5 Molecular Taxonomy and Phylogeny

Byron J. Adams1, Adler R. Dillman2 and Camille Finlinson1

1
Brigham Young University, Provo, Utah, USA; 2California Institute of Technology,
Pasadena, California, USA

5.1. Introduction 119

5.2. The History of Reconstructing Meloidogyne Phylogenetic History 119
5.3. Molecular Phylogenetics: Genetic Markers and Evolutionary Relationships 120
5.4. A Meloidogyne Supertree Analysis 127
5.5. Conclusions and Future Directions 130
5.6. References 135

5.1. Introduction cies descriptions) and, subsequently, phylogenetic

analyses. As molecular markers increasingly dem-
The genus Meloidogyne contains over 90 described onstrated improved resolving power, they became
species and each of these species typically has an more commonplace as diagnostic tools, eventually
extremely broad host range (as many as 3000 becoming more prominent as parts of formal
plant species; Trudgill and Blok, 2001). In add- taxonomic statements (including descriptions of
ition to their host-range diversity, they also exhibit new species) and phylogenetic analyses (e.g.
tremendous cytogenetic variation (aneuploidy and Castillo et al., 2003; Landa et al., 2008). With the
polyploidy) and mode of reproduction (from incorporation of molecular sources of characters
obligatory amphimixis to meiotic and mitotic par- and refinements to phylogenetic theory, the fields
thenogenesis) (Triantaphyllou, 1985; see Chitwood of taxonomy and evolutionary biology have now
and Perry, Chapter 8, this volume). In current become more completely integrated as a dis-
practice, identification of species is based primar- cipline, such that the terms molecular taxonomy
ily on the morphological features of females, and phylogenetics are (or should be) subsumed as
males and second-stage juveniles (Eisenback and a single research programme (systematics). In this
Triantaphyllou, 1991; see Hunt and Handoo, chapter we present a summary of early and con-
Chapter 3, this volume), as well as esterase and temporary research on the molecular systematics
malate dehydrogenase isozyme profiles derived of Meloidogyne.
from single females by polyacrylamide gel elecro-
phoresis (Esbenshade and Triantaphyllou, 1985,
1990; Carneiro et al., 2000) and DNA-based bar- 5.2. The History of Reconstructing
codes (Powers and Harris, 1993; Powers, 2004; Meloidogyne Phylogenetic History
Powers et al., 2005; see Blok and Powers, Chapter
4, this volume). Historically, the diagnostic fea- Some of the earliest work on evolutionary relation-
tures deemed most valuable for identification ships among species of Meloidogyne was based on
commonly preceded their use as important char- morphological characteristics and relied heavily on
acters for taxonomic statements (such as new spe- many of the characters used for identification

©CAB International 2009. Root-knot Nematodes (eds R.N. Perry, M. Moens and J.L. Starr) 119

Perry_Chap-05.indd 119 6/15/2009 3:28:26 PM

 120 B. Adams et al.

(Eisenback and Triantaphyllou, 1991). Subsequent monly explained as lineage sorting among mito-
efforts involved cytogenetics (Triantaphyllou, 1966, chondrial haplotypes, but, in addition, paralogous
1985), producing evidence that supported hypo- genes, such as would be expected for the rDNA
theses consistent with the idea that mitotic tandem array if intraspecific concerted evolution
parthenogens evolved from meiotic parthenoge- were non-uniform, or non-orthologous genes go
netic ancestors, following suppression of mei- undetected (Maddison, 1997; Maddison and
otic processes and establishing various ploidy Knowles, 2006). Phylogenomic analyses hold the
levels. Triantaphyllou also hypothesized that the promise of resolving problematic phylogenies by
amphimictic species in the genus, such as Meloidogyne swamping the data sets with signal, despite high
exigua, are highly specialized parasites and should noise, by including character information from
not be considered as ancestral forms (Triantaphyllou, numerous loci (Eisen, 1998; Eisen and Fraser,
1985). Cytogenetic studies, followed by protein 2003), but see Longhorn et al. (2007). Although
and DNA analyses, implied a unique origin of, and phylogenomic analyses that could exhaustively
monophyly among, the ameiotic species (Dickson sample all Meloidogyne are premature (if not unnec-
et al., 1971; Dalmasso and Bergé, 1978; Esbenshade essary), in a preliminary effort of this kind Scholl
and Triantaphyllou, 1987; Castagnone-Sereno and Bird (2005) identified numerous putative
et al., 1993; Baum et al., 1994; van der Beek et al., homologues and used them to generate a phylog-
1998). Later, studies based on mitochondrial genes eny for a subset of Meloidogyne species. Although
soon revealed that mitochondrial genes can be this effort was based on a small sample of taxa, the
hypervariable, both in patterns of sequence substi- major contribution was the elucidation of relation-
tution and in gene content and arrangement ships among three mitotic parthenogens (M. hapla,
(Powers and Sandall, 1988; Powers et al., 1993). M. incognita, M. javanica) that had been poorly
These properties are desirable for diagnostic or resolved in previous phylogenetic analyses.
population genetic markers, or for resolving phylo- Subsequent refinements to Meloidogyne phylogeny
genetic relationships among closely related species have consisted primarily of analyses that have
(see Blok and Powers, Chapter 4, this volume). added new or previously unsampled taxa to exist-
More recent phylogenetic analyses have utilized ing databases (Castillo et al., 2003; Landa et al.,
small ribosomal subunit (18S) rDNA sequences 2008).
(De Ley et al., 2002), large subunit (28S) rDNA
(Castillo et al., 2003), and mitochondrial DNA
(mtDNA) sequences (Tigano et al., 2005). Lunt
5.3. Molecular Phylogenetics:
(2008) performed separate analyses of four genes –
dystrophin, elongation factor 1-alpha, major sperm
Genetic Markers and Evolutionary
protein, and RNA polymerase 2 – as part of a Relationships
clever study to elucidate whether the origin of the
asexual Meloidogyne lineages was ancient or recent 5.3.1 Nuclear Ribosomal DNA Sequences
(Adams and Powers, 1996; Hugall et al., 1999).
A phenomenon that emerges from each of these Nuclear ribosomal DNA is currently the most
studies is the close relationships among the three extensively employed molecular marker for
major mitotic parthenogenetic species: Meloidogyne Meloidogyne molecular systematics. Variation in
arenaria, Meloidogyne javanica and Meloidogyne incognita. mutation rates observed among different genes
Regardless of the type of phylogenetic analysis per- and spacers within an rDNA transcription unit
formed, or the genetic locus examined, the pre- results in regions of adjacent DNA segments in
ponderance of evidence from single gene analyses the cistron that are useful across a wide range of
suggests that the mitotic parthenogens are clearly taxonomic hierarchical levels (Hillis and Dixon,
evolutionarily distinct from either the meiotic or 1991). This includes conserved and variable
obligatory amphimictic species. However, gene regions of the 18S and 28S subunits, and the
trees are not always concordant with the evolution- more highly variable ITS region. These three
ary history of independently evolving species, and rRNA gene regions are the most commonly used
the discordance between the two different histories genetic markers for nematode molecular system-
can confound phylogenetic inference. Discordance atics, and each of these regions has been employed
between gene trees and species trees is most com- for Meloidogyne phylogenetics (Landa et al., 2008).

