Population genomics of the Viking world

bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 1 Population genomics of the Viking world 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Ashot Margaryan1,2,3*, Daniel Lawson4*, Martin Sikora1*, Fernando Racimo1*, Simon Rasmussen5, Ida Moltke6, Lara Cassidy7, Emil Jørsboe6, Andrés Ingason1,58,59, Mikkel Pedersen1, Thorfinn Korneliussen1, Helene Wilhelmson8,9, Magdalena Buś10, Peter de Barros Damgaard1, Rui Martiniano11, Gabriel Renaud1, Claude Bhérer12, J. Víctor Moreno-Mayar1,13, Anna Fotakis3, Marie Allen10, Martyna Molak14, Enrico Cappellini3, Gabriele Scorrano3, Alexandra Buzhilova15, Allison Fox16, Anders Albrechtsen6, Berit Schütz17, Birgitte Skar18, Caroline Arcini19, Ceri Falys20, Charlotte Hedenstierna Jonson21, Dariusz Błaszczyk22, Denis Pezhemsky15, Gordon Turner-Walker23, Hildur Gestsdóttir24, Inge Lundstrøm3, Ingrid Gustin8, Ingrid Mainland25, Inna Potekhina26, Italo Muntoni27, Jade Cheng1, Jesper Stenderup1, Jilong Ma1, Julie Gibson25, Jüri Peets28, Jörgen Gustafsson29, Katrine Iversen5,64, Linzi Simpson30, Lisa Strand18, Louise Loe31,32, Maeve Sikora33, Marek Florek34, Maria Vretemark35, Mark Redknap36, Monika Bajka37, Tamara Pushkina15, Morten Søvsø38, Natalia Grigoreva39, Tom Christensen40, Ole Kastholm41, Otto Uldum42, Pasquale Favia43, Per Holck44, Raili Allmäe28, Sabine Sten45, Símun Arge46, Sturla Ellingvåg1, Vayacheslav Moiseyev47, Wiesław Bogdanowicz14, Yvonne Magnusson48, Ludovic Orlando49, Daniel Bradley7, Marie Louise Jørkov50, Jette Arneborg40,63, Niels Lynnerup50, Neil Price21, M. Thomas Gilbert3,51, Morten Allentoft1, Jan Bill52, Søren Sindbæk53, Lotte Hedeager54, Kristian Kristiansen55, Rasmus Nielsen1,56†, Thomas Werge1,57,58,59†, Eske Willerslev1,60,61,62† 1 Lundbeck Foundation GeoGenetics Centre, GLOBE Institute, University of Copenhagen, Øster Voldgade 5-7, 1350 Copenhagen K, Denmark. 2Institute of Molecular Biology, National Academy of Sciences, 7, Hasratian St., 0014, Yerevan, Armenia. 3Section for Evolutionary Genomics, GLOBE Institute, University of Copenhagen, Øster Voldgade 5-7, 1350 Copenhagen K, Denmark. 4MRC Integrative Epidemiology Unit, University of Bristol, Bristol, UK. 5Novo Nordisk Foundation Center for Protein Research, Faculty of Health and Medical Sciences, University of Copenhagen, Blegdamsvej 3B, 2200 Copenhagen, Denmark. 6Department of Biology, The Bioinformatics Centre, University of Copenhagen, 2200 Copenhagen N, Denmark. 7Smurfit Institute of Genetics, Trinity College Dublin, Dublin. 8Historical archaeology, Department of Archaeology and Ancient history, Lund University, PB 192, SE 22100 Lund, Sweden. 9Sydsvensk arkeologi AB, PB 134, SE 29122 Kristianstad, Sweden. 10Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala University, 751 08 Uppsala, Sweden. 11Department of Genetics, University of Cambridge, Downing Street, Cambridge CB2 3EH, UK. 12New York Genome Center, 101 Avenue of the Americas, New York, NY, USA, 10013. 13National Institute of Genomic Medicine (INMEGEN), Periférico Sur 4809, 14610 Mexico City, Mexico. 14Museum and Institute of Zoology, Polish Academy of Sciences, Wilcza 64, 00-679 Warsaw, Poland. 15Anuchin Research Institute and Museum of Anthropology, Moscow State University. 16Manx National Heritage, Kingswood Grove, Douglas, Isle of Man, British Isles IM1 3LY. 17Upplandsmuseet, Drottninggatan 7, 753 10 Uppsala, Sweden. 18 NTNU University Museum, Department of Archaeology and Cultural History Norway. 19 Arkeologerna. 20Thames Valley Archaeological Services (TVAS), Reading, UK. 21Department of Archaeology and Ancient History, Uppsala University, Box 626, 751 26 Uppsala, Sweden. 22Institute of Archaeology, University of Warsaw, ul. Krakowskie Przedmieście 26/28, 00-927 Warsaw, Poland. 23 Department of Cultural Heritage Conservation, National Yunlin University of Science and Technology, Douliou, Taiwan. 24Institute of Archaeology, Iceland. Bárugata 3, 101 Reykjavík, Iceland. 1 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 25 UHI Archaeology Institute, University of the Highlands and Islands, Orkney College, Kirkwall, Orkney, KW15 1LX. 26Department of Bioarchaeology, Institute of Archaeology of Natioanal Academy of Sciences of Ukraine, 12 Geroiv Stalingrada Ave. 04210 Kyiv, Ukraine. 27Soprintendenza Archeologia, Belle Arti e Paesaggio per le Province di Barletta - Andria - Trani e Foggia, Via Alberto Alvarez Valentini, 8 - 71121 Foggia, Italy. 28Archaeological Research Collection, Tallinn University, Rüütli 10, Tallinn 10130, Estonia. 29Jönköping county museum, Jönköping, Sweden. 30Trinity College Dublin. 31Oxford Archaeology, Janus House, Osney Mead, Oxford OX2 0ES, UK. 32Heritage Burial Services, Oxford Archaeology, Janus House, Osney Mead, Oxford OX2 0ES, UK. 33National Museum of Ireland, Kildare Street, Dublin 2, Ireland. 34Institute of Archaeology, Maria Curie-Sklodowska University in Lublin, Pl. M. Curie-Sklodowska 4, 20-035 Lublin, Poland. 35Västergötlands museum, Box 253, 532 23 Skara Sweden. 36National Museum Cardiff. 37"Trzy Epoki" Archaeological Service, Poland. 38Museum of Southwest Jutland. 39Institute for the history of material culture, Russian Academy of Sciences, Dvotsovaya Emb., 18, Saint-Petersburg, Russia, 191186. 40National Museum of Denmark, Frederiksholms Kanal 12, DK-1220 Copenhagen, Denmark. 41Roskilde Museum, Museum Organization ROMU, Sankt Ols Stræde 3, DK-4000 Roskilde, Denmark. 42Langelands Museum, Jens Winthersvej 12. 5900 Rudkøbing, Langeland, Denmark. 43Department of Humanities, University of Foggia, Via Arpi, 176, 71121 Foggia, Italy. 44Department of Molecular Medicine, Faculty of Medicine, University of Oslo. 45Department of Archaeology and Ancient History, Uppsala University Campus Gotland. 46Tjóðsavnið - Faroe Islands National Museum. Kúrdalsvegur 15. Postboks 1155. FO-110 Tórshavn. 47Peter the Great Museum of Anthropology and Ethnography (Kunstkamera), Russian Academy of Science, University Emb, 3, SPb, Russia, 199034. 48Malmö Museum, Box 406, 201 24 Malmö, Sweden. 49Laboratoire d’Anthropobiologie Moléculaire et d’Imagerie de Synthèse, CNRS UMR 5288, Université de Toulouse, Université Paul Sabatier, 31000 Toulouse, France. 50Department of Forensic Medicine, University of Copenhagen, Frederik V's vej 11, 2100 Copenhagen. 51Department of Natural History, NTNU. 52Museum of Cultural History, University of Oslo, P.O. Box 6762 St. Olavs plass, 0160 Oslo, Norway. 53Centre for Urban Network Evolutions (UrbNet), Aarhus University, School of Culture and Society, Moesgård Allé 20, building 4215, DK-8270 Højbjerg, Denmark. 54Institute of Archaeology, Conservation and History, Pb. 1019 Blindern, 0315 Oslo, Norway. 55Department of Historical Studies, University of Gothenburg. 56 Departments of Integrative Biology and Statistics, UC Berkeley, Berkeley, CA 94720, USA. 57 Department of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark. 58Institute of Biological Psychiatry, Mental Health Services Copenhagen, Copenhagen, Denmark. 59The Lundbeck Foundation Initiative for Integrative Psychiatric Research, iPSYCH, Denmark. 60Department of Zoology, University of Cambridge, UK. 61The Danish Institute for Advanced Study, University of Southern Denmark. 62The Wellcome Trust Sanger Institute, Cambridge, UK. 63School of GeoSciences, University of Edinburgh. 64Department of Health Technology, Section for Bioinformatics, Technical University of Denmark, DTU, 2800 Kgs. Lyngby, Denmark *These authors contributed equally to this work. † e-mail: [email protected]; [email protected]; [email protected] 2 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 87 Abstract 88 The Viking maritime expansion from Scandinavia (Denmark, Norway, and Sweden) marks one 89 of the swiftest and most far-flung cultural transformations in global history. During this time 90 (c. 750 to 1050 CE), the Vikings reached most of western Eurasia, Greenland, and North 91 America, and left a cultural legacy that persists till today. To understand the genetic structure 92 and influence of the Viking expansion, we sequenced the genomes of 442 ancient humans from 93 across Europe and Greenland ranging from the Bronze Age (c. 2400 BC) to the early Modern 94 period (c. 1600 CE), with particular emphasis on the Viking Age. We find that the period 95 preceding the Viking Age was accompanied by foreign gene flow into Scandinavia from the 96 south and east: spreading from Denmark and eastern Sweden to the rest of Scandinavia. 97 Despite the close linguistic similarities of modern Scandinavian languages, we observe genetic 98 structure within Scandinavia, suggesting that regional population differences were already 99 present 1,000 years ago. We find evidence for a majority of Danish Viking presence in England, 100 Swedish Viking presence in the Baltic, and Norwegian Viking presence in Ireland, Iceland, and 101 Greenland. Additionally, we see substantial foreign European ancestry entering Scandinavia 102 during the Viking Age. We also find that several of the members of the only archaeologically 103 well-attested Viking expedition were close family members. By comparing Viking Scandinavian 104 genomes with present-day Scandinavian genomes, we find that pigmentation-associated loci 105 have undergone strong population differentiation during the last millennia. Finally, we are able 106 to trace the allele frequency dynamics of positively selected loci with unprecedented detail, 107 including the lactase persistence allele and various alleles associated with the immune response. 