into Late Neolithic central Europeans, as it was not of purely eastern European hunter-
gatherer descent but had additional ancestry of a farmer-related population that split off before the
differentiation of European farmers, perhaps from the Caucasus or Near East, although more work is
needed to identify the source of this population. Taking the Yamnaya as a source of the eastern
admixing population, the level of population replacement between the Middle Neolithic and the
Corded Ware becomes ~3/4, suggesting a massive influence of eastern populations in the formation of
the Late Neolithic in central Europe. This influence is evident in other Late Neolithic populations (at a
lower level), as well as in present-day Europeans, who differ from the Late Neolithic central
Europeans by having additional farmer or west European hunter-gatherer ancestry. The robustness of
our inference is supported by the fact that by adding a different set of outgroups from ancient Europe,
the ~3/4 estimate was little affected. Further research of populations from central Europe, Eastern
Europe, and the Caucasus may reveal the formation of steppe groups that succeeded the eastern
European hunter-gatherers and how they affected (and were affected by) the populations surrounding
them. The fact that sizeable admixture accompanied major transitions in European prehistory in at
least three locations (Spain, Germany, and European Russia) suggests that the formation of present-
day European populations was effected by sizeable local population replacement. Archaeological
models of the past must accommodate the possibility of migration and admixture at a massive scale
long after the Neolithic transition in Europe.

123
Figure S9.27: Adding BedouinB as a 5th reference population improves residuals for Maltese,
Sicilians, Ashkenazi Jews, and Spanish.

124
Table S9.3: Populations that improve resnorm for European outlier populations when added to a model of EN/WHG/EHG admixture as a 4th
ancestral population. We show the top 20 populations that reduce resnorm the most when added to the mixture model. For Maltese, Sicilians, and Ashkenazi
Jews these populations tend to be from the Middle East and North Africa. For Finns, Russians, and Mordovians they tend to be populations from Central
Asia, Siberia, and Turkic/Uralic populations from Western Eurasia.

Maltese Sicilian Ashkenazi_Jew Finnish Russian Mordovian

4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm 4th anc. pop. resnorm
Moroccan_Jew 0.000006 Turkish_Jew 0.000006 Cypriot 0.000006 Chuvash 0.000013 Chuvash 0.000016 Turkmen 0.000011
Lebanese 0.000010 Cypriot 0.000006 Iraqi_Jew 0.000018 Mansi 0.000020 Turkmen 0.000020 Uzbek 0.000012
Syrian 0.000010 Moroccan_Jew 0.000009 Turkish_Jew 0.000018 Uzbek 0.000020 Uzbek 0.000021 Nogai 0.000012
Tunisian_Jew 0.000011 Druze 0.000012 Moroccan_Jew 0.000020 Even 0.000021 Nogai 0.000022 Chuvash 0.000014
Saudi 0.000014 Iraqi_Jew 0.000012 Druze 0.000025 Selkup 0.000021 Altaian 0.000025 Altaian 0.000017
Turkish_Jew 0.000016 Syrian 0.000014 Lebanese 0.000027 Dolgan 0.000021 Mansi 0.000025 Tubalar 0.000017
Libyan_Jew 0.000017 Lebanese 0.000016 Syrian 0.000028 Turkmen 0.000022 Tubalar 0.000025 Hazara 0.000017
Jordanian 0.000017 Tunisian_Jew 0.000019 Tunisian_Jew 0.000037 Nogai 0.000022 Selkup 0.000026 Kyrgyz 0.000019
Palestinian 0.000019 Saudi 0.000021 Saudi 0.000039 Nganasan 0.000022 Even 0.000026 Tuvinian 0.000019
Druze 0.000022 Jordanian 0.000022 Iranian_Jew 0.000039 Yakut 0.000022 Yakut 0.000026 Uygur 0.000019
Yemenite_Jew 0.000022 Libyan_Jew 0.000025 Jordanian 0.000039 Altaian 0.000023 Dolgan 0.000026 Yakut 0.000019
BedouinB 0.000022 Palestinian 0.000026 Palestinian 0.000044 Tubalar 0.000023 Tuvinian 0.000027 Oroqen 0.000020
BedouinA 0.000023 BedouinB 0.000027 Libyan_Jew 0.000044 Tuvinian 0.000024 Kyrgyz 0.000029 Kalmyk 0.000020
Tunisian 0.000024 Yemenite_Jew 0.000028 Yemenite_Jew 0.000045 Yukagir 0.000025 Hazara 0.000029 Even 0.000020
Mozabite 0.000024 Tunisian 0.000028 BedouinB 0.000045 Oroqen 0.000026 Nganasan 0.000029 Dolgan 0.000021
Algerian 0.000025 Mozabite 0.000028 BedouinA 0.000046 Kyrgyz 0.000026 Oroqen 0.000029 Selkup 0.000021
Egyptian 0.000025 BedouinA 0.000028 Tunisian 0.000046 Kalmyk 0.000027 Yukagir 0.000030 Hezhen 0.000021
Saharawi 0.000026 Egyptian 0.000028 Egyptian 0.000046 Hazara 0.000029 Kalmyk 0.000030 Daur 0.000021
Yemen 0.000028 Algerian 0.000029 Mozabite 0.000047 Hezhen 0.000030 Uygur 0.000032 Mansi 0.000022
Esan 0.000030 Saharawi 0.000029 Algerian 0.000047 Aleut 0.000031 Hezhen 0.000032 Xibo 0.000023