Perry_Chap-05.indd 120 6/15/2009 3:28:26 PM

 Molecular Taxonomy and Phylogeny 121

Whilst rRNA genes may provide optimal lev- bases, De Ley et al. (2002) were the first to use
els of variation for investigating Meloidogyne phy- this locus as part of a rigorous reconstruction of
logeny, they are not without significant theoretical Meloidogyne phylogeny. Their analysis included 12
and analytical drawbacks. The single biggest obs- species of Meloidogyne and four outgroup taxa
tacle in using rRNA genes is that, unlike protein- subject to phylogenetic analyses generated from
coding genes, they are not constrained to maintain three different multiple sequence alignment
codon fidelity or even an open reading frame. methodologies and three different tree-building
Whereas the length and composition of protein- optimality criteria (distance, parsimony and max-
coding genes are generally subject to selection by imum likelihood). Calculations of phylogenetic
codon usage, rRNA genes are not. Thus, for some signal (skewness of tree length distribution) were
rDNA regions, insertion and deletion events high and intraspecific sequence polymorphism
(indels) can be as frequent as transitions and trans- low, suggesting that the locus was appropriately
versions, often involving blocks of multiples of robust for resolving relationships among the sam-
nucleotides (Powers, 2004; Powers et al., 1997). pled species but with nodal support strongest at
Indel events can result in substantial rDNA size the deeper nodes. Their analysis showed strong
differences between sequences (taxa), which com- support for three clades, which they designated:
plicates the process of generating multiple sequence clade I, conscribing the mitotic parthenogens
alignments and reduces confidence in the hom- (M. incognita, M. arenaria and M. javanica); clade II,
ology statements for each nucleotide in the multi- including the obligatory amphimictic, meiotic
ple sequence alignment. In our experience, and as and mitotic parthenogens (Meloidogyne hapla races
shown by others, there is usually more variation in A and B, respectively) as well as Meloidogyne duytsi
tree topology due to differences in the multiple and Meloidogyne maritima; and clade III, containing
sequence alignment than there is among the differ- three meiotic parthenogens (Meloidogyne exigua,
ent methods used to generate the trees (i.e. parsi- Meloidogyne graminicola and Meloidogyne chitwoodi).
mony, maximum likelihood, Bayesian and distance As with earlier phylogenetic efforts (Triantaphyllou,
methods) (Morrison and Ellis, 1997). Approaches 1985; Castagnone-Sereno et al., 1993), their
to addressing this problem require thoughtful con- analysis supported the location of the ameiotic
sideration of the mechanics of how multiple species as distantly related to either the obligate
sequence alignments (homology statements) are amphimictic or meiotic species (Fig. 5.1).
constructed. These involve the nuts and bolts of
how computer algorithms generate multiple M. ichinohei
sequence alignments, removing the alignment-
M. artiellia
ambiguous regions based on an a priori metric (i.e.
remove ambiguous indels that lie between a prede- M. graminicola
termined number of invariant nucleotides (Nguyen M. chitwoodi
et al., 2001) ), direct optimization (Terry and
Whiting, 2005), comparison of secondary structure M. exigua
based on minimum energy models (Subbotin et al., M. incognita
2006), and minimum posterior probabilities among
M. arenaria
alternative placements of nucleotides (characters)
in the alignment (Loytynoja and Milinkovitch, M. javanica
2003). These problems are not unique to rDNA, M. hapla
as alignment ambiguity can also arise where
protein-coding genes have undergone tremendous M. microtyla
divergence, or for other non-coding sequences M. maritima
(such as non-coding regions of mitochondrial
M. duytsi
DNA; see section 5.3.3).
Fig. 5.1. 18 s rDNA phylogeny of De Ley et al.
5.3.1.1. 18S (small ribosomal subunit) (2002). The tree was generated from a
secondary-structure-based multiple sequence
Although 18 s sequences for several Meloidogyne alignment and resolved via maximum parsimony.
species had long been available in public data- (Adapted from De Ley et al., 2002.)

Perry_Chap-05.indd 121 6/15/2009 3:28:26 PM

 122 B. Adams et al.

Subsequent 18 s analyses of Tigano et al. (2005) M. artiellia

included 19 additional sequences, representing
M. baetica
12 nominal species, as well as several unknown
isolates from disparate geographic locations, in a M. hapla
effort that revealed interesting comparisons
M. dunensis
between the 18 s rDNA sequences and other tools
for diagnosing species, including morphological M. exigua
and isozyme phenotypes. Their 18 s rDNA ana-
M. graminicola
lysis proceeded from sequences profile-aligned to
the optimal sequence alignment of De Ley et al. M. trifoliophila
(2002) by distance, parsimony and maximum
M. chitwoodi
likelihood tree-building algorithms, all of which
produced congruent topologies (Fig. 5.2). Their M. hispanica
results revealed polymorphisms between isolates
M. arenaria

M. incognita
M. ichinohei
M. paranaensis
M. artiellia
M. konaensis
M. javanica
Fig. 5.3. 18 s rDNA phylogeny of Landa et al.
M. arenaria
(2008). The tree was generated from sequences
M. incognita aligned using Bionumerics software ver. 4.5
(Applied Maths, Kortrijk, Belgium) and resolved via
M. floridensis
maximum parsimony (also by Bionumerics).
M. morocciensis (Adapted from Landa et al., 2008.)
M. paranaensis
M. arabicida of the same species (as per morphological and
M. ethiopica isozyme diagnosis), including M. arenaria, M.
incognita, M. javanica, M. exigua and M. hapla. Some
M. enterolobii of the differences were slight, but several were
M. hapla substantial: M. arenaria of De Ley et al. (2002) dif-
fered from the Tigano et al. (2002) sequences by
M. microtyla
two substitutions and four insertions; two M.
M. maritima incognita isolates differed by one insertion and ten
M. duytsi substitution events; two M. javanica sequences dif-
fered by 24 substitutions, 18 ambiguities and four
M. chitwoodi insertions. However, 18 s rDNA sequences can
M. oryzae also be highly conserved between species. For
example, Meloidogyne hispanica and Meloidogyne ethi-
M. graminicola
opica yield identical 18 s sequences, so only M.
M. exigua hispanica is listed in the analysis of Landa et al.
(2008) (Fig. 5.3).
Fig. 5.2. 18 s rDNA phylogeny of Tigano et al.
(2005). The tree was generated from sequences
profile-aligned to the secondary-structure-based 5.3.1.2. 28S (large ribosomal subunit)
alignment of De Ley et al. (2002) and resolved
The first effort to use the 28 s rDNA region
via maximum likelihood. (Adapted from Tigano
et al., 2005.) Only nominal taxa were retained.
to resolve phylogenetic relationships among
Taxa with multiple representative sequences Meloidogyne spp. was that of Tenente et al. (2004).
were represented by a single semaphoront and Their analysis included nine species represented
relationships collapsed to their most inclusive by 12 sequences. Their most robust phylogenetic
clade. analyses included exhaustive, unrooted maximum