108 We conclude that the Viking diaspora was characterized by substantial foreign engagement: 109 distinct Viking populations influenced the genomic makeup of different regions of Europe, 110 while Scandinavia also experienced increased contact with the rest of the continent. 111 3 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 112 Introduction 113 Three centuries from approximately 750 to 1050 CE mark a pivotal change for the peoples of 114 Scandinavia. The maritime transformation commonly known as the Viking Age (VA) altered the 115 political, cultural and demographic map of Europe in ways that are evident even today. The Vikings 116 established systems of trade and settlement that stretched from the eastern American seaboard to the 117 Asian steppe1. They also exported new ideas, technologies, language, beliefs and practices to these 118 lands. In the process, they gradually developed new socio-political structures, assimilated cultural 119 influences, and adopted the Christian faith2. 120 121 Currently, most of our understanding of the VA is based on written sources and archaeological 122 evidence. The VA as a historical period has been framed by the first clearly documented raid on 123 Lindisfarne in 793 CE, and the defeat of a Norwegian army at Stamford Bridge in 1066 CE. More 124 recent perspectives emphasize long-term, multi-causal social processes with after-effects that varied 125 greatly by region3–5. Similarly, the notion of a Viking ‘expansion’, implying deliberate drive and 126 purpose, has been supplemented by the more fluid concept of a ‘diaspora’ that developed over time2. 127 Under this framework, however, the role of demographic dynamics has remained unclear, as has the 128 question of whether VA Scandinavia was genetically structured or represented a homogenous 129 population. Similarly, we still do not know to what extent Vikings mixed with local populations they 130 encountered and how much foreign ancestry was brought back to Scandinavia. 131 In order to explore the genomic history of the Viking era, we shotgun sequenced 442 ancient human 132 remains, from the Bronze Age c. 2400 BC to the Medieval Age c. 1600 AD (Fig. 1). The majority of 133 these individuals (n=376) were sequenced to between 0.1 and 11X average depth of coverage. The 134 dataset includes Bronze Age (n=2) and Iron Age (n=10) individuals from Scandinavia; Early Viking 135 Age (n=43) individuals from Estonia (n=34), Denmark (n=6) and Sweden (n=3); ancient individuals 136 associated with Norse culture from Greenland (n=23), VA individuals from Denmark (n=78), Faroe 137 Islands (n=1), Iceland (n=17), Ireland (n=4), Norway (n=29), Poland (n=8), Russia (n=33), Sweden 138 (n=118), UK (n=42), Ukraine (n=3) as well as medieval individuals from Faroe Islands (n=16), Italy 139 (n=5), Norway (n=7), Poland (n=2) and Ukraine (n=1). The VA individuals were supplemented with 140 additional published genomes (n=21) from Sigtuna, in Sweden6. The skeletons originate from major 141 archaeological sites of VA Scandinavian settlements and activities from Europe to Greenland 142 (Supplementary Table 1). The data from the ancient individuals were analyzed together with 4 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 143 previously published data from a total of 3,855 present-day individuals across two reference panels, 144 and data from 922 individuals of ancient origin (Supplementary Note 6). 145 146 Scandinavian genetic ancestry and the beginnings of the Viking era 147 Although VA Scandinavians shared a common cultural, linguistic and material background, there 148 was no common word for Scandinavian identity at that time1. The word ‘Viking’ is used in 149 contemporary sources to mean a ‘pirate’ or ‘sea warrior’2. As such, there is no single ‘Viking world’, 150 but a coalescence of ‘Viking worlds’ marked by rapidly growing maritime exploration, trade, war 151 and colonization, following the adoption of deep-sea navigation among the coastal populations of 152 Scandinavia and the Baltic Sea area7,8. Thus, it is unclear whether the Viking-phenomenon refers to 153 people with a recently shared genetic background and if foreign influence initiated or accompanied 154 the transition from the Scandinavian Iron Age into the Viking era. 155 To assess the genetic relationship of the VA Scandinavians with that of earlier European peoples, we 156 performed genetic clustering using multi-dimensional scaling (MDS) on a pairwise identity-by-state 157 (IBS) sharing matrix, as well as latent mixed-ancestry models (Admixture)9. We find that the majority 158 of our samples broadly cluster within the range of European Bronze Age (BA) and Iron Age (IA) 159 populations, characterized by an ancestry component that is related to pastoralist populations from 160 the Pontic-Caspian steppe (Fig. 2a and Extended Data Fig. 2) entering Europe around 5000 BP10,11. 161 A different dimensionality reduction technique using uniform manifold approximation and projection 162 (UMAP) revealed additional fine-scale genetic structure. European individuals from the Bronze Age 163 and onwards are generally distributed within a broad area anchored by four ancestry clusters across 164 the two UMAP dimensions: Early BA individuals from the Steppe; pre-BA Neolithic Europeans; 165 Baltic BA individuals; and Scandinavian IA and early VA individuals (Fig. 2b). We observe a wide 166 range of distributions for VA individuals within this broad area, with notable differences between 167 geographic regions (Fig. S8.10), indicating complex fine-scale structure among the different groups. 168 Modelling Scandinavian groups from the BA and onwards as mixtures of three ancestral components 169 (Mesolithic hunter-gatherers; Anatolian Neolithic; Steppe early BA), again revealed subtle 170 differences in their composition. We find that the transition from the BA to the IA is accompanied by 171 a reduction in Neolithic farmer ancestry, with a corresponding increase in both Steppe-like ancestry 172 and hunter-gatherer ancestry (Extended Data Fig. 6). While most groups show a slight recovery of 173 farmer ancestry during the VA, there is considerable variation in ancestry across Scandinavia. In 174 particular, we observe a wide range of ancestry compositions among individuals from Sweden, with 5 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 175 some groups in southern Sweden showing some of the highest farmer ancestry proportions (40% or 176 more in individuals from Malmö, Kärda or Öland). Ancestry proportions in Norway and Denmark on 177 the other hand appear more uniform (Extended Data Fig. 6). Finally, we detect an influx of low levels 178 of “eastern” ancestry starting in the early VA, mostly constrained among groups from eastern and 179 central Sweden as well as some Norwegian groups (Extended Data Fig. 6). Testing of putative source 180 groups for this “eastern” ancestry revealed differing patterns among the Viking Age target groups, 181 with contributions of either East Asian- or Caucasus-related ancestry (Supplementary Note 10). 182 Overall, our findings suggest that the genetic makeup of VA Scandinavia derives from mixtures of 183 three earlier sources: Mesolithic hunter-gatherers, Neolithic farmers, and Bronze Age pastoralists. 184 Intriguingly, our results also indicate ongoing gene flow from the south and east into Iron Age 185 Scandinavia. Thus, these observations are consistent with archaeological claims of wide-ranging 186 demographic turmoil in the aftermath of the Roman Empire with consequences for the Scandinavian 187 populations during the late Iron Age12,13. We caution, however, that our sampling for the periods 188 preceding the VA is still sparse, and hence do not provide a full picture of the genetic diversity across 189 Scandinavia during that period. 190 191 Genetic structure within Viking-Age Scandinavia 192 By the end of the Iron Age in the 8th century CE, Scandinavia formed a patchwork of conflicting and 193 competing kingdoms with a shared cultural background. For centuries, a political economy based on 194 raiding, trading and gifts had been common5. However, the cause for the development of this 195 economic and political system into the more organized maritime society of the Viking era remains 196 debated5. It is commonly argued that seafaring8,14 contributed to create a densely interlinked 197 Scandinavia during the Viking era2,15,16. 