125
References

1. Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry in

unrelated individuals. Genome Res. 19, 1655-1664, (2009).
2. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).
3. Reich, D., Thangaraj, K., Patterson, N., Price, A. L. & Singh, L. Reconstructing Indian
population history. Nature 461, 489-494, (2009).
4. Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065-1093, (2012).
5. Fraley, C. & Raftery, A. E. Model-Based Clustering, Discriminant Analysis, and Density
Estimation. Journal of the American Statistical Association 97, 611-631, (2002).
6. Fraley, C., Raftery, A. E., Murphy, T. B. & Scrucca, L. mclust Version 4 for R: Normal
Mixture Modeling for Model-Based Clustering, Classification, and Density Estimation.
Department of Statistics, University of Washington Technical Report No. 597, (2012).
7. Keller, A. et al. New insights into the Tyrolean Iceman's origin and phenotype as inferred by
whole-genome sequencing. Nat. Commun. 3, 698, (2012).
8. Sikora, M. et al. Population Genomic Analysis of Ancient and Modern Genomes Yields New
Insights into the Genetic Ancestry of the Tyrolean Iceman and the Genetic Structure of
Europe. PLoS Genet. 10, e1004353, (2014).
9. Skoglund, P. et al. Origins and genetic legacy of Neolithic farmers and hunter-gatherers in
Europe. Science 336, 466-469, (2012).
10. Hellenthal, G. et al. A genetic atlas of human admixture history. Science 343, 747-751,
(2014).
11. Ralph, P. & Coop, G. The geography of recent genetic ancestry across Europe. PLoS Biol. 11,
e1001555, (2013).
12. Skoglund, P. et al. Genomic Diversity and Admixture Differs for Stone-Age Scandinavian
Foragers and Farmers. Science 344, 747-750, (2014).
13. Tambets, K. et al. The Western and Eastern Roots of the Saami—the Story of Genetic
“Outliers” Told by Mitochondrial DNA and Y Chromosomes. The American Journal of
Human Genetics 74, 661-682, (2004).

126
Supplementary Information 10
Number of streams of migration into ancient Europe

Nick Patterson*, Iosif Lazaridis*, David Reich

* To whom correspondence should be addressed

([email protected] or [email protected])

Number of streams of migration

Take a set of left populations U and a set of right populations V and consider the matrix

X(u, v) = F4(u0, u; v0, v)

where u0, v0 are some fixed populations of U and V, and u, v range over all choices of populations of
U, V. We can assume that u u≠u0, and v≠v0, so that the matrix X is (a − 1) × (b − 1), where a, b is the
cardinality of U, V respectively. We have shown1 that if X had rank r and there had been n waves of
migration from V to U with no back-migration from U to V then:

r+1≤n

We describe our computational strategy in a little more detail. We compute ܺ෠ an estimate of X so that
in the notation of ref. 2:

ܺ෠(u, v) = f4(u0, u; v0, v)

Thus:
‫ܧ‬ሺܺሻ ൌ ܺ෠

where expectations are calculated from the true (and unknown) phylogeny. We can use the block
jackknife3 to compute Q an estimate of the error covariance of X. To test if ܺ෠ has rank r we write:
ܺ෠ ൌ ‫ܣ‬Ǥ ‫ ܤ‬൅ ‫ܧ‬

where A is (a-1)×r and B is r×(b-1) and E is a matrix of residuals. The (log) likelihood for (A, B) and
implicitly r is:
ͳ ିଵ
ࣦሺ‫ܣ‬ǡ ‫ܤ‬ሻ ൌ െ ෍ ܳ௜ǡ௝Ǣ௞ǡ௟ ‫ܧ‬௜ǡ௝ ‫ܧ‬௞ǡ௟
ʹ
௜ǡ௝ǡ௞ǡ௟

where the residual matrix E is defined by:

‫ ܧ‬ൌ ܺ෠ െ ‫ܣ‬Ǥ ‫ܤ‬

For each rank r we set A, B initially by an SVD analysis of X, and then iterate, minimizing ࣦ with
respect to A, B in turn. For fixed A, ࣦሺ‫ܣ‬ǡ ‫ܤ‬ሻ is quadratic in B and can be minimized by solving linear
equations, and similarly if we fix B. Since A, B only enter into the likelihood through a matrix
product:

‫ܣ‬Ǥ ‫ ܤ‬ൌ ሺ‫ܣ‬Ǥ ܼሻǤ ሺܼ ିଵ Ǥ ‫ܤ‬ሻ

127
for any non-singular r×r matrix Z. Thus, the number of degrees of freedom is:

d(r) = ((a−1) + (b−1))r – r2 = r(a+b−(r+2))

As a check, if r is the maximal rank Min(a-1, b-1), then d(r)=(a-1)(b-1) which is obviously correct.
This is the saturated model, where we fit the data perfectly. We compute statistics with a likelihood
ratio test (LRT), obtaining statistics that are asymptotically ߯ ଶ distributed. The ߯ ଶ statistics here,
using the LRT, are computed using a fixed covariance Q. It would be more correct to re-estimate Q
simultaneously with A, B. This would greatly increase complexity, without adding much precision.

Finding mixture coefficients

Let T be a target population, and S = {s1, s2, …, sn} a set of source populations. In the easiest case to
consider, in which T is an admixture of populations of S, we can write symbolically:
௡

ܶ ൌ ෍ ‫ݓ‬௜ ‫ݏ‬௜
௜ୀଵ

where wi is the admixture coefficient (ancestry proportion) from the ith source population. It then
follows that for any populations o1, o2:

෍ ‫ݓ‬௜ ‫ܨ‬ସ ሺܶǡ ‫ݏ‬௜ ǡ ‫݋‬ଵ ǡ ‫݋‬ଶ ሻ ൌ ‫ܨ‬ସ ሺܶǡ ܶǢ ‫݋‬ଵ ǡ ‫݋‬ଶ ሻ ൌ Ͳ
௜

A little thought shows that this is true even if the populations si are descendants of the true source
populations, provided that there has been no gene flow between the most recent ancestor of T, S on
the one hand and the most recent ancestor of o1, o2 on the other. (We previously used f3 statistics to
derive mixing coefficients4. The methods there require samples of the actual source and admixed
populations, but do not require outgroup populations as we do here with the oi.)

Thus, if T is admixed, as above, pick a set of outgroup populations O, and

1. Check, setting left populations L = S, and right populations O that the matrix X has full rank n-1.
2. Check, again setting L = {T, S} that there is no strong evidence that the rank of X increases with
the addition of T.