Perry_Chap-05.indd 122 6/15/2009 3:28:26 PM

 Molecular Taxonomy and Phylogeny 123

parsimony searches and maximum likelihood groups I and III form a clade with the more
searches based on various ClustalX alignments. ancestral lineage comprising Meloidogyne artiellia
The data for each of the generated alignments and Meloidogyne baetica (representatives of clade II
revealed strong phylogenetic signal as based on were not included in the analysis). Like Tenente
g1 statistics, and most nodes, particularly for the et al. (2004), Castillo et al. (2003) and Landa et al.
deeper clades, were strongly supported by boot- (2008) were unable to resolve relationships among
strap resampling estimates. Their analysis found species of clade I but resolved deeper nodes with
fairly strong support for two monophyletic clades much greater support (Figs 5.5 and 5.6).
that are compatible with clades I and III of the
18 s analysis of De Ley et al. (2002) (but did not 5.3.1.3. ITS (internally transcribed
include any of the clade II taxa) (Fig. 5.4). The spacer region)
weakest supported nodes, and the ones that dif-
fered the most by alignment and tree search Early work by Hugall et al. (1999) on Meloidogyne
strategy, involved the relationships among the ITS rDNA sequences revealed a somewhat sur-
mitotic parthenogenetic species M. arenaria, M. prising phenomenon: although amphimictic spe-
incognita, Meloidogyne konaensis, and Meloidogyne para- cies exhibited only a single ITS lineage, the
naensis. They conclude that the D2/D3 region of ameiotic species M. hapla, M. arenaria and M.
this marker, which showed high degrees of varia- incognita exhibited numerous lineages, even within
tion between two species of Acrobleoides that are individuals. In fact, Hugall et al. (1999) showed
morphologically virtually indistinguishable (De that up to 90% of the total ITS diversity could be
Ley et al., 1999), were ‘simply too conserved for found within an individual nematode, which con-
the phylogenetic analysis of mitotic parthenoge- tained as many as 9–13 different sequence vari-
netic Meloidogyne species’ (Tenente et al., 2004). ants. Such variation, and the way it is partitioned
A contemporary analysis by Castillo et al. (2003) across the genus, poses challenges to using it to
also reflected these sentiments. Their analysis, infer phylogenetic relationships but does provide
which included the Tenente et al. sequences, also strong evidence for the hybrid origins of M. hapla,
included several other unpublished Meloidogyne M. arenaria and M. incognita (Hugall et al., 1999).
and outgroup (Pratylenchus) sequences, and was
based on simple ClustalX default alignment
M. arenaria
parameters under the maximum parsimony opti-
mality criterion. Their results are consistent with M. paranaensis
Tenente et al. (2004), whereby monophyletic
M. incognita
M. arenaria
M. konaensis
M. konaensis
M. trifoliophila
M. incognita
M. graminicola
M. paranaensis
M. exigua
M. chitwoodi
M. chitwoodi
M. graminicola
M. artiellia
M. trifoliophila
M. baetica
M. exigua Fig. 5.5. 28 s rDNA phylogeny of Castillo et al.
(2003). The tree is a strict consensus of five
Fig. 5.4. 28 s rDNA phylogeny of Tenente et al. equally parsimonious trees generated from
(2004). The tree was generated from sequences sequences aligned using ClustalX default
aligned using ClustalX (Thompson et al., 1997) parameters (Thompson et al., 1997) and resolved
default parameters and resolved using maximum via maximum parsimony (Swofford, 2002).
parsimony. (Adapted from Tenente et al., 2004.) (Adapted from Castillo et al., 2003.)

Perry_Chap-05.indd 123 6/15/2009 3:28:26 PM

 124 B. Adams et al.

M. fallax M. trifoliophila
M. minor
M. chitwoodi M. chitwoodi
M. oryzae
M. graminicola M. hapla
M. naasi
M. maritima M. arenaria
M. hapla
M. partityla M. incognita
M. dunensis
M. microtyla M. javanica
M. arenaria
M. incognita M. baetica
M. avanica
M. morocciensis M. artiellia
M. floridensis
M. hispanica Fig. 5.7. ITS rDNA phylogeny of Castillo et al.
M. ethiopica (2003). The tree was generated from sequences
M. arabicida aligned using ClustalX default parameters
M. exigua (Thompson et al., 1997) and resolved via
M. enterolobii maximum parsimony (Swofford, 2002). The original
M. paranaensis solution was presented as unrooted, we root it
M. duytsi here with Meloidogyne artiellia and Meloidogyne
M. graminis baetica. (Adapted from Castillo et al., 2003.)
M. ardenensis

Fig. 5.6. 28 s rDNA phylogeny of Landa et al.

(2008). The tree was generated from sequences
aligned using Bionumerics software ver. 4.5
(Applied Maths, Kortrijk, Belgium) and resolved via M. panyuensis
maximum parsimony (also by Bionumerics). M. artiellia
(Adapted from Landa et al., 2008.)
M. baetica
M. fallax
M. chitwoodi
One of the by-products of exploring ITS sequence
M. minor
variation in search of diagnostic markers is a
large, publicly available database of sequences. In M. trifoliophila
an early effort, Castillo et al. (2003) generated an M. naasi
unrooted maximum parsimony tree from ITS M. dunensis
sequences for nine species. Subsequently, Landa M. hapla
et al. (2008) used these and additional sequences
M. hispanica
to perform a phylogenetic analysis of 16 species
from 29 different isolates. Although details of M. enterolobii
their sequence alignment process is not explicit, M. javanica
the maximum parsimony tree they generated M. incognita
depicts a completely resolved tree but with vary-
M. arenaria
ing levels of support, particularly through the
intermediate nodes. Clades III and I are mono- M. thailandica
phyletic. Clade II is depicted as paraphyletic, but
Fig. 5.8. ITS rDNA phylogeny of Landa et al.
the discordant nodes are also those that are most (2008). The tree was generated from sequences
weakly supported. Also evident in the resulting aligned using Bionumerics software ver. 4.5
tree is the apparent paraphyletic nature of the (Applied Maths, Kortrijk, Belgium) and resolved via
ITS lineages within and among the ameiotic spe- maximum parsimony (also by Bionumerics).
cies (Figs 5.7 and 5.8). (Adapted from Landa et al., 2008.)

Perry_Chap-05.indd 124 6/15/2009 3:28:26 PM

 Molecular Taxonomy and Phylogeny 125

5.3.2. Orthologous Nuclear Genes M. chitwoodi

As with gene families, concerted evolution does M. fallax

not always work fast enough to homogenize all
M. hapla
copy variants in the rDNA cistron. The result is
that paralogous gene genealogies can contradict
M. enterolobii
actual phylogenetic relationships of species. This
is evident by lineage sorting events for mtDNA M. arenaria
(discussed below) and also by species that arose
via hybridization events, such as the mitotic par- M. incognita
thenogenetic species of Meloidogyne (Hugall et al.,
1999). Thus, an optimal phylogenetic analysis M. javanica
will include as many nuclear, single-copy genes in
the transformation series as possible. In an effort Fig. 5.9. Dystrophin exon gene genealogy of Lunt
to distinguish the relative timing of the hybridiza- (2008). The tree was generated from exon
tion events involved in the origin of the ameiotic sequences aligned with Clustal X (Thompson
lineages, Lunt (2008) explored the evolution of et al., 1997) (unambiguous alignment due to
codon structure) and resolved using maximum
four different single-copy nuclear genes. Prior to
likelihood as implemented in PhyML (Guindon
this effort, Scholl and Bird (2005) undertook a
and Gascuel, 2003).
phylogenomic approach in order to tease out the
relationships between the mitotic parthenogenetic
Meloidogyne species. Each of these efforts is dis- the cell’s motility (Roberts, 2005). Lunt (2008)
cussed below. studied Meloidogyne msp gene genealogies in order
to see if the gene underwent an accelerated
5.3.2.1. Dystrophin rate of mutation after the evolution of mitotic
parthenogenesis and, therefore, putatively no
Dystrophin is a muscle protein that connects the longer under selection pressure to maintain its
cytoskeleton of a muscle fibre to the surrounding function; it did not. Maximum parsimony analy-
extracellular matrix. It is the longest gene in the sis from multiple populations of five Meloidogyne
human genome but exists in single-copy form in species (M. incognita, M. javanica, M. arenaria,
invertebrates (Roberts et al., 1995; Roberts and M. enterolobii, M. hapla) and an outgroup taxon
Bobrow, 1998). Lunt (2008) designed primers (G. pallida) yielded a phylogenetic tree that is
that amplified a 670–770 bp product of three consistent with the three clade designation of De
exons and two introns from multiple populations Ley et al. (2003), except that the positions of
of five Meloidogyne species (M. incognita, M. javanica, clade II and III are reversed relative to clade I,
M. arenaria, Meloidogyne enterolobii (=Meloidogyne but the node involved in this reversal is not well
mayaguensis) and M. hapla) and an outgroup taxon supported (Fig. 5.10).
(Globodera pallida) for phylogenetic analyses using
maximum likelihood. The results for this gene 5.3.2.3. Elongation factor 1-alpha (EF1-a)
are consistent with the clade designations of De
Ley et al. (2002) but with failure to resolve unam- Elongation factor 1-alpha (EF-1a), the GTP
biguously relationships among the members of binding protein involved in catalysing the bind-
clade I (M. enterolobii, M. arenaria, M. incognita, M. ing of the aminoacyl-transfer RNAs to the ribos-
javanica). However, none of the paraphyletic ome, is an essential component of eukaryotic
nodes is well supported by approximate likeli- translation (Watson, 2008). Lunt (2008) pursued
hood ratio tests (Fig. 5.9). this gene as an independent estimate of phylogen-
etic relationships but, upon inspection of the
5.3.2.2. Major sperm protein (msp) resulting phylogenetic tree, suspected that instead
of being single copy, a gene duplication event
Major sperm protein is the most abundant pro- might have occurred, resulting in paralogous loci.
tein in nematode sperm and is responsible for In support of this, Lunt identified two copies in