198 To disentangle the fine-scale population structure within VA Scandinavia, we performed genotype 199 imputation on a subset of 300 individuals with sufficient coverage (>0.5X) and inferred genomic 200 segments shared via identity-by-descent (IBD) within the context of a reference panel of 1,464 201 present-day Europeans, using IBDseq. We find that VA Scandinavians on average cluster into three 202 groups according to their geographic origin, shifted towards their respective present-day counterparts 203 in Denmark, Sweden and Norway (Fig. 3a). Closer inspection of the distributions for the different 204 groups reveals additional complexity in their genetic structure (Fig. S10.1). We find that the 205 ‘Norwegian’ cluster includes Norwegian IA individuals, who are distinct from both Swedish and 206 Danish IA individuals which cluster together with the majority of central and eastern Swedish VA 6 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 207 individuals. Many individuals from southwestern Sweden (e.g. Skara) cluster with Danish present- 208 day individuals from the eastern islands (Funen, Zealand), skewing towards the ‘Swedish’ cluster 209 with respect to early and more western Danish VA individuals (Jutland). Some individuals have 210 strong affinity with Eastern Europeans, particularly those from the island of Gotland in eastern 211 Sweden. The latter likely reflects individuals with Baltic ancestry, as clustering with Baltic BA 212 individuals is evident in the IBS-UMAP analysis (Fig. 2b) and through f4-statistics (Extended Data 213 Fig. 4). 214 To further quantify the within-Scandinavia population structure, we used ChromoPainter17 to identify 215 long, shared haplotypes among sequenced individuals using a reference panel enriched with 216 Scandinavian populations (n=1,675 individuals, see Supplementary Notes 6 and 11). Our approach 217 detects subtle population structure present during the VA in Scandinavia. Supplementary Figures 218 S11.1-10 and Supplementary Note 11 describe the supervised method that we used to obtain power 219 to robustly identify local ancestry variation in the presence of sequencing rate variation. We find at 220 least four major ancestry components in Scandinavia, each with affinities with a present-day 221 population (Fig. S11.11): a Danish-like, a Swedish-like, a Norwegian-like and a British-like 222 component. Henceforth, we call this latter component ‘North Atlantic’, and we suspect it may reflect 223 originally Celtic individuals that occupied the British Isles and were brought into Scandinavia. We 224 refer to the first three ancestries as ‘Danish-like’, ‘Swedish-like’ and ‘Norwegian-like’, though we 225 emphasize that the correspondence between these ancestries and present-day inhabitants of the 226 respective Scandinavian countries is by no means exact or exclusive. During the VA, we mostly find 227 high levels of Norwegian-like and Swedish-like components in Norway and Sweden, respectively, 228 while Danish-like and ‘North Atlantic’ components are more widespread within Scandinavia (Fig. 229 S11.12 and Supplementary Table 6). Notably, the ‘Swedish-like’ component is higher in Salme, 230 Estonia, than in Sweden, because our sampling scheme included several individuals from the famous 231 Salme Viking ship burial, of which archaeological and isotopic data suggest a Scandinavian 232 origin18,19. While in general individuals from most of the Scandinavian VA settlements show mixed 233 (Danish, Norwegian and Swedish) genetic ancestries, VA individuals from Jutland (Denmark) do not 234 have significant Swedish-like or Norwegian-like genetic components. Furthermore, gene flow within 235 Scandinavia appears to be broadly northwards, dominated by Danish Vikings moving into what are 236 now Norway and Sweden (Table S11.2; see Supplementary Note 11). 237 Although the majority of the Viking genomes within Scandinavia and abroad show affinities to 238 Danish, Norwegian, Swedish or British populations, there are some notable exceptions. We identified 7 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 239 two ancient individuals (VK518 and VK519) originating from northern regions of Norway 240 (Nordland), which have affinities to present-day Saami. This signal is weaker for VK519, indicating 241 that he might have also had Norwegian-like ancestors. Given the geographic provenance of these 242 samples, it was not unexpected to find individuals with Saami-like ancestry among the VA samples. 243 However, as VK519 is indeed an admixed individual with both Norwegian-like and Saami-like 244 ancestries, it appears that genetic contacts between these groups were already underway in VA 245 Norway. 246 Importantly, present-day country boundaries are not always well reflected in the genetic data. Thus, 247 the south-western part of Sweden in the VA is genetically more similar to Danish VA populations 248 than the eastern regions of mainland Sweden (i.e. the area around the Mälaren Valley), likely due to 249 geographic barriers that prevented gene flow in Sweden. 250 We quantified genetic diversity in our samples using two measures: conditional nucleotide diversity 251 (Supplementary Note 9) and variation in inferred ancestry (Supplementary Note 11; Extended Data 252 Fig. 5 and Fig. S11.13). We find overall high nucleotide diversity among most Viking-Age groups, 253 with diversity values exceeding those of earlier Neolithic or BA groups, and only slightly lower than 254 the highly diverse IA individuals from the British Isles (Fig. S9.1). Both measures of diversity vary 255 significantly across locations. Denmark and Gotland in Sweden have the highest genetic diversity in 256 the region, suggesting that these regions may have been centers of interaction and trade during this 257 time. They also possess high diversity in inferred ancestries. North Norway also has high diversity in 258 inferred ancestry due to its mixture of ‘North Atlantic’ and ‘Norwegian-like’ ancestry. 259 Interestingly, on Gotland, there are much more Danish-like, British-like and Finnish-like genetic 260 components than Swedish-like components, supporting the notion that the island may have been 261 marked by extensive maritime contacts during the VA. Our two Gotland sampling sites, Fröjel and 262 Kopparsvik, have traditionally been argued to contain non-local individuals20, but recent Sr-isotope 263 analyses have suggested otherwise21,22. 264 On Öland in Sweden, we observe high genetic diversity and the most variable patterns of recent 265 ancestry (Extended Data Fig. 5) in Scandinavia. Sr and O isotope variation in these samples, and 266 more contemporary samples from Öland have concluded that there is: (i) a high proportion (68%) of 267 non-locals, (ii) high diversity in geographical origins and (iii) long distance migration23. Thus, the 268 genetic diversity observed for Öland in the VA fits well with all of these results. 269 In conclusion, the results for Gotland and Öland agree with the archaeological record, suggesting that 270 Öland and Gotland were important trading posts from the Roman period onwards24,25. A similar 8 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 271 pattern is observed at a few archaeological sites from the central Danish islands, such as Langeland, 272 although at a lower scale. Interestingly, genetic diversity here increases from the early to the late VA, 273 suggesting increasing interregional interaction. 274 Our findings do not agree with the view of an overall highly connected population in Viking 275 Scandinavia 276 evidence of a few cosmopolitan centers to the south, in southern Sweden and Denmark, where we see 277 higher diversity of ancestries than in the rest of Scandinavia. These patterns are consistent with a 278 restricted number of sea routes between the different Scandinavian areas and beyond. 2,8,14–16 . Rather, we find clear genetic population structure within Scandinavia. We see 279 280 Viking migrations 281 Viking society is particularly famous for its ship technology, allowing for fast transport of large 282 numbers of individuals in a single vessel26. These vessels enabled the Vikings not only to carry out 283 lucrative raids and extended trade routes across Western Eurasia, but also to reach and settle lands in 284 the North Atlantic27–30. Based on historical and archaeological data, Viking presence extended into 285 both western and eastern Europe, reaching perhaps as far as the Pontic Steppe and the Middle 286 East31,32. It is commonly believed that the westward migrations and raids were mainly carried out by 287 people from what are now Norway and Denmark in the 9th and 10th centuries CE. In contrast to 288 western movements, eastward expansions are commonly believed to have been carried out by 289 Swedish Vikings, trading along navigable river systems and overland caravan routes32. Swedish 290 Vikings (the ‘Rus’) are also credited for being active in the formation of the first Russian state33,34. 291 Overall, our fine-scale ancestry analysis based on genomic data largely support the Viking expansion 292 patterns inferred from archaeology (Figs. 3, 4 and S11.12). The eastward movements mainly involved 293 individuals with Swedish-like ancestry, while the Viking individuals with Norwegian-like ancestry 294 travelled to Iceland, Greenland, Ireland and the Isle of Man. A Danish-like ancestry component is 295 more pronounced in present-day England, which is also in accordance with historical records35 and 296 still visible in place-names34, and modern genetics36,37. Importantly, however, it is currently 297 impossible for us to distinguish Danish-like ancestry in the British Isles from that of the Angles and 298 Saxons, who migrated in the 5th-to-6th centuries CE from Jutland and Northern Germany. 