We now will take T as the base population of L = {T, S}, which simplifies the algebra. We calculate
matrices A, B as above, with the rank set to n-1 (corank 1). Thus, the recovered A is of dimension
n×(n-1). We have n source populations so L has n+1 populations after including T. It then follows that
estimates w = (w1,w2, …,wn) of the admixture weights can be found by solving the equations:

࢝Ǥ ࡭ ൌ ૙
௡

෍ ‫ݓ‬௜ ൌ ͳ
௜ୀଵ

We can use the block jackknife to compute a covariance matrix for the errors. Formally, we should re-
estimate Q, the estimated covariance of X as we delete blocks in the jackknife. This is not done at
present, as it would add complexity and seems unlikely to make a material difference.

128
Application to ancient Europe
We use this methodology to study some of the proposed admixture events discussed in the paper. We
begin by using the same 15 World Outgroups as in SI9 (this is the right population list V).

World Foci 15 outgroups: Ami, Biaka, Bougainville, Chukchi, Eskimo, Han,

Ju_hoan_North, Karitiana, Kharia, Mbuti, Onge, Papuan, She, Ulchi, Yoruba
In the following discussion we will follow the idea described above of first testing a pair of source
populations (left list U) to ensure that they have rank=1 with respect to the outgroups, which shows
that they are descended from two independent streams of ancestry. We will then add a third
population Test, making U={U, Test}, showing that it does not increase the rank, and can thus be
modelled as a mixture of the initial pair.

Western and Eastern European hunter-gatherers not a clade with respect to world outgroups
We first test the pair (Loschbour, Karelia_HG). Model with rank=0 is rejected (p=1.6e-43), thus the
WHG and EHG are descended from at least two ancestral populations.

Scandinavian hunter-gatherers not a third ancestral population independent of WHG and EHG
We next add Scandinavian hunter-gatherers to (Loschbour, Karelia_HG) (Table S10.1). Rank=0 is
excluded, but rank=1 is not. Two ancestral populations suffice for Scandinavian hunter-gatherers.

Table S10.1: Testing (SHG, Loschbour, Karelia_HG)

SHG population added p for rank=0 p for rank=1
Motala_HG 8.93E-44 7.09E-01
SwedenSkoglund_MHG 4.26E-13 2.15E-01
SwedenSkoglund_NHG 7.49E-44 4.93E-01

Early Neolithic Europeans and Western European hunter-gatherers (EN and WHG) are not a
clade with respect to world outgroups
We first show that the pair of an early Neolithic population EN and Loschbour are descended from
two streams of migration (Table S10.2). We add all Early Neolithic populations in turn, and in each
case rank=0 is rejected. Thus, EN and WHG are descended from at least two ancestral populations.

Table S10.2: Testing (EN, Loschbour)

EN population added p for rank=0
Starcevo_EN 9.49E-09
LBKT_EN 4.16E-04
LBK_EN 7.17E-15
Spain_EN 1.84E-11
HungaryGamba_EN 6.24E-13
Stuttgart 9.73E-10

Middle Neolithic Europeans are not a 3rd ancestral population independent of WHG and EN
We next add a Middle Neolithic European population MN to (Loschbour, LBK_EN) (Table S10.3).
With the exception of the Iceman, most Middle Neolithic Europeans appear to be well-modeled as a
mixture of Loschbour and LBK_EN.

129
Table S10.3: Testing (MN, LBK_EN, Loschbour)
MN population added p for rank=0 p for rank=1
Baalberge_MN 1.92E-11 7.59E-01
Esperstedt_MN 4.57E-15 4.28E-01
Spain_MN 2.18E-14 1.52E-01
HungaryGamba_CA 4.49E-12 2.44E-01
SwedenSkoglund_MN 4.85E-13 1.46E-01
Iceman 1.34E-15 1.87E-04
Yamnaya and Eastern European hunter-gatherers not a clade with respect to world outgroups
We test the pair (Yamnaya, Karelia_HG). The model with rank=0 is rejected (p= 2.5e-28). Thus the
Yamnaya and EHG are descended from at least two ancestral populations. This is related to the
process of dilution of EHG ancestry discussed in SI7, SI9.

Yamnaya can be modeled as a mixture of Armenians and Karelia_HG

To understand the source of the dilution of EHG ancestry in the Yamnaya, we test the triple
(Yamnaya, Karelia_HG, Armenian). We do not strongly reject rank=1 for (Yamnaya, Karelia_HG,
Armenian) (p=0.0365). The p-value is low, which may be driven by complexity in the ancestry of
present-day Armenians. We may not currently have a good surrogate for the population that diluted
the EHG to form the Yamnaya, although the analysis of f-statistics (SI7) suggests a source related to
Near Eastern populations, and the Yamnaya show evidence of such admixture (SI7, SI9).

Table S10.4: Testing (MN, Eastern)

Eastern MN p for rank=0
Yamnaya Baalberge_MN 4.95E-29
Karelia_HG Baalberge_MN 4.57E-68
Yamnaya Esperstedt_MN 5.31E-21
Karelia_HG Esperstedt_MN 7.82E-56
Yamnaya Spain_MN 4.58E-61
Karelia_HG Spain_MN 3.68E-103
Yamnaya HungaryGamba_CA 4.55E-18
Karelia_HG HungaryGamba_CA 4.66E-52
Yamnaya SwedenSkoglund_MN 1.60E-35
Karelia_HG SwedenSkoglund_MN 4.26E-76
Yamnaya Iceman 3.29E-36
Karelia_HG Iceman 1.91E-78

Middle Neolithic Europeans and Yamnaya or Karelia_HG not a clade versus world outgroups
We next test whether a Middle Neolithic European population MN is a clade with the populations of
eastern Europe (Yamnaya or Karelia_HG). Rank=0 is universally rejected (Table S10.4).