Perry_Chap-05.indd 125 6/15/2009 3:28:26 PM

 126 B. Adams et al.

M. hapla 5.3.2.4. RNA polymerase 2

RNA polymerase 2 is the eukaryotic enzyme

M. chitwoodi responsible for synthesis of mRNA during tran-
scription (Kornberg, 2007). Lunt (2008) designed
primers that amplified an approximately 710 bp
M. javanica fragment that included coding sequence from two
exons and one intron. Using the exon sequences
only resulted in the maximum likelihood solution
M. incognita in Fig. 5.12. The tree is consistent with the clade
designations of De Ley et al. (2002) but fails to
resolve M. chitwoodi and M. fallax lineages within
M. arenaria
Clade III, and where they are resolved, the M.
Fig. 5.10. Major sperm protein exon gene javanica, M. incognita or M. arenaria relationships
genealogy of Lunt (2008). The tree was generated within Clade I are paraphyletic.
from exon sequences aligned with Clustal X
(Thompson et al., 1997) (unambiguous alignment
due to codon structure) and resolved using 5.3.3. Mitochondrial DNA
maximum likelihood as implemented in PhyML
(Guindon and Gascuel, 2003).
Despite apparent low genetic diversity (or because
of it; see Blok and Powers, Chapter 4, this vol-
the Caenorhabditis elegans genome, yet the maximum ume) among populations of M. arenaria, M. javanica
likelihood solution among his sampled species and M. incognita (Hugall et al., 1994, 1997; Stanton
was still congruent with the clade designations of et al., 1997), mtDNA sequences spanning the
De Ley et al. (2002). However, the EF-1a gene COII through lRNA genes have been intensively
genealogy among the ameiotic species is para- studied for Meloidogyne molecular diagnostics (Blok
phyletic, poorly supported and too conserved to et al., 2002; Powers, 2004; Brito et al., 2004;
distinguish Meloidogyne fallax from M. chitwoodi Powers et al., 2005). The diagnostic utility of the
(Fig. 5.11). marker persuaded Tigano et al. (2005) to explore

M. chitwoodi M. chitwoodi

M. fallax
M. hapla

M. hapla
M. enterolobii
M. arenaria
M. javanica
M. incognita

M. incognita
M. javanica

M. enterolobii M. arenaria

Fig. 5.11. Elongation factor 1-alpha exon gene Fig. 5.12. RNA polymerase 2 exon gene
genealogy of Lunt (2008). The tree was generated genealogy of Lunt (2008). The tree was generated
from exon sequences aligned with Clustal X from exon sequences aligned with Clustal X
(Thompson et al., 1997) (unambiguous alignment (Thompson et al., 1997) (unambiguous alignment
due to codon structure) and resolved using due to codon structure) and resolved using
maximum likelihood as implemented in PhyML maximum likelihood as implemented in PhyML
(Guindon and Gascuel, 2003). (Guindon and Gascuel, 2003).

Perry_Chap-05.indd 126 6/15/2009 3:28:26 PM

 Molecular Taxonomy and Phylogeny 127

M. hapla sible problems associated with gene tree/species

tree discordance, a reasonable expectation might
M. enterolobii be that if one simply looks at enough characters,
M. arabicida or in this case enough DNA sequence from
enough genes, eventually the phylogenetic signal
M. paranaensis will swamp noise. Such is the idea of using
genomic data to resolve phylogenetic relation-
M. floridensis
ships, where it has been shown that using small
M. incognita numbers of genes can produce support for incor-
rect phylogenies, but that support and resolution
M. ethiopica become optimized at about 20 genes. Taking
this into consideration, Scholl and Bird (2005)
M. javanica
sampled 47 orthologous genes from several
M. arenaria tylenchid nematodes, including M. chitwoodi, M.
hapla, M. arenaria, M. javanica and M. incognita,
M. morocciensis with the goal of resolving the relationships
between the apomicts (M. arenaria, M. javanica
Fig. 5.13. Mitochondrial DNA phylogeny of Tigano
et al. (2005). The tree was generated from DNA
and M. incognita) where previous analyses
sequences that span the COII through lRNA appeared to show conflicting relationships: based
region, including the complete sequence for tRNA- on mtDNA, Powers and Sandall (1988) suggested
His, and the AT-rich region. Sequences were the relationship to be (M. arenaria (M. javanica +
aligned using Clustal X (Thompson et al., 1997), M. incognita), whereas the 18 s best estimate of De
adjusted by eye using MacClade (Maddison and Ley et al. (2002) was (M. incognita (M. javanica +
Maddison, 2002), and resolved by maximum M. arenaria) ). To construct their phylogenetic
parsimony as implemented in PAUP* (Swofford, trees, Scholl and Bird (2005) first performed
2002). rigorous screens to identify orthologous genes
from EST sequence databases that had a homo-
logue in C. elegans (and thus conserved across
its ability to resolve phylogenetic relationships
large phylogenetic distance, reflecting evolution-
among the very closely related Clade I taxa. By
ary constraint). The orthologues were aligned
parsimony, distance and maximum likelihood
based on their inferred amino acid sequence and
approaches they analysed the complete region,
then back-translated to their DNA sequence to
including partial COII and partial lRNA
maintain open reading frame fidelity. Phylogenies
sequence, the complete sequence for tRNA-His
were constructed using multiple alignments of
and the AT-rich region, and obtained a single
the individual genes and concatenated full-
optimal solution. The concatenation of the differ-
length data sets containing all the genes, or dif-
ent gene regions was justified by failure to reject
ferent subsets of genes, via Bayesian inference,
shared evolutionary histories by way of an incon-
maximum likelihood and minimum evolution
gruence length difference test. Although there is
(via neighbour joining). The optimal solution
some discordance between this tree and their
based on all of the analyses favoured the hypoth-
rDNA solution, the relationship among the amei-
esis of Powers and Sandall (1988) (Fig. 5.14).
otic species is congruent, if poorly supported,
with the 18 s rDNA analyses of Tigano et al.
(2005) and De Ley et al. (2002) (Fig. 5.13).
5.4. A Meloidogyne Supertree
Analysis
5.3.4. Phylogenomics
It would be excellent if we could just take all the
Optimally, phylogenetic construction proceeds DNA sequences for all of the Meloidogyne taxa
from consideration of as many independently ever generated, compile them into a single giant
evolving, heritable characters as possible (Farris, multiple sequence alignment, crunch it through
1983; Kluge, 1997). In order to address the pos- some tree-building algorithms and confidently