299 Interestingly, the ancient individuals from two execution sites in England (Dorset and Oxford) have 300 significant local ‘North Atlantic’ ancestry, as well as Danish-like and Norwegian-like ancestries. If 301 these represent Viking raiding parties coming to grief, as has been suggested38,39, this implies such 302 forces were composed of individuals from different places of origin. This pattern is also suggested 9 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 303 by isotopic data from the warrior cemetery in Trelleborg, Denmark40. Similarly, the presence of 304 Danish-like ancestry in an ancient sample from Gnezdovo (Eastern Europe) indicates that the eastern 305 migrations were not entirely composed of Vikings from Sweden. 306 However, in some cases, localities seem to have taken up Viking culture while incorporating little to 307 no Scandinavian ancestry components, suggesting that the “Viking” identity was not always 308 necessarily associated with Scandinavian genetic ancestry. Archaeological evidence indicates that the 309 six higher coverage VA individuals from three different archaeological sites in the Orkney Islands 310 have Scandinavian cultural links41,42. However, four (VK201, VK202, VK203 and VK207) of these 311 samples have over 85% “UK” ancestry and are genetically similar to present-day Irish and Scottish 312 populations (Figs 3a and S10.1, Supplementary Table 6), which is in contrast to the isotopic 313 evidence43. Haplotype-based analyses corroborate that four of these samples possessed local genetic 314 ancestries, with little Scandinavian contribution. Only two individuals - VK204 and VK205 315 - displayed c. 50% Norwegian and Danish ancestries (Supplementary Table 6), respectively, which 316 may indicate admixture between the locals and Scandinavians on the Orkney Islands during the VA. 317 The four ancient genomes of Orkney individuals with little Scandinavian ancestry may be the first 318 ones of Pictish people published to date (Supplementary Note 12). Yet a similar (>80% “UK” 319 ancestry) individual was found in Ireland (VK545) and five in Scandinavia, implying that Pictish 320 populations were integrated into Scandinavian culture by the Viking Age. 321 322 Gene flow into Scandinavia during the Viking era 323 Archaeological findings and the written sources support the hypothesis that Viking back migrations 324 and interaction between the newly settled areas and Scandinavia occurred as part of the process44. 325 Presumably, if these migrations took place, native ancestry from these areas must have also been 326 introduced into Scandinavia. We therefore aimed to assess the levels of non-Scandinavian ancestry 327 emerging in Scandinavia during the VA. 328 Using fineStructure17, we find that the levels of non-Scandinavian ancestry in the Danish, Norwegian 329 and Swedish Vikings agree with known trading routes (Supplementary Notes 11 and 12). The most 330 obvious genetic signals are from Finnish and Baltic sources into the area of what is now modern 331 Sweden, including Gotland. These ancestries are present at considerably lower levels or are 332 completely absent in most individuals from Denmark and Norway. A substantial interaction across 333 the Baltic Sea is also suggested by objects from Finland found in graves in Middle Sweden, albeit 334 recent Sr-isotope analyses are inconclusive regarding the origin of the buried individuals 45,46 . In 10 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 335 comparison, western regions of Scandinavia have much higher levels of ancestry from the British 336 Isles, in comparison with the eastern regions of Sweden (Supplementary Notes 11 and 12). We also 337 observe several individuals (Supplementary Table 6) with large amounts of South European ancestry 338 in Denmark and southwest Sweden during the Viking period (Fig. 4). No such individuals are found 339 among our Scandinavian Iron Age samples, though we stress that our sampling for this period is more 340 limited than for the other two. The discovery of individuals with ancestry from Southern Europe and 341 the British Isles is the first direct evidence for movement into Scandinavia from these regions. The 342 directions of interaction marked by these individuals is consistent with the major directions of gene 343 flows outwards from Scandinavia also seen in the data. 344 Surprisingly, three individuals from the Kärda site show much higher genetic similarity to Late 345 Neolithic/Early Bronze Age Danish individuals than to all other VA individuals in the dataset. The 346 site is located far inland, in south-west Sweden. This similarity is quite unexpected, given that the 347 samples are AMS-dated to the middle of the VA, and consistent with the presence of Caucasus-related 348 ancestry inferred in the qpAdm ancestry modelling. Studies of VA burial customs suggest that the 349 Småland area was characterized by locally confined cultural groups47. The genetic data suggest that 350 this pattern of cultural isolation was sustained in marked contrast to contemporary coastal and island 351 communities. Consistent with this hypothesis we find that the individuals from Kärda show a marked 352 reduction in nucleotide diversity compared to other VA groups (Fig. S9.1), although they also have 353 high amounts of Southern European ancestry. 354 355 Disappearance of the Greenlandic Norse 356 From around 980 to 1440 CE South-west Greenland was settled by peoples of Scandinavian (Norse) 357 descent. They likely originated from Icelandic Vikings who established a colony there at the end of 358 the 9th century CE29,48. It is believed that the Norse also reached Labrador, North America, from 359 Greenland around 1000, although no permanent settlement was established30. The fate of the Norse 360 in Greenland remains debated, but probable causes of their disappearance are social or economic 361 processes in Europe (e.g. political relations within Scandinavia and changed trading systems) and 362 natural processes, like climatic changes29,49,50. 363 We see no evidence of long-term inbreeding in the Greenlandic Norse genomes, though we note that 364 we only have one high-coverage genome from the later period of occupation of Greenland 365 (Supplementary Note 10; Figs. S10.2 and S10.3). This suggests a depopulation scenario over 366 approximately 100 years which would be in line with previous demographic models51, as well as the 11 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 367 archaeology. Indeed, the latter indicates that marginal farms in the Western Settlement and the 368 northern and southern parts of the Eastern Settlement were abandoned from about 1200 CE, with no 369 converse intensification of settlement in the central areas. 370 We also find no evidence of ancestry from local populations from the Western Atlantic (Paleo 371 Eskimo, Inuit or Native American) in the Norse genomes. This is in accordance with previous 372 physical anthropological studies of the skeletal remains51. This suggests that either sexual interactions 373 did not take place or that, if they did, then on a very small and incidental scale with the children 374 remaining in the native communities. In terms of genetic ancestry of the Greenlandic Norse, we find 375 evidence of admixture between Scandinavians (mostly from Norway) and individuals from the British 376 Isles, similar to the first settlers of Iceland52, which supports the archaeological and historical links 377 between the Greenlandic Norse and the Icelandic Vikings. 378 379 Genetic composition of the earliest Viking expedition and kinship findings 380 Maritime raiding has been a constant of seafaring cultures for millennia. However, the VA is unusual 381 in that it is partly defined by such activity53. Despite the historical importance of Viking raiding, the 382 exact nature and composition of these war parties is unknown5. Only one Viking raiding or diplomatic 383 expedition has left direct archaeological traces, at Salme in Estonia, where 41 Swedish Vikings who 384 died violently were buried in two boats accompanied by high-status weaponry18,19. Importantly, the 385 Salme boat-burial predates the first textually documented raid (in Lindisfarne in 793) by nearly half 386 a century. 