The Corded Ware are modeled as a mixture of Middle Neolithic and Eastern Europeans
We have determined that eastern European populations (Karelia_HG or Yamnaya) and Middle
Neolithic populations from the rest of Europe are descended from two ancestral populations. Next, we
add the Corded Ware. In Table S10.5 we list the p-values for models that involve Esperstedt_MN,
Baalberge_MN, i.e., the populations that precede the Corded Ware in Germany. The p-value for
rank=1 is ≥.179, thus rank=1 is not excluded, and the Corded Ware can be modeled as a mixture of an

130
eastern population (Yamnaya or Karelia_HG) and a Middle Neolithic population from Germany
(Baalberge_MN or Esperstedt_MN).

Table S10.5: The Corded Ware can be modeled as mixtures of Eastern European and Middle
Neolithic central Europeans
Eastern population MN population p for rank=0 p for rank=1
Yamnaya Baalberge_MN 6.24E-29 2.38E-01
Karelia_HG Baalberge_MN 5.59E-68 1.79E-01
Yamnaya Esperstedt_MN 2.38E-22 4.21E-01
Karelia_HG Esperstedt_MN 8.34E-59 7.07E-01

Modeling the Corded Ware as a mixture of Middle Neolithic Europeans and Yamnaya is robust
to choice of outgroups, while modeling it as a mixture of Middle Neolithic Europeans and
Eastern European hunter-gatherers is not
Table S10.5 shows that the Corded Ware can be modeled as mixtures of Middle Neolithic Europeans
and either Yamnaya or Karelia_HG. To test the robustness of this inference, we added Eurasian
hunter-gatherers and Early Neolithic farmers as additional outgroups. The full set of outgroups is:

8 Ancient Eurasian Outgroups: HungaryGamba_EN, Kostenki14, LBK_EN, Loschbour,

MA1, Motala_HG, Spain_EN, Ust_Ishim

This set includes all Upper Paleolithic Eurasians, representatives of early European farmers from
Germany, Spain, and Hungary, as well as Western and Scandinavian hunter-gatherers. We add each
of these 8 populations to the set of 15 World Outgroups, thus resulting in 8 different sets of 16
outgroups. In Table S10.6 we show the p-value for rank=1 for these different sets.

Rank=1 for models involving Karelia_HG and a Middle Neolithic population can be excluded with
p<7.64E-05 for at least one added outgroup. Rank=1 involving Yamnaya and Baalberge_MN can be
very weakly excluded (p=.0234, when adding Motala_HG as an outgroup). However, rank=1 for the
pairing (Yamnaya, Esperstedt_MN) cannot be excluded (p≥0.272 for all outgroups added), suggesting
that Corded Ware can be modeled as a simple 2-way mixture of these two groups.

To further test the choice of (Yamnaya, Esperstedt_MN), we modeled all Late Neolithic / Bronze Age
populations as such mixtures and show the p-value for rank=1 in Table S10.7. All such p-values are
≥0.017 for the Late Neolithic / Bronze Age populations of Germany, suggesting that Late
Neolithic/Bronze Age populations in general can be reasonably well-modelled as such mixtures.

Table S10.6: p-values for rank=1 of U {Corded_Ware_LN, Eastern population, MN population}

using outgroups WorldFoci15 plus an ancient Eurasian indicated in the column names.
Motala_HG
Gamba_EN
Kostenki14

Loschbour
Ust_Ishim

Spain_EN
LBK_EN

Hungary
MA1

Eastern population MN population

Karelia_HG Baalberge_MN 1.98E-01 3.91E-01 5.51E-01 3.30E-03 6.32E-03 7.64E-05 8.53E-05 1.80E-02
Karelia_HG Esperstedt_MN 7.15E-01 7.94E-01 2.88E-01 3.70E-05 4.04E-04 4.96E-03 3.24E-02 5.07E-05
Yamnaya Baalberge_MN 2.84E-01 1.93E-01 5.68E-01 2.18E-01 2.35E-01 2.99E-01 2.57E-01 2.34E-02
Yamnaya Esperstedt_MN 4.86E-01 6.44E-01 2.72E-01 4.83E-01 5.13E-01 4.79E-01 4.25E-01 3.99E-01

131
Table S10.6: p-values for rank=1 of U {Late Neolithic / Bronze Age, Yamnaya, Esperstedt_MN}
using outgroups WorldFoci15 plus an ancient Eurasian indicated in the column names.

Motala_HG
Gamba_EN
Kostenki14

Loschbour
Ust_Ishim

Spain_EN
LBK_EN

Hungary
MA1
Late Neolithic / Bronze Age population
Karsdorf_LN 8.98E-01 8.40E-01 7.98E-01 9.02E-01 8.70E-01 9.00E-01 9.12E-01 8.67E-01
Corded_Ware_LN 4.86E-01 6.44E-01 2.72E-01 4.83E-01 5.13E-01 4.79E-01 4.25E-01 3.99E-01
Bell_Beaker_LN 2.70E-01 1.28E-01 4.98E-01 1.67E-02 4.35E-02 1.54E-01 1.93E-01 2.10E-01
Alberstedt_LN 2.34E-01 2.71E-01 3.87E-01 1.65E-01 2.65E-01 2.02E-01 1.97E-01 5.74E-02
BenzigerodeHeimburg_LN 8.07E-02 8.19E-02 1.56E-01 3.32E-02 4.90E-02 5.65E-02 5.46E-02 7.22E-02
Unetice_EBA 3.57E-01 1.95E-01 6.88E-01 2.48E-01 2.25E-01 3.56E-01 1.49E-01 1.74E-01
Halberstadt_LBA 1.34E-01 4.84E-02 2.69E-01 1.45E-01 1.42E-01 1.40E-01 3.27E-02 1.44E-01
HungaryGamba_BA 8.66E-01 7.79E-01 9.49E-01 8.43E-03 3.66E-03 6.02E-01 8.13E-01 2.32E-01
We estimate mixture proportions for all Late Neolithic / Bronze Age of Germany populations in Table
S10.8, showing estimates when using 15 world outgroups, 8 ancient Eurasians as outgroups, and 23
outgroups (combining both of the above).