Perry_Chap-05.indd 127 6/15/2009 3:28:27 PM

 128 B. Adams et al.

M. chitwoodi whether the phylogenies are independent. As

opposed to using morphological or molecular
characters to build phylogenies, the raw data
M. hapla for generating supertrees are lifted from the
topological arrangements of two or more partially
overlapping phylogenetic trees. So just as using
M. arenaria the character ‘male stylet length’ five times in a
character matrix would be redundant, using three
different 18 s trees to build a supertree that was
M. javanica to include seven other genetic loci could artifi-
cially bias the overall supertree topology in favour
of the 18 s topology. In other words, the three
M. incognita 18 s trees are not independent estimates of phylo-
genetic relationships and must be dealt with
Fig. 5.14. Phylogeny of Scholl and Bird (2005). somehow so as not to bias the analysis unfairly.
The tree was generated from a concatenated Similarly, what about different trees that are gen-
matrix of 47 putative orthologous genes aligned
erated by the same author? If Lunt (2008) used
using Clustal W (Thompson et al., 1994) and
resolved using Bayesian analysis as implemented
the same general methodology, from alignment
in MRBAYES (Huelsenbeck and Ronquist, 2001). to tree-building strategy, to construct all of his
phylogenetic trees, is there an element of non-
independence among his different gene trees? In
report the one true tree of Meloidogyne spp. evolu- fact, the majority of Meloidogyne phylogenetic
tionary history. While perhaps optimal, such an analyses to date employed several different
exercise is obviated by the range of sequence approaches in the same publication, from align-
variation that spans the gamut of evolutionary ment strategy to tree-building algorithms and
rates of change on the loci under study. For optimality criteria. More often than not, these
example, for many of the mtDNA and ITS different approaches generated several different
sequences, the divergence and length differences hypotheses of relationships for each data set. Of
prohibit making unambiguous multiple sequence these, which do we choose to use as source trees
alignments among all taxa in the genus. Similarly, to generate a supertree?
not all taxa are present in all of the data sets, and With the above caveats in mind, we gener-
the inclusion of large amounts of missing charac- ated a supertree based on the best estimates of
ters can result in spurious phylogenetic inference phylogenetic relationships from the most robust
(Maddison, 1993; Wiens, 1998). One solution and comprehensive molecular phylogenetic ana-
to this problem is to generate a supertree – an lyses we could find in the literature. In order to
evolutionary tree that is assembled from a bunch account for duplication of gene trees (three 18 s,
of smaller trees that share some, but not neces- three 28 s, and two ITS gene trees) we generated
sarily all, common taxa (Bininda-Emonds, 2004). ‘mini-supertrees’ from each set of redundant gene
Although it may sound straightforward to trees and then used the ‘mini-supertree’ as the
take several phylogenetic trees and spin them source tree for the main analysis (Bininda-Emonds
into a single tree that is the sum of all the parts, et al., 1999). In choosing among the alternative
it is not. What should not be surprising is that the phylogenetic arrangements published in most
quality of the obtained supertree is a function of papers (i.e. the maximum parsimony tree versus
the quality of the phylogenies used to build it. the maximum likelihood tree), we did what any
Even if the phylogenies used to build the super- good taxonomist would do – appeal to authority.
tree are robust, there are still several important We simply chose the hypothesis that was most
aspects of building a supertree to consider; we favoured by the authors.
will only touch on a few here that are relevant to To construct the supertree we mined the
Meloidogyne (but see Bininda-Emonds et al., 2004). optimal topologies from the following publica-
In the case for building a Meloidogyne supertree tions: De Ley et al. (2002), Castillo et al. (2003),
one must consider not only the quality of the dif- Tenente et al. (2004), Scholl and Bird (2005),
ferent phylogenies used to build the tree but also Tigano et al. (2005), Landa et al. (2008) and Lunt

Perry_Chap-05.indd 128 6/15/2009 3:28:27 PM

 Molecular Taxonomy and Phylogeny 129

(2008). To account for non-independent estimates 2005) (Fig. 5.15). Even though these reconstruc-
we generated mini-supertrees from the data sets tions involved the analysis of only nine different
that were represented more than once (18 s, 28 s topologies, their sum of possible unrooted solu-
and ITS). The mtDNA data set was treated as a tions is 2.92156 × 1040, presenting a fairly com-
single estimate, even though it comprised a con- putationally intensive effort. Each of the two
catenation of several genes. While potentially supertree analyses yielded multiple equally parsi-
misleading, we justify this because Tigano et al. monious (MRP) or costly (dfit) trees (1736 and 7,
(2005) tested for shared evolutionary history and respectively), which we represent here using two
common inheritance of each of the individual different consensus approaches (combinable
genes. The phylogenomic analysis of Scholl and components and majority rule) (Fig. 5.15). For
Bird (2005) presents an interesting case because both types of supertree construction methods,
their analysis consisted of 47 concatenated genes. the consensus trees differed only in terms of reso-
In theory, for the present analysis we should be lution (combinable components being more con-
analysing each of the 47 gene tree topologies servative and less resolved). Overall the obtained
independently and then using each one of those topologies among the different supertree meth-
topologies as an independent estimate of relation- ods are quite similar. Both approaches identified
ships (source tree) to construct the supertree. Meloidogyne ichinohei as the lineage that shares a
When Scholl and Bird performed their analyses most recent common ancestry with the remain-
on the concatenated data set of 47 genes, the ing members of the genus, with the next lineage
data matrix was treated as if it were a single to branch being the monophyletic clade of
gene, assuming a single model of evolution (M. baetica (Meloidogyne panyuensis + M. artiellia) ).
(general time-reversible with four categories of Membership in clade III is congruent between
gamma-distributed rate heterogeneity). In reality, the two analyses, with the exception that the
since it is unlikely that all 47 genes evolved under dfit solution suggests sister relationships for
the same model of sequence evolution, it would (Meloidogyne oryzae + M. graminicola) and (Meloidogyne
have been more appropriate to partition the con- trifoliophila + Meloidogyne naasi). Clade II member-
catenated data set by gene, each partition with its ship is monophyletic in the dfit topology, differ-
own most appropriate model of evolution, but ing from the MRP solution, which suggests that
that approach was still under development at the Meloidogyne dunensis and Meloidogyne microtyla are
time of their analyses (Huelsenbeck et al., 2008). sister taxa. Also, the MRP topology differs
However, Scholl and Bird did do independent radically from the dfit solution as Meloidogyne
analyses of subsets of single genes and reported graminis and Meloidogyne ardenensis nest within
that they were congruent with the overall con- Clade II, but in the MRP solution they are
catenated gene phylogeny. Thus, as with the ancestral to Clades I, II and III. Clade I mem-
mtDNA topology, for our supertree analysis we bership is congruent between both analyses, and
used the Scholl and Bird (2005) topology as a lack of resolution is completely understandable
single, independent estimate of relationships, given that these relationships were poorly
acknowledging that were we to use all 47 gene resolved in most of the source trees. Both
topologies (which were identical) that there would approaches favour Meloidogyne floridensis and
be overwhelming support for a (M. arenaria (M. M. incognita as sister taxa (Fig. 5.15B,C), but
javanica + M. incognita) ) clade in the present ana- beyond this relationship there is only discord.
lysis. Similarly, by using the mtDNA data as a The most fundamental difference is that of the
single source tree, we are probably underestimat- relationship between M. arenaria, M. incognita and
ing the overall support for its topology in the M. javanica. The MRP solution favours (M. incog-
supertree. nita (M. arenaria + M. javanica) ), whereas the dfit
We used two supertree construction meth- solution is (M. arenaria (M. javanica + M. incognita) ).
ods: matrix representation of parsimony (Purvis, Because the actual number of ‘characters’ (tree
1995) with the matrix generated in RadCon topologies) is so small (n = 9), bootstrap support
1.1.6 (Thorley and Page, 2000) and implemented for any of the relationships is virtually non-
in PAUP* (Swofford, 2002), and the most similar existent (data not shown). However, phylogenetic
supertree method (dfit) as implemented in signal was significantly better than random (per-
CLANN ver. 3.0.0 (Creevey and McInerney, mutation tail probability test; P < 0.01).