387 Comparing the genomes of 34 individuals from the Salme burial using kinship analyses, we find that 388 these elite warriors included four brothers buried side by side and a 3rd degree relative of one of the 389 four brothers (Supplementary Note 4). In addition, members of the Salme group had very similar 390 ancestry profiles, in comparison to the profiles of other Viking burials (Supplementary Notes 10 and 391 11). This suggests that this raid was conducted by genetically homogeneous people of high status, 392 including close kin. Isotope analyses indicate that the crew descended from the Mälaren area in 393 Eastern Sweden19 thus confirming that the Baltic-Mid-Swedish interaction took place early in the 394 VA. 395 Intriguingly, we identified several additional pairs of kin among the other Viking genomes. One is a 396 pair of 2nd degree male relatives (i.e. half-brothers, nephew-uncle or grandson-grandfather) from two 397 locations separated by the North Sea: one of the samples (VK279) was excavated in Denmark 398 (Galgedil site on Funen; this cemetery was also analyzed for strontium with a group of non-locals 12 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 399 there) while the other individual (VK144) was found in the UK (Oxford site). Another pair of 400 individuals with 3rd or 4th degree relatedness (e.g. cousins) was discovered in Sweden, namely a male 401 sample excavated on the island of Öland (VK342) and a female individual from Skämsta, Uppsala 402 (VK527), some 300-400 kilometers apart. Interestingly, the female from Uppsala (VK527) also had 403 a brother (VK517), and both siblings display a rare genetic disorder of abnormal skeletal 404 development: spondyloepiphyseal dysplasia. Given the very low frequency of this disorder, the close 405 family ties between these individuals were expected by the archaeologists22. Such long-distance 406 relationships in our dataset underscore the degree of individual-level mobility during the Viking era. 407 408 Positive selection in Europe in the last 10,000 years 409 The availability of hundreds of genomes from the IA and VA - in combination with previously 410 published Mesolithic, Neolithic and Bronze Age genomes10,11,54,55 - permit us to directly investigate 411 the role of positive selection using time series of allele frequencies from the last ten millennia of 412 European history. We looked for SNPs whose allele frequencies changed significantly in the last 413 10,000 years, using a newly developed method called “neoscan” that is implemented in the Ohana 414 software package56,57, and that can detect strong allele frequency shifts in time that cannot be 415 explained by temporal changes in genome-wide genetic ancestry alone (Supplementary Note 14). 416 Figure 5a shows the resulting likelihood ratio scores in favor of selection looking at the entire 10,000- 417 year period (top, “general” scan), the period up to 4,000 BP (middle, “ancient” scan) and the period 418 from 4,000 BP up to the present (bottom, “recent” scan). The strongest candidate for selection - 419 especially in the “recent” scan - is a cluster of SNPs near the LCT gene - a signal that has been 420 extensively characterized in the past58,59. The rise in frequency of the lactase persistence allele to its 421 present-day levels in Northern Europe is, however, poorly understood. We know that this rise must 422 have occurred after the Bronze Age, a time at which this allele was still segregating at low 423 frequencies10,54. Based on the archaeological record, we also know that VA Scandinavians used a 424 variety of dairy products as an essential part of their daily food intake. Our dataset allows us, for the 425 first time, to directly assess the frequency of the lactase persistence allele (at SNP rs4988235, 426 upstream of the LCT gene) in Scandinavia during the Iron Age and VA, and trace its evolution since 427 the Bronze Age. 428 Figure 5b shows that VA groups had very similar allele frequencies at the LCT lactase persistence 429 SNP to those found in present-day northern European populations. In contrast, the persistence allele 430 was at low frequencies in Bronze Age Scandinavians, as well as Corded Ware and Bell Beaker 13 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 431 cultures from central Europe, even though there is evidence for milk consumption in these regions by 432 that time. The allele frequency in Iron Age samples is at intermediate levels (c. 37.5%), suggesting 433 this rise in frequency must indeed have occurred during the Iron Age (c. 1500-2500 years ago), but 434 was largely complete at the onset of the VA. Interestingly, the allele frequency of the allele is much 435 higher (c. 40%) in the Bronze Age population from the neighboring Baltic Sea region than in Bronze 436 Age Scandinavia. Given the geographic and cultural proximities between Scandinavia and the Baltic 437 region, this may suggest gene flow between the two regions resulting in increased frequency of lactase 438 persistence in Scandinavia during the Iron Age. 439 Other candidates for selection include previously identified regions like the TLR1/TLR6/TLR10 440 region, the HLA region, SLC45A2 and SLC22A454. We also find several new candidate regions for 441 selection in the “ancient” scan, some of which contain SNPs where the selected allele rose in 442 frequency early in the Holocene but then decreased later on (Supplementary Note 14). These 443 candidate regions include a region overlapping the DCC gene, which has been implicated in 444 colorectal cancer60 and another overlapping the AKNA gene, which is involved in the secondary 445 immune response by regulating CD40 and its ligand61. This highlights the utility of using ancient 446 DNA to detect signatures of selection that may have been erased by recent selective dynamics. 447 448 Pigmentation-associated SNPs 449 Exploring twenty-two SNPs with large effect associated with eye color and hair pigmentation, we 450 observe that their frequencies are very similar to those of present-day Scandinavians (Supplementary 451 Note 13). This suggests that pigmentation phenotype in VA Scandinavians may not have differed 452 much from the present-day occupants of the region (although see section on complex traits below for 453 an analysis including alleles of small effect). Nevertheless, it is important to stress that there is quite 454 a lot of variation in the genotypes of these SNPs across the sequenced samples, and that there is 455 therefore not a single ‘Viking phenotype’. For example, two of the ancient samples with the highest 456 coverage have different pigmentation phenotypes: VA individual VK42 from Skara, Sweden has 457 alleles associated with brown eyes and darker hair coloration while VK1 from Greenland was likely 458 to have had blue eyes and lighter hair. 459 460 Evolution of complex traits in Scandinavia 461 To search for signals of recent population differentiation of complex traits, we compared genotypes 462 of Viking age samples with those of a present-day Scandinavian population for a range of trait- 14 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 463 associated SNP markers. We selected 16 traits for which summary statistics from well-powered 464 genome-wide association studies (GWAS) were available through the GWAS ATLAS 465 (https://atlas.ctglab.nl)62. For comparison with the Viking age samples we used a random population 466 subset of the IPSYCH case-cohort study of individuals born in Denmark between 1981-201163. We 467 derived polygenic risk scores (PRS) for the 16 traits, based on independent genome-wide significant 468 allelic effects and tested for a difference in the distribution of polygenic scores between the two 469 groups, correcting for sex and ancestry-sensitive principal components (Supplementary note S15). 470 We observed a significant difference between the polygenic scores of VA samples and current Danish 471 population samples for three traits; black hair color (P = 0.00089), standing height (P = 0.019) and 472 schizophrenia (P = 0.0096) (Extended Data Fig. 5). For all three traits, the polygenic score was higher 473 in the VA group than in the present-day Danish group. The observed difference in PRS for height and 474 schizophrenia between the groups did however not remain significant after taking into account the 475 number of tests. A binomial test of the number of black hair color risk alleles found in higher 476 frequency in the VA sample and the present-day sample, also returned a significant difference (65/41; 477 P = 0.025), which suggests that the signal is not entirely driven by a few large-effect loci. 478 Thus, we only find evidence for systematic changes in combined frequencies of alleles affecting hair 479 color (and possibly also height and schizophrenia), among all the anthropometric traits and complex 480 disorders we tested. Also, we cannot conclude whether the observed difference in allele frequencies 481 are due to selection acting on these alleles between the Viking Age and the present time or to some 482 other factors (such as more ethnic diversity in the VA sample), nor can we conclude whether a similar 483 change has occurred in other Nordic populations than the Danish. 484 485 Genetic legacy of the Vikings in present-day populations 486 To test whether present-day Scandinavians share increased ancestry with their respective ancient 487 Viking counterparts, we first inferred D-statistics of the form D(YRI, ancient; present-day X, present- 488 day DK), which contrast allele sharing of a test ancient individual with a present-day test population 489 X and present-day Danes. We find subtle but noticeable shifts of ancient individuals towards their 490 present-day counterparts in the distributions of these statistics (Extended Data Fig. 3). We further 491 examined variation in present-day populations using fineSTRUCTURE, and then described these 492 present-day groups by their ancestry from ancient populations (Fig. S11.14). 