Table S10.8: Yamnaya mixture proportions in Late Neolithic / Bronze Age Germans
LN/BA population WorldFoci15 Outgroups 8 Ancient Outgroups All 23 Outgroups
Halberstadt_LBA 0.556 +/- 0.108 0.488 +/- 0.041 0.520 +/- 0.032
Alberstedt_LN 0.422 +/- 0.098 0.568 +/- 0.040 0.531 +/- 0.034
Bell_Beaker_LN 0.387 +/- 0.071 0.534 +/- 0.030 0.555 +/- 0.023
BenzigerodeHeimburg_LN 0.472 +/- 0.097 0.650 +/- 0.032 0.618 +/- 0.026
Unetice_EBA 0.543 +/- 0.062 0.632 +/- 0.024 0.625 +/- 0.019
Corded_Ware_LN 0.764 +/- 0.066 0.743 +/- 0.027 0.731 +/- 0.022
Karsdorf_LN 0.823 +/- 0.128 0.821 +/- 0.081 0.745 +/- 0.064

The standard errors are highest with the WorldFoci15 set of outgroups, as one might expect given that
these populations share little genetic drift with those of Europe. They are smaller when using Ancient
Eurasians as outgroups, and smaller still when combining both sets. The standard errors are also larger
for populations that consist of single individuals (such as Halberstadt_LBA or Karsdorf_N), as one
might expect due to the limited data.

Overall, there is reasonable agreement between the inferences using the two disjoint sets of outgroups,
which increases our confidence that our inference that Late Neolithic / Bronze Age populations were
formed by this type of mixture between local Middle Neolithic and migrant Yamnaya-related
ancestors. All estimates agree that a sizeable portion of the ancestry of LN/BA populations is related
to the Yamnaya, and quantify this portion to ~3/4 for the Corded Ware, in agreement with the
inference of SI9 with a related method and the results of PCA and ADMIXTURE analysis (Fig. 2)
and FST analysis (Extended Data Table 3) that also reveal the close relationship between the Corded
Ware of Late Neolithic Germany with the Yamnaya population of the Samara district sampled
~2,600km to the east.

132
References

1. Reich, D. et al. Reconstructing Native American population history. Nature 488, 370-374,
(2012).
2. Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065-1093, (2012).
3. Kunsch, H. R. The Jackknife and the Bootstrap for General Stationary Observations. Ann.
Statist. 17, 1217-1241, (1989).
4. Patterson, N. et al. Genetic structure of a unique admixed population: implications for
medical research. Hum. Mol. Genet. 19, 411-419, (2010).

133
Supplementary Information 11
Relevance of ancient DNA to the problem of Indo-European language dispersals

Iosif Lazaridis, Wolfgang Haak, Nick Patterson, David Anthony and David Reich*
* To whom correspondence should be addressed ([email protected])

Main hypotheses of Indo-European language dispersals

Indo-European languages are widely spoken from the Atlantic Ocean to the Indian subcontinent1, and,
through more recent migrations, have expanded into much of the rest of the world. The set of living
Indo-European languages includes, among others, English, Greek, Russian, Italian, Farsi, and Hindi,
and is complemented by a number of extinct languages like Hittite and later Phrygian in Anatolia,
Tocharian in the Xinjiang province of China2,3, and the languages of the Scythians of the Eurasian
steppe. With the advent of the written word and modern means of travel and communication, it has
become possible to acquire a new language by cultural transmission alone. However, in the past,
languages must have spread through direct contact4 and, at least to some extent, with migrants, and
thus the study of ancient DNA, which can trace the migration of people, is relevant to evaluating
models of language dispersal.

Ancient DNA can roll back the clock and reveal earlier stages in the formation of human populations,
just as comparative linguistics can reveal earlier stages in the formation of human languages. Most
experts agree that Proto-Indo-European, a reconstructed ancestral language to living and historically
attested Indo-European languages, must have once been spoken in a geographically circumscribed and
relatively small part of Eurasia1,5-9. Many theories have been proposed since the discovery of the Indo-
European language family in the eighteenth century as to where that homeland was and how the
language was transmitted. Such theories invariably postulate at least some migrations of people.
These hypotheses can be tested with the methods of genetics applied to ancient individuals.

The two leading theories of Indo-European language dispersal are:

1. The “Steppe hypothesis”8,10-14, which derives Indo-European languages from the people
buried under kurgans on the Pontic-Caspian steppe in modern Ukraine and southern Russia.

2. The “Anatolian hypothesis”9,15, which suggests that early farmers who migrated from Anatolia
to Europe during the Early Neolithic brought Indo-European languages with them.

Other hypotheses have also been formulated and we will discuss the following two where relevant:

3. The “Balkan hypothesis”6 that resembles the Anatolian hypothesis in linking the migration of
Indo-Europeans to farmers, but with the difference that Indo-European languages spread from
southeastern Europe at a later period.

4. The “Armenian plateau hypothesis”7,16 which resembles the Steppe hypothesis in postulating a
role for the steppe in the dispersal of languages into Europe, but places the homeland of Proto-
Indo-European speakers south of the Caucasus.

134
A key argument in favor of the Anatolian hypothesis has been that the migration of early farmers from
Anatolia ~8,000-7,000 years ago was a major transformation that affected most of Europe. A massive
population turnover is the most widely accepted and easiest explanation for the introduction of a new
language across a large geographic area. There are of course other ways by which languages can
spread, such as elite dominance7, whereby a politically dominant minority transmits its language to
the resident majority (e.g. Hungarian). Another argument for the Anatolian hypothesis is the inferred
topology of the Indo-European family tree, in which Anatolian languages like Hittite occupy a basal
position, which could plausibly be explained by an Anatolian homeland and migrations from Anatolia
to Europe through the Balkans17. Finally, the recent application of phylogenetic methods adopted
from biology to the estimation of language divergence times18,19 has been used to argue in favor of an
earlier breakup of Proto-Indo-European into daughter languages.