Perry_Chap-05.indd 129 6/15/2009 3:28:27 PM

 130 B. Adams et al.

mrp outgroup
M. chitwoodi
M. fallax
M. minor
M. oryzae
M. graminicola
M. trifoliophila
M. naasi
M. exigua
M. hapla
M. partityla
M. dunensis
M. microtyla
M. maritima
M. duytsi
M. arenaria
M. javanica
M. incognita
M. enterolobii
M. arabicida
M. paranaensis
M. floridensis
M. ethiopica
M. morocciensis
M. hispanica
M. konaensis
M. thailandica
M. graminis
M. ardenensis
M. artiellia
M. panyuensis
M. baetica
M. ichinohei

Fig. 5.15. Supertree solutions for seven independent Meloidogyne source trees. A: MRP combinable
components consensus;

5.5. Conclusions and Future Following the clade designations of De Ley et al.
Directions (2002), clade I includes the mitotic parthenogens,
M. arenaria, M. incognita and M. javanica, as well
Overall, the topologies of the phylogenies we as M. enterolobii, Meloidogyne morociensis, M. ethiopica,
used as source trees are remarkably similar. M. hispanica, M. konaensis, M. paranaensis, Meloidogyne

Perry_Chap-05.indd 130 6/15/2009 3:28:27 PM

 Molecular Taxonomy and Phylogeny 131

mrp outgroup
M. chitwoodi
M. fallax
M. minor
M. oryzae
M. graminicola
M. trifoliophila
M. naasi
M. exigua
M. hapla
M. partityla
M. dunensis
M. microtyla
M. maritima
M. duytsi
M. arenaria
M. javanica
M. incognita
M. floridensis
M. konaensis
M. morocciensis
M. thailandica
M. enterolobii
M. paranaensis
M. arabicida
M. ethiopica
M. hispanica
M. graminis
M. ardenensis
M. artiellia
M. panyuensis
M. baetica
M. ichinohei

Fig. 5.15. (Continued) B: MRP majority rule consensus tree;

thailandica and Meloidogyne arabicida. The only sister to clade II, which contains M. hapla,
known exceptional member of the clade is M. Meloidogyne partityla, M. dunensis, M. microtyla, M.
floridensis, which is described as a meiotic parthe- maritima and M. duytsi. The dfit analysis also
nogen (Handoo et al., 2004) but counter-intuitively includes in this group M. graminis and M. ardenensis.
nests as sister taxon to M. incognita. Clade I is Clades I and II form a clade with respect to clade

Perry_Chap-05.indd 131 6/15/2009 3:28:27 PM

 132 B. Adams et al.

M. ichinohei
M. baetica
M. panyuensis
M. artiellia
M. exigua
M. trifoliophila
M. naasi
M. oryzae
M. graminicola
M. minor
M. chitwoodi
M. fallax
M. duytsi
M. graminis
M. ardenensis
M. dunensis
M. maritima
M. microtyla
M. partityla
M. hapla
M. enterolobii
M. hispanica
M. arabicida
M. thailandica
M. paranaensis
M. konaensis
M. floridensis
M. incognita
M. ethiopica
M. javanica
M. arenaria
M. morocciensis

Fig. 5.15. (Continued) C: dfit combinable components consensus;

III, which contains M. graminicola, M. chitwoodi, their sampled taxa (52 differences between M.
M. exigua, M. trifoliophila, M. naasi, Meloidogyne chitwoodi and its nearest taxon). The analyses of
minor, M. fallax and M. oryzae. Tigano et al. (2005) and De Ley et al. (2002) reveal
Tenente et al. (2004) showed that there was even more amazing variation at the 18 s locus. At
considerable variation in the 28 s region among the conservative end of this spectrum, M. hispanica

Perry_Chap-05.indd 132 6/15/2009 3:28:27 PM

 Molecular Taxonomy and Phylogeny 133

M. ichinohei
M. baetica
M. panyuensis
M. artiellia
M. exigua
M. trifoliophila
M. naasi
M. oryzae
M. graminicola
M. minor
M. chitwoodi
M. fallax
M. duytsi
M. graminis
M. ardenensis
M. dunensis
M. maritima
M. microtyla
M. partityla
M. hapla
M. enterolobii
M. hispanica
M. arabicida
M. paranaensis
M. konaensis
M. thailandica
M. floridensis
M. incognita
M. ethiopica
M. javanica
M. arenaria
M. morocciensis

Fig. 5.15. (Continued) D: dfit majority rule consensus tree.

and M. ethiopica sequences are identical (Landa Many of the differences are attributable to base-
et al., 2008). At the other end is the fact that there call ambiguity that could be resolved with more
can be more variation between two species of persistent sampling and sequencing effort (prob-
Meloidogyne (say, M. maritima and M. artiellia) than ably attributable to intraspecific and intra-
between a human and a platypus (63 differences). individual variation). Still, such discrepancies are