493 We find that within Scandinavia, present-day populations are still structured according to the ancient 494 Viking population groups. The component that we associated as Norwegian-like is present at 45-65% 15 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 495 in present-day Norway. Similarly, the ancient Swedish-like ancestry is present at 15-30% within 496 Sweden. Of the four Swedish clusters, one is more related to the ancient Finnish than the Swedish- 497 like ancestry, and a second is more related to Danes and Norwegians. Danish-like ancestry is now 498 high across the whole region. 499 Outside of Scandinavia, the genetic legacy of the Vikings is consistent, though limited. A small 500 component is present in Poland (up to 5%) and the south of Europe. Within the British Isles, it is 501 difficult to assess how much of the Danish-like ancestry is due to pre-existing Anglo-Saxon ancestry; 502 however, the Norwegian-like ancestry is consistently around 4%. The Danish-like contribution is 503 likely to be similar in magnitude and is certainly not larger than 16% as found in Scotland and Ireland. 504 The lack of strong variation in ancestry from Scandinavia makes sense if the Vikings did not maintain 505 a diaspora identity over time but instead integrated into the respective societies in which they settled. 506 The genetic impacts are stronger in the other direction. The ‘British-like’ populations of Orkney 507 became ‘Scandinavian’ culturally, whilst other British populations found themselves in Iceland and 508 Norway, and beyond. Present-day Norwegians vary between 12 and 25% in their ‘British-like’ 509 ancestry, whilst it is still (a more uniform) 10% in Sweden. Separating the VA signals from more 510 recent population movements is difficult, but these numbers are consistent with our VA estimates. 511 512 Discussion 513 Until now, our main understanding of the VA was largely based on a combination of historical sources 514 and archaeological evidence. These often characterize the VA as a period of high mobility and 515 interaction between peoples. Networks of trade were established, connecting distant regions within 516 Scandinavia through established waterways with significant movement between regions. It has also 517 been viewed as a time where links were created to regions outside Europe, from the Pontic Steppe in 518 the east to North America in the west. 519 Our genomic analyses add complex layers of nuance to this simple picture. We largely reconstruct 520 the long-argued movements of Vikings outside Scandinavia: Danish Vikings going to Britain, 521 Norwegian Vikings moving to Ireland, Iceland, and Greenland, and Swedish Vikings sailing east 522 towards the Baltic and beyond. However, we also see evidence of individuals with ancient Swedish 523 and Finnish ancestry in the westernmost fringes of Europe, whilst Danish-like ancestry is also found 524 in the east, defying our modern notions of historical groupings. It is likely that many such individuals 525 were from communities with mixtures of ancestries, likely thrown together by complex trading, 526 raiding and settling networks that crossed cultures and the continent. 16 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 527 Our observations also suggest that the different parts of Scandinavia were not as evenly connected, 528 as has often been assumed. Despite relatively fast and easy communication between the coastal 529 regions of Denmark, Norway, and Sweden, we find that clear genetic structure was present in Viking- 530 Age Scandinavia. In fact, our data indicate that Viking Scandinavia consisted of a limited number of 531 transport zones and maritime enclaves64 where contact was made with Europe, while the remaining 532 regions had limited external gene flow with the rest of the Scandinavian continent. Some Viking-Age 533 Scandinavian locations are relatively homogeneous both in terms of genetic diversity and patterns of 534 ancestry; particularly mid-Norway, Jutland, and the Atlantic settlements, which contain 535 predominantly Norwegian-like and ‘North Atlantic’ (including pre-Anglo Saxon British) ancestry. 536 Indeed, one of the clearest vectors of contrast observed in this study is between the strong genetic 537 variation seen in relatively populous coastal trading communities such as in the islands Gotland and 538 Öland, and the reduced diversity in less populated (mostly inland) areas in Scandinavia. Such high 539 genetic heterogeneity, which was likely due to increased population size, extends the urbanization 540 model of Late Viking Age city of Sigtuna proposed by Krzewińska et al.6 both spatially and further 541 back in time. 542 Interestingly, our findings correspond with paleodemographic studies based on place-name evidence 543 and archaeological distributions suggesting population density was higher in Denmark than elsewhere 544 in Viking-Age Scandinavia65. Gene flow from Denmark to the north is also paralleled by the linguistic 545 affinities of the medieval Scandinavian languages: The 12th-century Icelandic law text Grágás states 546 that the common language of Swedes, Norwegians, Icelanders, and Danes was dǫnsk tunga (‘Danish 547 tongue’)66. It appears that the formation of large-scale trading and cultural networks that spread 548 people, goods and warfare took time to affect the heartlands of Scandinavia, which received much 549 more restricted gene flow, retaining pre-existing genetic differences between Scandinavian 550 populations. This pattern of behavior seems to prevail from the beginning of the Viking diaspora to 551 its end at the beginning of the medieval period. 552 Our findings also show that Vikings are not simply a direct continuation of the Scandinavian Iron 553 Age groups. Rather than simple continuity, we observe foreign gene flow from the south and east 554 into Scandinavia, starting in the Iron Age, and continuing throughout the duration of the Viking 555 period from an increasing number of sources. Our findings also contradict the myth of the Vikings as 556 peoples of pure local Scandinavian ancestry. In fact, we found many Viking Age individuals with 557 high levels of foreign ancestry, both within and outside Scandinavia, suggesting ongoing gene flow 558 with different peoples across Europe. Indeed, it appears that some foreign peoples contributed more 17 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 559 genetic ancestry to Scandinavia during this period than the Vikings contributed to them which could 560 partially be due to smaller effective population size of the VA Scandinavians as opposed to their 561 continental and British neighbors. 562 563 Acknowledgements 564 This work was supported by the Mærsk Foundation, the Lundbeck Foundation, the Novo Foundation, 565 the Danish National Research Foundation, KU2016, and the Wellcome Trust (grant nos. 566 WT104125MA). The authors thank the iPSYCH Initiative, funded by the Lundbeck Foundation 567 (grant nos. R102-A9118 and R155-2014-1724), for supplying SNP frequency estimates from the 568 present-day Danish population for comparison with Viking Age samples. EW would like to thank St. 569 John’s College, Cambridge for providing an excellent environment for scientific thoughts and 570 collaborations. SR was supported by the Novo Nordisk Foundation (NNF14CC0001). FR was 571 supported by a Villum Young Investigator Award (project no. 000253000). The authors are grateful 572 to Marisa Corrente for providing access to the skeletal remains from Cancarro and Nunzia M. 573 Mangialardi and Marco Maruotti for the useful suggestion. G.S. and E.C. were supported by a Marie 574 Skłodowska-Curie Individual Fellowship “PALAEO-ENEO”, a project funded by the European 575 Union EU Framework Programme for Research and Innovation Horizon 2020 (Grant Agreement 576 number 751349). RM was supported by an EMBO Long-Term Fellowship (ALTF 133-2017). We 577 thank Mattias Jakobsson and Anders Götherström for providing preliminary access to the sequencing 578 data of 23 Viking Age samples from Sigtuna; L. Vinner, A. Seguin-Orlando, K. Magnussen, L. 579 Petersen and C. Mortensen at the Danish National Sequencing Centre for producing the analysed 580 sequences; P.V. Olsen and T. Brand for technical assistance in the laboratories. We thank Richard M. 581 Durbin and James H. Barrett for comments and suggestions. 582 583 18 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 584 References 585 1. Brink, S. & Price, N. The Viking World. (Routledge, 2008). 586 2. Jesch, J. The Viking Diaspora. (Routledge, 2015). 587 3. Ashby, S. P. What really caused the Viking Age? The social content of raiding and 588 589 exploration. Archaeological Dialogues 22, 89–106 (2015). 4. Barrett, J. Rounding up the usual suspects: Causation and the Viking Age diaspora. in The 590 Global Origins and Development of Seafaring 289–302 (McDonald Institute for 591 Archaeological Research, 2010). 592 5. 593 594 (Routledge, 2016). 6. 595 596 7. 8. 9. 10. 11. Haak, W. et al. Massive migration from the steppe was a source for Indo-European languages in Europe. Nature 522, 207–211 (2015). 12. 607 608 Allentoft, M. E. et al. Population genomics of Bronze Age Eurasia. Nature 522, 167–172 (2015). 605 606 Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 19, 1655–1664 (2009). 