The Steppe hypothesis dates the breakup of Proto-Indo-European to several millennia later: to 4,000-
3,000 BCE. One of the prominent arguments in favor of the Steppe hypothesis is that reconstructions
of the ancestral Proto-Indo-European language have suggested a shared vocabulary associated with
wheeled vehicles, with cognates in Indo-European languages ranging from Europe to India, although
not all linguists agree that it is possible to make meaningful inferences about the material culture of an
ancient people based on descendant languages20. If this reconstruction is right, Proto-Indo-European
could not have been the language spread by the first farmers moving into Europe, as wheels and
wheeled vehicles were only invented thousands of years after the spread of farming. While the earliest
finds of wheels are not from the steppe21, the intensification of the use of wheeled vehicles and
domestic horses is linked to the spread of steppe pastoralism. Another attractive feature of the Steppe
hypothesis is that the steppe geography could account for the evidence of loan words between Proto-
Indo-European and Uralic languages to the north, and Kartvelian languages to the south5.

Hypotheses of Indo-European origins in light of the new genetic data presented in this paper
Genetic data is a valuable source of information that is useful for evaluating competing hypotheses of
Indo-European language dispersals, to the extent that theories of such dispersals invoke migratory
movements to explain their extensive distribution. Such hypotheses can be tested both with
archaeology (which tests for the spread of material culture, which can spread not only by migration
but also through trade or an exchange of ideas), and with genetics (which can test directly whether the
movement of people accompanied perceived changes in the material record).

Past genetic data from ancient DNA has confirmed one of the major predictions of the Anatolian
hypothesis – the migration of early farmers from the Near East (inclusive of Anatolia) to Europe –
using both mitochondrial DNA22-24 and whole genome analysis25,26. The results of our study are
consistent with these findings, and also extend them by showing that not only the early farmers of
central Europe (Germany and Hungary) and Scandinavia, but also of Iberia were descended from a
common stock. In this sense, ancient DNA is consistent with migrations following the predictions of
the Anatolian hypothesis, and indeed our ancient DNA results match the scenario outlined by
Bellwood for the initial dispersal of farming into Europe remarkably well27. This is also true for the
Balkan hypothesis, as geographically, southeastern Europe is a plausible place where early farmers
could have diverged into an inland Danubian route toward central Europe, and a Mediterranean route
toward Iberia. The evidence of a relatively homogeneous population of early European farmers with
substantial Near Eastern ancestry25 is indeed a reasonable candidate for the spread of a single
language family across Europe.

135
Our new genetic data are important in showing that a second major migration from the steppe into
Europe occurred at the end of the Neolithic period (between 5000-4500 years ago). Moreover, we
have demonstrated that these migrants accounted for at least ~3/4 of the ancestry of the Corded Ware
people of Germany, and much of the ancestry of other Late Neolithic / Bronze Age populations of
Germany and present-day northern Europeans (Fig. 3, SI9, SI10). Thus, the main argument in favor of
the Anatolian hypothesis (that major language change requires major migration) can now also be
applied to the Steppe hypothesis. While we cannot go back in time to learn what languages the
migrants spoke, it seems more likely than not that the Corded Ware people we sampled spoke the
languages of the people who contributed the great majority of their ancestry (Yamnaya), rather than
the local languages of the people who preceded them. Thus, our results increase the plausibility that
the Corded Ware people and those genetically similar groups who followed them in central Europe
spoke a steppe-derived Indo-European language. More generally, our results level the playing field
between the two leading hypotheses of Indo-European origins, as we now know that both the Early
Neolithic and the Late Neolithic were associated with major migrations.

While our results do not settle the debate about the location of the proto-Indo European homeland,
they increase the plausibility of some hypotheses and decrease the plausibility of others as follows:

1. The Steppe hypothesis gains in plausibility by our discovery of a migration during the Late
Neolithic from the steppe into central Europe. This migration was predicted by some
proponents of the Steppe hypothesis and we have now shown (definitively) that it occurred.
We also note that our results help to differentiate between variants of the steppe hypothesis:
we do not find evidence of an influence of steppe migrants earlier than the Corded Ware,
although we cannot rule out the possibility that such evidence might be found with larger
sample sizes and more sampling locations in central Europe. However, we can definitely
reject that the breakup of Indo-European occurred as late as 4000 years ago28, as by ~4500
years ago the migration into Europe had already taken place. Moreover, this migration clearly
resulted in a large population turnover, meaning that the Steppe hypothesis does not require
elite dominance9 to have transmitted Indo-European languages into Europe. Instead, our
results show that the languages could have been introduced simply by strength of numbers:
via major migration in which both sexes participated (SI2, SI4)

2. The Anatolian hypothesis becomes less plausible as an explanation for the origin of all Indo-
European languages in Europe, as it can no longer claim to correspond to the only major
population transformation in European prehistory, and it must also account for the language
of the steppe migrants. However, the Anatolian hypothesis cannot be ruled out entirely by our
data, as it is possible that it still accounts for some of the major branches of the Indo-
European language family in Europe, especially the branches of the south where the
proportion of steppe ancestry today is smaller than in central and northern Europe (Figure 3).

3. The Balkan hypothesis faces similar difficulties as the Anatolian. If the early farmers of
southeastern Europe were genetically similar to their descendants in central and western
Europe, a spread of Indo-European speaking migrants from the Balkans to the rest of Europe
would simply introduce another layer of “Early Neolithic” genes similar to those present
elsewhere in Europe, but would not account for the migration from the steppe and its
associated language. Furthermore if the steppe immigrants spoke Indo-European languages,
these languages are unlikely to have been acquired by migration from Europe, as our

136
 Yamnaya samples show no sign of a major component of ancestry derived from European
Early or Middle Neolithic farmers (Fig. 2).