Perry_Chap-05.indd 133 6/15/2009 3:28:27 PM

 134 B. Adams et al.

astonishing when one considers that two organ- What will it take to achieve a fully resolved,
isms identified by nematode taxonomists as the robust Meloidogyne phylogeny with near complete
same species (i.e. two isolates of M. javanica) have representation of all its species? First, and per-
far greater 18 s rDNA sequence divergence (24 haps most importantly, it will take a rigorous
substitutions, 4 indels) than an Australian wom- sampling effort of both genes and taxa. This
bat and a North American possum (3 substitu- Herculean step requires thorough field sampling
tions, 2 indels). Certainly nematodes are reported of genetic variation across the globe, collabora-
to have high rates of sequence divergence relative tive research involving experts in both morpho-
to morphological evolution (Stein et al., 2003; logical and molecular identification, and the
Sudhaus and Kiontke, 2007), and Meloidogyne spp. resources and will to generate enormous amounts
are part of a rapidly evolving clade (Holterman of DNA sequence data for each species. Secondly,
et al., 2006), but such high divergence at the 18 s it must be recognized that because the relative
locus among closely related species, and even amount of DNA sequence divergence is so varied
among individuals within a species, is sufficient between taxa (some extremely high, some
cause for further investigation into the origin(s) of extremely low), phylogenetic analysis of the whole
Meloidogyne species and the morphological and group will require use of suites of genes that
molecular characters considered informative for evolve very slowly for deep nodes and very rap-
diagnosis and species delimitation. idly for shallow nodes of the tree. It is likely that
There are several reasons why the published genes appropriate for resolving relationships
18 s and mtDNA analyses might have returned among closely related species will be inappropri-
different topologies for the Clade I mitotic ate, if not completely alignment ambiguous,
parthenogens. The first, and most obvious, is that among distantly related species. Thus, phyloge-
they represent discordant evolutionary histories. nomic and total evidence analyses using concat-
This notion was first empirically tested by Tigano enated data sets will probably be highly
et al. (2005), who performed an incongruence informative for resolving relationships among
length difference test (Farris et al., 1994, 1995) (but deeper nodes, but unless they can also sample
see Hipp et al., 2004; Barker and Lutzoni, 2002; variation within and between populations, meta-
Darlu and Lecointre, 2002) on their mtDNA and analyses and supertree construction are likely to
18 s rDNA sequences and found significant differ- be required to assemble a tree with the greatest
ences, sufficient to conclude that the two data sets explanatory power.
did not share a common evolutionary history. This The observed inability to obtain mono-
observation could simply be due to the fundamen- phyletic relationships among ITS rDNA sequences
tal nature of the two markers. 18 s sequences reside (Hugall et al., 1999) and/or ‘alleles’ of putative
within a tandemly repeated cistron that is subject single-copy nuclear loci (Lunt, 2008) from hybri-
to mutation and gene conversion through con- dogenic lineages is completely consistent with the
certed evolution, which could result in paralogous, expected fate of such genes upon phylogenetic
and not orthologous, gene sequences (Slowinski analysis. In fact, it is highly likely that even the
and Page, 1999). Alternatively, it is possible that resolution of these genealogies exhibited by the
lineage sorting of mitochondrial haplotypes, muta- phylogenomic analysis of Scholl and Bird (2005)
tion rate heterogeneity or sampling error from the would dissolve upon further sampling of ‘allelic’
small number of phylogenetically informative variation among additional individuals of these
nucleotide bases examined could result in discord- species from disparate populations. Simply put,
ance between gene and species trees (Maddison, the evolutionary lineages that comprise these lin-
1997; Funk and Omland, 2003; Avise, 2007).The eages may not have unique evolutionary origins
problem of resolving phylogenetic relationships or fates and thus are not only intractable phylo-
among the mitotic parthenogenetic species goes genetically but also ontologically (Ghiselin, 1997;
far beyond differential lineage sorting and gene Adams, 2001). Regardless, the resolution of his-
conversion, and is most certainly compounded by torical relationships among the genes that com-
their probable hybrid origins. It has long been sus- prise the ‘species’ remains the single most
pected that M. incognita, M. arenaria and M. javanica powerful tool in the arsenal of comparative meth-
arose through hybridization events between sexual ods for understanding the origin and evolution of
or meiotic parthenogenic taxa (Triantaphyllou, what are arguably the most perplexing, and vex-
1985; Castagnone-Sereno et al., 1993). ing, nematodes on earth.

Perry_Chap-05.indd 134 6/15/2009 3:28:27 PM

 Molecular Taxonomy and Phylogeny 135

5.6. References

Adams, B.J. (2001) The species delimitation uncertainty principle. Journal of Nematology 33, 153–160.
Adams, B.J. and Powers, T.O. (1996) An rDNA phylogenetic analysis of the genus Meloidogyne: a recent
origin of mitotic parthenogenetic species. Nematropica 26, 193–194 (abstract).
Avise, J.C. (2007) On Evolution. The Johns Hopkins University Press, Baltimore, Maryland.
Barker, F.K. and Lutzoni, F.M. (2002) The utility of the incongruence length difference test. Systematic
Biology 51, 625–637.
Baum, T.J., Gresshoff, P.M., Lewis, S.A. and Dean, R.A. (1994) Characterization and phylogenetic analysis
of 4 root-knot nematode species using DNA amplification fingerprinting and automated polyacrylamide
gel electrophoresis. Molecular Plant-Microbe Interactions 7, 39–47.
Bininda-Emonds, O.R.P. (2004) Phylogenetic Supertrees: Combining Information to Reveal the Tree of Life.
Kluwer Academic Publishers, Dordrecht, The Netherlands.
Bininda-Emonds, O.R.P., Gittleman, J.L. and Purvis, A. (1999) Building large trees by combining phylogen-
etic information: a complete phylogeny of the extant Carnivora (Mammalia). Biological Reviews 74,
143–175.
Bininda-Emonds, O.R.P., Jones, K.E., Price, S.A., Cardillo, M.G.R. and Purvis, A. (2004) Garbage in, gar-
bage out: data issues in supertree construction. In: Bininda-Emonds, O.R.P. (ed.) Phylogenetic
Supertrees: Combining Information to Reveal the Tree of Life. Kluwer Academic Publishers, Dordrecht,
the Netherlands, pp. 267–280.
Blok, V.C., Wishart, J., Fargette, M., Berthier, K. and Phillips, M.S. (2002) Mitochondrial DNA differences
distinguishing Meloidogyne mayaguensis from the major species of tropical root-knot nematodes.
Nematology 4, 773–781.
Brito, J., Powers, T.O., Mullin, P.G., Inserra, R.N. and Dickson, D.W. (2004) Morphological and molecular
characterization of Meloidogyne mayaguensis isolates from Florida. Journal of Nematology 36,
232–240.
Carneiro, R.M.D.G., Almeida, M.R.A. and Quénéhervé, P. (2000) Enzyme phenotypes of Meloidogyne spp.
populations. Nematology 2, 645–654.
Castagnone-Sereno, P., Piotte, C., Uijthof, J., Abad, P., Wajnberg, E., Vanlerberghe Masutti, F., Bongiovanni, M.
and Dalmasso, A. (1993) Phylogenetic relationships between amphimictic and parthenogenetic
nematodes of the genus Meloidogyne as inferred from repetitive DNA analysis. Heredity 70, 195–204.
Castillo, P., Vovlas, N., Subbotin, S. and Troccoli, A. (2003) A new root-knot nematode, Meloidogyne baetica
n. sp (Nematoda : Heteroderidae), parasitizing wild olive in Southern Spain. Phytopathology 93,
1093–1102.
Creevey, C.J. and Mcinerney, J.O. (2005) Clann: investigating phylogenetic information through supertree
analyses. Bioinformatics 21, 390–392.
Dalmasso, A. and Bergé, J.B. (1978) Molecular polymorphism and phylogenetic relationship in some
Meloidogyne spp.: application to taxonomy of Meloidogyne. Journal of Nematology 10, 323–332.
Darlu, P. and Lecointre, G. (2002) When does the incongruence length difference test fail? Molecular
Biology and Evolution 19, 432–437.
De Ley, I.T., De Ley, P., Vierstraete, A., Karssen, G., Moens, M. and Vanfleteren, J. (2002) Phylogenetic
analyses of Meloidogyne small subunit rDNA. Journal of Nematology 34, 319–327.
De Ley, P., Felix, M.A., Frisse, L.M., Nadler, S.A., Sternberg, P.W. and Thomas, W.K. (1999) Molecular and
morphological characterisation of two reproductively isolated species with mirror-image anatomy
(Nematoda : Cephalobidae). Nematology 1, 591–612.
Dickson, D.W., Huisingh, D. and Sasser, J.N. (1971) Dehydrogenases, acid and alkaline phosphatases,
and esterases for chemotaxonomy of selected Meloidogyne, Ditylenchus, Heterodera, and
Aphelenchus spp. Journal of Nematology 3, 1–16.
Eisen, J.A. (1998) Phylogenomics: improving functional predictions for uncharacterized genes by evolution-
ary analysis. Genome Research 8, 163–167.
Eisen, J.A. and Fraser, C.M. (2003) Phylogenomics: intersection of evolution and genomics. Science 300,
1706–1707.
Eisenback, J.D. and Triantaphyllou, H.H. (1991) Root-knot nematodes: Meloidogyne species and races. In:
Nickle, W.R. (ed.) Manual of Agricultural Nematology. Marcel Dekker, New York, pp. 191–274.
Esbenshade, P.R. and Triantaphyllou, A.C. (1985) Use of enzyme phenotypes for identification of
Meloidogyne species. Journal of Nematology 17, 6–20.
Esbenshade, P.R. and Triantaphyllou, A.C. (1987) Enzymatic relationships and evolution in the genus
Meloidogyne (Nematoda, Tylenchida). Journal of Nematology 19, 8–18.