603 604 Sindbæk, S. M. & Trakadas, A. The World in the Viking Age. (Viking Ship Museum in Roskilde, 2014). 601 602 Eriksen, M. H., Pedersen, U., Rundberget, B. & Axelsen, I. Viking Worlds: Things, Spaces and Movement. (Oxbow Books, 2014). 599 600 Krzewińska, M. et al. Genomic and Strontium Isotope Variation Reveal Immigration Patterns in a Viking Age Town. Curr. Biol. 28, 2730–2738.e10 (2018). 597 598 Barrett, J. H. & Gibbon, S. J. Maritime Societies of the Viking and Medieval World. Hedeager, L. Scandinavia before the Viking Age. in The Viking World 35–46 (Routledge, 2008). 13. 609 Hedeager, L. Iron Age Myth and Materiality: An Archaeology of Scandinavia AD 400-1000. (Routledge, 2011). 610 14. Bill, J. Viking ships and the sea. in The Viking World 170–180 (Routledge, 2008). 611 15. Sindbæk, S. M. The Small World of the Vikings: Networks in Early Medieval 612 613 Communication and Exchange. Norwegian Archaeological Review 40, 59–74 (2007). 16. Hilberg, V. & Kalmring, S. Viking Age Hedeby and Its Relations with Iceland and the North 614 Atlantic: Communication, Long-distance Trade, and Production. in Viking Archaeology in 615 Iceland: Mosfell Archaeological Project 221–245 (Brepols Publishers, 2014). 19 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 616 17. 617 618 dense haplotype data. PLoS Genet. 8, e1002453 (2012). 18. 619 620 19. Douglas Price, T., Peets, J., Allmäe, R., Maldre, L. & Oras, E. Isotopic provenancing of the Salme ship burials in Pre-Viking Age Estonia. Antiquity 90, 1022–1037 (2016). 20. 623 624 Peets, J. et al. Research results of the Salme ship burials in 2011-2012. Archaeological fieldwork in Estonia 2012, 43–60 (2012). 621 622 Lawson, D. J., Hellenthal, G., Myers, S. & Falush, D. Inference of population structure using Thunmark-Nylén, L. Visby-ett pussel med gamla och nya pusselbitar. Fornvännen 99, 285– 297 (2004). 21. Peschel, E. M., Carlsson, D., Bethard, J. & Beaudry, M. C. Who resided in Ridanäs?: A study 625 of mobility on a Viking Age trading port in Gotland, Sweden. Journal of Archaeological 626 Science: Reports 13, 175–184 (2017). 627 22. Arcini, C. The Viking Age: A Time with Many Faces. (Oxbow Books, 2018). 628 23. Wilhelmson, H. & Price, T. D. Migration and integration on the Baltic island of Öland in the 629 630 Iron Age. Journal of Archaeological Science: Reports 12, 183–196 (2017). 24. Hansen, U. L. Römischer Import im Norden: Warenaustausch zwischen dem Römischen 631 Reich und dem freien Germanien während der Kaiserzeit unter besonderer Berücksichtigung 632 Nordeuropas. (Det Kongelige nordiske Oldskriftselskab, 1987). 633 25. Andersson, K. I skuggan av Rom: romersk kulturpåverkan i Norden. (Atlantis, 2013). 634 26. Crumlin-Pedersen, O. Maritime Topography and the Medieval Town. in Ships as indicators of 635 trade in Northern Europe 600-1200 11–20 (National Museum of Denmark, Department of 636 Danish Collections, 1999). 637 27. Sigurðsson, J. V. Iceland. in The Viking World 571–587 (Routledge London, UK, 2008). 638 28. Sigurðsson, G. The North Atlantic expansion. in The Viking World 562–570 (Routledge 639 640 London, UK, 2008). 29. 641 642 London, UK, 2008). 30. 643 644 647 Wallace, B. The discovery of Vinland. in The Viking World 604–612 (Routledge London, UK, 2008). 31. 645 646 Arneborg, J. The Norse settlement in Greenland. in The Viking World 588–603 (Routledge Androshchuk, F. Vikings in the East : Essays on Contacts along the Road to Byzantium (8001100). (Coronet Books Incorporated, 2013). 32. Duczko, W. Viking Rus: studies on the presence of Scandinavians in Eastern Europe. (Brill, 2004). 20 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 648 33. 649 650 UK, 2008). 34. 651 652 35. 36. 37. 38. Loe, L., Boyle, A., Webb, H. & Score, D. ‘Given to the Ground’: A Viking Age Mass Grave on Ridgeway Hill, Weymouth. (Dorset Natural History and Archaeological Society, 2014). 39. 661 662 Athanasiadis, G. et al. Nationwide Genomic Study in Denmark Reveals Remarkable Population Homogeneity. Genetics 204, 711–722 (2016). 659 660 Leslie, S. et al. The fine-scale genetic structure of the British population. Nature 519, 309– 314 (2015). 657 658 Downham, C. Viking ethnicities: a historiographic overview. History Compass 10, 1–12 (2012). 655 656 Fellows-Jensen, G. Scandinavian place names in the British Isles. in The Viking World 391– 400 (Routledge London, UK, 2008). 653 654 Androshchuk, F. The Vikings in the East. in The Viking World 517–542 (Routledge London, Wallis, S. The Oxford Henge and Late Saxon Massacre: With Medieval and Later Occupation at St John’s College, Oxford. (Thames Valley Archaelogical Services Limited, 2014). 40. Douglas Price, T., Frei, K. M., Dobat, A. S., Lynnerup, N. & Bennike, P. Who was in Harold 663 Bluetooth’s army? Strontium isotope investigation of the cemetery at the Viking Age fortress 664 at Trelleborg, Denmark. Antiquity 85, 476–489 (2011). 665 41. Morris, C. D. The Birsay Bay Project Volume 1. Coastal Sites Beside the Brough Road, 666 Birsay, Orkney: Excavations 1976–1982. (Durham: University of Durham Department of 667 Archaeology Monograph Series, 1989). 668 42. 669 670 Ritchie, A. Excavation of Pictish and Viking-age farmsteads at Buckquoy, Orkney. Society of Antiquaries of Scotland 108, 174–227 (1976). 43. Montgomery, J., Grimes, V., Buckberry, J., Evans, J.A., Richards, M.P. and Barrett, J.H. 671 Finding Vikings with isotope analysis: The view from wet and windy islands. Journal of the 672 North Atlantic 7, 54–70 (2014). 673 44. 674 675 Tsigaridas Glørstad, Z. Sign of the Times? The Transfer and Transformation of Penannular Brooches in Viking-Age Norway. Norwegian Archaeological Review 45, 30–51 (2012). 45. Gustin, I. Contacts identity and Hybridity: Objects from South-western Finland in the Birka 676 graves. in Identity Formation and Diversity in the Early Medieval Baltic and Beyond 205–258 677 (BRILL, 2017). 21 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 678 46. Price, T. D., Arcini, C., Gustin, I., Drenzel, L. & Kalmring, S. Isotopes and human burials at 679 Viking Age Birka and the Mälaren region, east central Sweden. Journal of Anthropological 680 Archaeology 49, 19–38 (2018). 681 47. 682 683 Age. (Lund: Lund University, 2003). 48. 684 685 Price, T. D. & Arneborg, J. The Peopling of the North Atlantic: Isotopic Results from Greenland. Journal of the North Atlantic 7, 164–185 (2014). 49. 686 687 Svanberg, F. Death Rituals in South-East Scandinavia Ad 800-1000: Decolonizing the Viking Dugmore, A. J. et al. Cultural adaptation, compounding vulnerabilities and conjunctures in Norse Greenland. Proc. Natl. Acad. Sci. U. S. A. 109, 3658–3663 (2012). 50. Arneborg, J. Norse Greenland: Research into abandonment. in Medieval Archaeology in 688 Scandinavia and Beyond: History, trends and tomorrow 247–271 (Aarhus Universitetsforlag, 689 2015). 690 51. 691 692 Tusculanum Press, 1998). 52. 693 694 Ebenesersdóttir, S. S. et al. Ancient genomes from Iceland reveal the making of a human population. Science 360, 1028–1032 (2018). 53. 695 696 Lynnerup, N. The Greenland Norse: a biological-anthropological study. (Museum Sindbæk, S. M. Urbanism and exchange in the North Atlantic/Baltic, 600-1000CE. in Routledge Handbook of Archaeology and Globalization 553–565 (Routledge, 2016). 54. 697 Mathieson, I. et al. Genome-wide patterns of selection in 230 ancient Eurasians. Nature 528, 499–503 (2015). 698 55. Mathieson, I. et al. The genomic history of southeastern Europe. Nature 555, 197–203 (2018). 699 56. Cheng, J. Y., Racimo, F. & Nielsen, R. Ohana: detecting selection in multiple populations by 700 701 modelling ancestral admixture components. bioRxiv 546408 (2019). 57. 702 703 1319–1326 (2019). 58. 704 705 708 Enattah, N. S. et al. Identification of a variant associated with adult-type hypolactasia. Nat. Genet. 30, 233–237 (2002). 59. 706 707 Alves, J. M. et al. Parallel adaptation of rabbit populations to myxoma virus. Science 363, Bersaglieri, T. et al. Genetic signatures of strong recent positive selection at the lactase gene. Am. J. Hum. Genet. 74, 1111–1120 (2004). 60. Fearon, E. R. et al. Identification of a chromosome 18q gene that is altered in colorectal cancers. Science 247, 49–56 (1990). 22 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 709 61. 710 711 AKNA. Nature 410, 383–387 (2001). 62. 712 713 Siddiqa, A. et al. Regulation of CD40 and CD40 ligand by the AT-hook transcription factor Watanabe, K. et al. A global view of pleiotropy and genetic architecture in complex traits. bioRxiv 500090 (2018). 63. Pedersen, C. B. et al. The iPSYCH2012 case-cohort sample: new directions for unravelling 714 genetic and environmental architectures of severe mental disorders. Mol. Psychiatry 23, 6–14 715 (2018). 716 64. Westerdahl, C. The maritime cultural landscape. Int. J. Naut. Archaeol. 21, 5–14 (1992). 717 65. Gammeltoft, P., Jakobsen, J. G. G. & Sindbæk, S. M. Vikingetidens bebyggelse omkring 718 Kattegat og Skagerrak: Et forsøg på kortlægning. in Et fælles hav: Skagerrak og Kattegat i 719 vikingetiden. 6–23 (Nationalmuseet, 2015). 720 721 66. Foote, P., Perkins, R. & Dennis, A. Laws of Early Iceland: Grágás: the Codex Regius of Grágás with Material from Other Manuscripts. (University of Manitoba Press, 1980). 