4. The Armenian plateau hypothesis gains in plausibility by the fact that we have discovered
evidence of admixture in the ancestry of Yamnaya steppe pastoralists, including gene flow
from a population of Near Eastern ancestry for which Armenians today appear to be a
reasonable surrogate (SI4, SI7, SI9). However, the question of what languages were spoken
by the “Eastern European hunter-gatherers” and the southern, Armenian-like, ancestral
population remains open. Examining ancient DNA from the Caucasus and Near East may be
able to provide further insight about the dynamics of the interaction between these regions
and the steppe. Our results show that southern populations diluted the ancestry of populations
from the steppe, but also that ancestry related to Ancient North Eurasians forms a major
ancestral component of the populations of the present-day Caucasus25. Thus, both south-north
and north-south genetic influence across the Caucasus is plausible.

Pitfalls in using genetic data to make inferences about language spread

The study of Indo-European origins and language dispersals has been controversial, in part because of
the history of misuse of the concept of the Proto-Indo-European homeland for ideological reasons29.

In the early 20th century, Gustav Kossinna proposed the idea of ‘settlement archaeology’: that a
material culture identified by archaeology, specifically the Corded Ware, might correspond to a
genetically well-defined people and homogeneous language group, specifically the Proto-Indo-
Europeans. Our data directly contradict Kossinna’s theories in showing that the Corded Ware are not
a locally derived central European population but instead are to a significant degree descended from
eastern migrants. V. Gordon Childe12, following linguistic arguments by Otto Schrader14, proposed a
migration from the steppe into Europe, which seems, in view of the results of our study, to have been
closer to the mark. However, following the Second World War, and especially in the 1960s and
1970s, archaeologists responded to the history of misuse of archaeology by rejecting sweeping
migrations and the settlement archaeology framework altogether30, and suggesting that in practice, it
would not ever be possible to show that archaeological, linguistic, and genetic groupings overlap31,32.
This climate in the archaeological community has made it challenging to propose migration as an
explanation for similarities or differences in material culture across time and space. Although
migration is today accepted more widely by archaeologists33 than it was 30 years ago11, it is usually
discussed in connection with demographic-growth models linked to the expansion of agriculture27,
while migrations linked to the evolution of new socio-political structures among long-established
food-producing populations are less understood, less recognizable and often viewed with skepticism.

Genetic data is important to this debate as it changes the equation, bringing to bear a new type of data
that can directly speak to whether or not migration occurred. This type of fact could never be clearly
established before the advent of ancient DNA, except by use of stable isotope analysis which only
works to detect migration if the studied samples are from the first generation of migrants. While
migration from the steppe had been proposed on the basis of human skeletal morphological data by
Childe and others, a more recent study34, which dealt with precisely the problem of relationships
between the Corded Ware and steppe Kurgan (Pit Grave) groups, concluded that “The local [Corded
Ware] groups of the core area (Central Germany, Czechoslovakia, Poland) form a very homogeneous
block, issued from the local "Old Europe" substratum and persisting until Aunjetitz (Unetice) at least.
This block show[s] no biological affinities to the Ukrainian Kurgan populations. There is no evidence
for physical presence of Kurgan tribes in this area.” Similarly, on the basis of archaeology it has been

137
claimed by a key proponent of the steppe hypothesis8 that “there is no real case for an expansion of
Yamnaya invaders across the North European plain, producing the Corded Ware horizon”,
highlighting how controversial the idea of migration has become. With ancient DNA, one can
establish whether or not migrations occurred, and tie migrations to well-defined archaeological
cultures by sequencing DNA from radiocarbon dated skeletons buried with diagnostic grave goods.
Thus, it is possible, using ancient DNA, to evaluate directly whether a particular material culture
could have spread through migration or whether cultural transmission occurred. Our paper has now
made an overwhelming case for just such a migration, showing the power of genetics to contribute
meaningfully to debates about past cultural interactions.

An important caveat to using ancient DNA to make arguments about the origins of languages is that
prior to the invention of writing, we have no way to directly tie ancient cultures to a language.
Nevertheless, by establishing that major migrations or exchanges of genes occurred, we identify
movements of people that would have been plausible vectors for the spread of languages, and we can
establish some periods in time as the most plausible ones for language spread. Thus, genetic data can
change the balance of probabilities among competing hypotheses as we outline above.

Although in this study we have focused on the genetic findings, our data are also interesting from the
point of view of archaeological methodology. Specifically, our findings challenge the idea of a limited
role of migration in human population history11 by providing unambiguous evidence of two major
episodes of migration and population turnover in Europe. By documenting not only that these major
migrations occurred, but that they were both followed to a degree by the genetic resurgence of the
local populations (SI7, Fig. 3), we hope that our study will help to spur new debate on the interactions
between migrants and indigenous peoples long after the occurrence of migration.

Prospects for further genetic insight into Indo-European language origins

The ultimate question of the Proto-Indo-European homeland is unresolved by our data.

One important future direction for genetic research into Indo-European origins is to obtain ancient
DNA data from India, Iran, northwestern China, and intervening regions to test hypotheses about the
spread of Indo-European languages to the east. While obtaining such data is likely to be more difficult
than European ancient DNA work (because of poorer conditions for preservation), it could provide
crucial clues. For example, if the people who brought substantial steppe ancestry to northern Europe
(via the Yamnaya and Corded Ware) also spread Indo-European languages into Iran and India, we
should expect to find steppe ancestry in prehistoric samples from the Iranian plateau and South Asia,
unless the mode of language transgression was different there (e.g. elite dominance).

A second direction for future genetic research is to study additional ancient European populations.
Present-day populations of southern Europe have lower ancestry related to the Yamnaya than those
from northern Europe (Fig. 3). This could be explained either by migration from the Yamnaya that
affected southern Europe to a lesser extent than northern Europe, the later dilution of this ancestry in
southern Europe, or the later introduction of such ancestry only indirectly, from northern Europe as a
result of much later admixture events35,36. It is still possible that the steppe migration detected by our
study into Late Neolithic Europe might account for only a subset of Indo-European languages in
Europe, and other Indo-European languages arrived in Europe not from the steppe but from either an
early “Neolithic Anatolian” or later “Armenian plateau” homeland. We highlight Anatolia and the
Balkans as particularly promising places to study to make further progress toward understanding these
open questions. The four different theories of Indo-European origins make very different predictions

138
about the population history of Anatolia and southeastern Europe where two of the earliest attested
Indo-European languages, Hittite and Mycenaean Greek are found. By examining ancient samples
from these regions it should become possible to determine if there are genetic discontinuities prior to
the appearance of these languages that may be correlated with the migration of a new population.
Ancient DNA provides evidence independent of archaeology or linguistics, and thus has the potential
to resolve continuing controversies and to contribute to progress in answering the centuries’ old riddle
of Indo-European origins.

References

1. Mallory, J. P. & Adams, D. Q. The Oxford Introduction to Proto-Indo-European and the

Proto-Indo-European World. (Oxford University Press, 2006).
2. Mair, V. H. Genes, geography, and glottochronology: the Tarim Basin during late prehistory
and history. The Journal of Indo-European Studies Monograph Series. No. 50, 1-47, (2005).
3. Mallory, J. P. & Mair, V. H. The Tarim Mummies: Ancient China and the Mystery of the
Earliest Peoples from the West. (Thames & Hudson, 2000).
4. Thomason, S. G. & Kaufman, T. Language contact, creolization and genetic linguistics.
(University of California, Press, 1988).
5. Anthony, D. W. The Horse, the Wheel, and Language: How Bronze-Age Riders from the
Eurasian Steppes Shaped the Modern World. (Princeton University Press, 2007).
6. Diakonov, I. M. On the original home of the speakers of Indo-European. Journal of Indo-
European Studies 13, 92-174, (1985).
7. Gamkrelidze, T. V. & Ivanov, V. V. The Early History of Indo-European Languages. Sci. Am.
262, 110-116, (1990).
8. Mallory, J. P. In Search of the Indo-Europeans: Language, Archaeology and Myth. (Thames
and Hudson, 1991).
9. Renfrew, C. Archaeology and Language: The Puzzle of Indo-European Origins. (Pimlico,
1987).
10. Anthony, D. W. The "Kurgan Culture," Indo-European Origins, and the Domestication of the
Horse: A Reconsideration. Curr. Anthropol. 27, 291-313, (1986).
11. Anthony, D. W. Migration in Archaeology: the Baby and the Bathwater. American
Anthropologist 92, 895-914, (1990).
12. Childe, V. G. The Aryans: A study of Indo-European origins. (Kegan Paul, Trench, Trubner,
1926).
13. Gimbutas, M. Vol. Bulletin 20 (American School of Prehistoric Research, Harvard
University, 1956).
14. Schrader, O. Prehistoric antiquities of the Aryan peoples: A manual of comparative philology
and the earliest culture. (Charles Griffin, 1890).
15. Bellwood, P. First Farmers: The Origins of Agricultural Societies. (Wiley-Blackwell, 2004).
16. Grigoriev, S. A. Ancient Indo-Europeans. (Rifei, 2002).
17. Renfrew, C. in The Bronze Age & Early Iron Age Peoples of Eastern Central Asia (ed V. H.
Mair) (Journal of Indo-European Studies Monograph #26, Vol. 1, 1998).
18. Bouckaert, R. et al. Mapping the Origins and Expansion of the Indo-European Language
Family. Science 337, 957-960, (2012).
19. Gray, R. D. & Atkinson, Q. D. Language-tree divergence times support the Anatolian theory
of Indo-European origin. Nature 426, 435-439, (2003).
20. Heggarty, P. in The Encyclopedia of Global Human Migration (ed I. Ness) (Blackwell
Publishing Ltd., 2013).
21. Menotti, F. Living on the Lake in Prehistoric Europe. (Routledge, 2004).
22. Bramanti, B. et al. Genetic Discontinuity Between Local Hunter-Gatherers and Central
Europe’s First Farmers. Science 326, 137-140, (2009).
23. Brandt, G. et al. Ancient DNA Reveals Key Stages in the Formation of Central European
Mitochondrial Genetic Diversity. Science 342, 257-261, (2013).

139
24. Haak, W. et al. Ancient DNA from European Early Neolithic Farmers Reveals Their Near
Eastern Affinities. PLoS Biol 8, e1000536, (2010).
25. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-
day Europeans. Nature 513, 409-413, (2014).
26. Skoglund, P. et al. Origins and Genetic Legacy of Neolithic Farmers and Hunter-Gatherers in
Europe. Science 336, 466-469, (2012).
27. Bellwood, P. et al. Examining the farming/language dispersal hypothesis. (McDonald Inst.
for Archaeological Research, 2002).
28. Beckwith, C. I. Empires of the Silk Road: A History of Central Eurasia from the Bronze Age
to the Present. (Princeton University Press, 2011).
29. Poliakov, L. Le mythe aryen - Essai sur les sources du racisme et des nationalismes.
(Calmann-Lévy, 1971).
30. Renfrew, C. & Bahn, P. G. Archaeology : theories, methods and practice. 5. edn, (Thames &
Hudson, 2008).
31. Lawler, A. From the Horse's Mouth. Humanities, 11-13, (2008).
32. Soares, P. et al. The archaeogenetics of Europe. Current Biology 20, R174-183, (2010).
33. Burmeister, S. Archaeology and migration: approaches to an archaeological proof of
migration. Current Anthropology 41, 554-555, (2000).
34. Menk, R. A Synopsis of the Physical Anthropology of the Corded Ware Complex on the
Background of the Expansion of the Kurgan Cultures. Journal of Indo-European Studies 8,
361-392, (1980).
35. Hellenthal, G. et al. A genetic atlas of human admixture history. Science 343, 747-751,
(2014).
36. Ralph, P. & Coop, G. The Geography of Recent Genetic Ancestry across Europe. PLoS Biol
11, e1001555, (2013).

140