722 723 724 23 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 Fig. 1: Map of the “Viking World” from 8th till 11th centuries. Different symbols on the map (a) 751 correspond to ancient sites of a specific age/culture. The ancient samples are divided into the 752 following five broad categories: Bronze Age (BA) - c. 2500 BC - 900 BC; Iron Age (IA) - c. 900 BC 753 to 700 CE; Early Viking Age (EVA) - c. 700 to 800 CE; VA - c. 800 to 1100 CE; Medieval - c. 1100 754 to 1600 CE. b, All ancient individuals from this study (n=442) and published VA samples (n=21) 755 from Sigtuna6 are categorized based on their spatio-temporal origin. 756 24 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 757 758 759 760 761 762 763 764 765 766 767 Fig. 2: Genetic structure of VA samples. a, Multidimensional scaling (MDS) plot based on a 768 pairwise identity-by-state (IBS) sharing matrix of the VA and other ancient samples (Supplementary 769 Table 3). b, Uniform manifold approximation and projection (UMAP) analysis of the same dataset 770 as in plot (a). 771 25 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 Fig. 3: Genetic structure and diversity of ancient samples. a, Uniform manifold approximation 788 and projection (UMAP) analysis of the ancient and modern Scandinavian individuals based on the 789 first 10 dimensions of MDS using identity-by-descent (IBD) segments of imputed individuals. Large 790 symbols indicate median coordinates for each group. b, Genetic diversity in major Scandinavian VA 791 populations. Plots next to the map show MDS analysis based on a pairwise IBS sharing matrix. Here 792 “Norway” represents all the sites from Norway. The scale is identical for all the plots. 793 26 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 Fig. 4: Spatiotemporal patterns of Viking and non-Viking ancestry in Europe during the IA, 817 EVA and VA. UK = ‘British-like’ / ‘North Atlantic’ ancient ancestry component. Sweden = 818 ‘Swedish-like’ ancient ancestry component. Denmark = ‘Danish-like’ ancient ancestry component. 819 Norway = ‘Norwegian-like’ ancient ancestry component. Italy = ‘Southern European-like’ ancestry 820 component. See Table S11.2 for statistical tests. The ‘Swedish-like’ ancestry is the highest in present- 821 day Estonia due to the ancient samples from the Salme ship burial, which originated from the Mälaren 822 Valley of Sweden, according to archaeological sources. 823 27 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 824 825 826 Fig. 5: Positive selection in Europe. a, Manhattan plots of the likelihood ratio scores in favor of 827 selection looking at the entire 10,000-year period (top, “general” scan), the period up to 4,000 BP 828 (middle, “ancient” scan) and the period from 4,000 BP up to the present (bottom, “recent” scan). The 829 highlighted SNPs have a score larger than the 99.9% quantile of the empirical distribution of log- 830 likelihood ratios, and have at least two neighboring SNPs (+/- 500kb) with a score larger than the 831 same quantile. b, Frequencies of the derived ‘A’ allele rs4988235 SNP responsible for lactase 832 persistence in humans for different Viking-Age groups, present-day populations from the 1000 833 Genomes Project as well as relevant Bronze Age population panels. The numbers at the top of the 834 bars denote the sample size on which the allele frequency estimates are based. 835 836 28 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 837 Extended Data Figures 838 839 Extended Data Fig. 1: Viking Age archaeological sites. 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 Examples of a few archaeological Viking Age sites and samples used in this study. a, Salme II ship burial site of Early Viking Age excavated in present-day Estonia: schematic representation of skeletons (upper left-hand corner image) and aerial images of skeletons (upper right-hand corner and lower images). b, Ridgeway Hill mass grave dated to the 10th or 11th century, located on the crest of Ridgeway Hill, near Weymouth, on the South coast of England. Around 50 predominantly young adult male individuals were excavated. c, The site of Balladoole: around AD 900, a Viking was buried in an oak ship at Balladoole, Arbory in the south east of the Isle of Man. d, Viking Age archaeological site in Varnhem, Sweden: Schematic map of the church foundation (left) and the excavated graves (red markings) at the early Christian cemetery in Varnhem; foundations of the Viking Age stone church in Varnhem (middle) and the remains of a 182 cm long male individual (no. 17) buried in a lime stone coffin close to the church foundations (right). 29 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 856 Extended Data Fig. 2: Model-based clustering analysis 857 858 859 860 861 862 863 864 Admixture plot (K=2 to K=5) for 517 ancient individuals spanning 60 different populations. This figure is a subset of most relevant individuals and populations from Figure S7.2, see Supplementary Note 7 for details. 30 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 865 866 Extended Data Fig. 3: Symmetry tests of genetic affinity of ancient individuals with contemporary populations. 867 868 869 870 871 872 873 874 Panels show D-statistics of the form D(YRI,Y; X,Denmark), which contrast allele sharing of an ancient individual Y with either contemporary population X or Denmark. Plot symbols show point estimates, and density plots distributions across all individuals per analysis group. 31 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 875 876 Extended Data Fig. 4: Symmetry tests for genetic affinity with Baltic Bronze Age 877 878 879 880 881 882 883 884 885 Panels show f -statistics of the form f (Mbuti,Baltic_BA;Y, Salme.SG_EVA), which contrast allele sharing of Baltic_BA with either a test individual Y or Salme.SG_EVA. a, point estimates and error bars (± 3 standard errors) for each target individual, aggregated by analysis group. Individuals with significant f -statistics (|Z| ≥ 3) are indicated without transparency and respective sample IDs. b, as in (a), with density plot for distributions across all individuals per analysis group. 4 4 4 32 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 886 Extended Data Fig. 5: Ancestry diversity of different population groups 887 888 889 890 891 892 893 894 895 Diversity of different labels (i.e. sample locations combined with historical age) are shown as a function of their sample size. The Diversity measure is the Kullback-Leibler divergence from the label means, capturing the diversity of a group with respect to the average of that group; see text for details. Larger values are more diverse, though a dependence on sample size is expected. The simulation expectation for the best-fit to the data (0=0.2) is shown. 33 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 896 Extended Data Fig. 6: Ancestry modelling using qpAdm 897 898 899 900 901 902 903 904 905 906 907 a, Ternary plots of ancestry proportions for a three-way model of Mesolithic hunter-gatherer (Loschbour.SG_M), Neolithic farmer (Barcin.SG_EN) and Bronze Age Steppe herders (Yamnaya.SG_EBA). b, Bar plots with ancestry proportions as in (a), with error bars indicating standard errors and transparency/text colors indicating p-value for model fit (no transparency/black: p ≥ 0.05; light transparency/blue: 0.05 > p ≥ 0.01; strong transparency/red: p ≤ 0.01). c, Ancestry proportions of four-way models including additional putative source groups for target groups for which three-way fit was rejected (p ≤ 0.01); transparency/text colors as in (b) 34 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 908 Extended Data Fig. 7: Polygenic risk scores 909 910 911 912 913 914 915 916 917 918 919 Polygenic risk scores (PRS) for 16 complex human traits in Viking Age samples from Denmark, Sweden and Norway compared against a reference sample of >20,000 Danish-ancestry individuals randomly drawn from all individuals born in Denmark in 1981-2011. The PRS is in each case based on allelic effects for >100 independent genome-wide significant SNPs from recent GWAS of the respective traits. Only PRS for black hair colour is significantly different between the groups after taking account of multiple testing, although PRS for height and schizophrenia are considerably elevated as well in the Viking Age samples. 35 bioRxiv preprint doi: https://doi.org/10.1101/703405. this version posted July 17, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license. 920 921 Extended Data Fig. 8: Kinship analysis of ancient samples from Sandoy Church 2 site in Faroe Islands. 922 923 924 925 926 927 928 929 930 931 932 a, Reconstruction of four most likely pedigree networks for one (Family-1) of the three families in Sandoy Church 2 site in Faroe Islands. b, Five most likely pedigree networks for the Family-2: the most “parsimonious” network (top left) is likely to represent the true family relationship between the individuals (i.e. grandparents and grandsons) based on the burial pattern of the graves as shown at the bottom image (c). Ages of the individuals are approximate to help pedigree reconstructions. Blue diamond shapes and lines in each possible pedigree reconstruction represent the same individual. 36