Perry_Chap-05.indd 135 6/15/2009 3:28:27 PM

 136 B. Adams et al.

Esbenshade, P.R. and Triantaphyllou, A.C. (1990) Isozyme phenotypes for the identification of Meloidogyne
species. Journal of Nematology 22, 10–15.
Farris, J.S. (1983) The logical basis of phylogenetic analysis. In: Platnick, N.I. and Funk, V.A. (eds) Advances
in Cladistics. Columbia University Press, New York, pp. 570–572.
Farris, J.S., Kallersjo, M., Kluge, A.G. and Bult, C. (1994) Testing significance of incongruence. Cladistics
10, 315–319.
Farris, J.S., Källersjö, M., Kluge, A.G. and Bult, C. (1995) Constructing a significance test for incongruence.
Systematic Biology 44, 570–572.
Funk, D.J. and Omland, K.E. (2003) Species-level paraphyly and polyphyly: frequency, causes, and conse-
quences, with insights from animal mitochondrial DNA. Annual Review of Ecology Evolution and
Systematics 34, 397–423.
Ghiselin, M.T. (1997) Metaphysics and the Origin of Species. State University of New York Press, Albany,
New York.
Guindon, S. and Gascuel, O. (2003) A simple, fast, and accurate algorithm to estimate large phylogenies
by maximum likelihood. Systematic Biology 52, 696–704.
Handoo, Z.A., Nyczepir, A.P., Esmenjaud, D., van der Beek, J.G., Castagnone-Sereno, P., Carta, L.K.,
Skantar, A.M. and Higgins, J.A. (2004) Morphological, molecular, and differential-host characteriza-
tion of Meloidogyne floridensis n. sp (Nematoda : Meloidogynidae), a root-knot nematode parasitizing
peach in Florida. Journal of Nematology 36, 20–35.
Hillis, D.M. and Dixon, M.T. (1991) Ribosomal DNA: molecular evolution and phylogenetic inference. The
Quarterly Review of Biology 66, 411–453.
Hipp, A.L., Hall, J.C. and Sytsma, K.J. (2004) Congruence versus phylogenetic accuracy: revisiting the
incongruence length difference test. Systematic Biology 53, 81–89.
Holterman, M., van der Wurff, A., van den Elsen, S., van Megen, H., Bongers, T., Holovachov, O., Bakker,
J. and Helder, J. (2006) Phylum-wide analysis of SSU rDNA reveals deep phylogenetic relationships
among nematodes and accelerated evolution toward crown clades. Molecular Biology and Evolution
23, 1792–1800.
Huelsenbeck, J.P. and Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogenetic trees.
Bioinformatics 17, 754–755.
Huelsenbeck, J.P., Joyce, P., Lakner, C. and Ronquist, F. (2008) Bayesian analysis of amino acid substitution
models. Philosophical Transactions of the Royal Society B – Biological Sciences 363, 3941–3953.
Hugall, A., Moritz, C., Stanton, J. and Wolstenholme, D.R. (1994) Low, but strongly structured
mitochondrial-DNA diversity in root-knot nematodes (Meloidogyne). Genetics 136, 903–912.
Hugall, A., Stanton, J. and Moritz, C. (1997) Evolution of the AT-rich mitochondrial DNA of the root knot
nematode, Meloidogyne hapla. Molecular Biology and Evolution 14, 40–48.
Hugall, A., Stanton, J. and Moritz, C. (1999) Reticulate evolution and the origins of ribosomal internal tran-
scribed spacer diversity in apomictic Meloidogyne. Molecular Biology and Evolution 16, 157–164.
Kluge, A.G. (1997) Testability and the refutation and corroboration of cladistic hypotheses. Cladistics – the
International Journal of the Willi Hennig Society 13, 81–96.
Kornberg, R.D. (2007) The molecular basis of eukaryotic transcription. Proceedings of the National
Academy of Sciences of the USA 104, 12955–12961.
Landa, B.B., Rius, J.E.P., Vovlas, N., Carneiro, R., Maleita, C.M.N., Abrantes, I.M.D. and Castillo, P. (2008)
Molecular characterization of Meloidogyne hispanica (Nematoda, Meloidogynidae) by phylogenetic
analysis of genes within the rDNA in Meloidogyne spp. Plant Disease 92, 1104–1110.
Longhorn, S.J., Foster, P.G. and Vogler, A.P. (2007) The nematode-arthropod clade revisited: phylogenomic
analyses from ribosomal protein genes misled by shared evolutionary biases. Cladistics 23,
130–144.
Loytynoja, A. and Milinkovitch, M.C. (2003) A hidden Markov model for progressive multiple alignment.
Bioinformatics 19, 1505–1513.
Lunt, D.H. (2008) Genetic tests of ancient asexuality in root knot nematodes reveal recent hybrid origins.
Bmc Evolutionary Biology 8, 194.
Maddison, W.P. (1993) Missing data versus missing characters in phylogenetic analysis. Systematic Biology
42, 576–581.
Maddison, W.P. (1997) Gene trees in species trees. Systematic Biology 46, 523–536.
Maddison, W.P. and Knowles, L.L. (2006) Inferring phylogeny despite incomplete lineage sorting. Systematic
Biology 55, 21–30.
Maddison, W.P. and Maddison, D.R. (2002) MacClade version 4.0.5. Sinauer, Sunderland, Massachusetts.

Perry_Chap-05.indd 136 6/15/2009 3:28:27 PM

 Molecular Taxonomy and Phylogeny 137

Morrison, D.A. and Ellis, J.T. (1997) Effects of nucleotide sequence alignment on phylogeny estimation:
a case study of 18S rDNAs of Apicomplexa. Molecular Biology and Evolution 14, 428–441.
Nguyen, K.B., Maruniak, J. and Adams, B.J. (2001) Diagnostic and phylogenetic utility of the rDNA internal
transcribed spacer sequences of Steinernema. Journal of Nematology 33, 73–82.
Powers, T.O. (2004) Nematode molecular diagnostics: from bands to barcodes. Annual Review of
Phytopathology 42, 367–383.
Powers, T.O. and Harris, T.S. (1993) A polymerase chain reaction method for identification of five major
Meloidogyne species. Journal of Nematology 25, 1–6.
Powers, T.O. and Sandall, L.J. (1988) Estimation of genetic divergence in Meloidogyne mitochondrial DNA.
Journal of Nematology 20, 505–511.
Powers, T.O., Harris, T.S. and Hyman, B.C. (1993) Mitochondrial DNA sequence divergence among
Meloidogyne incognita, Romanomermis culicivorax, Ascaris suum, and Caenorhabditis elegans.
Journal of Nematology 25, 564–572.
Powers, T.O., Todd, T.C., Burnell, A.M., Murray, P.C.B., Fleming, C.C., Szalanski, A.L., Adams, B.A. and
Harris, T.S. (1997) The rDNA internal transcribed spacer region as a taxonomic marker for nema-
todes. Journal of Nematology 29, 441–450.
Powers, T.O., Mullin, P.G., Harris, T.S., Sutton, L.A. and Higgins, R.S. (2005) Incorporating molecular iden-
tification of Meloidogyne spp. into a large-scale regional nematode survey. Journal of Nematology 37,
226–235.
Purvis, A. (1995) A modification to Baum and Ragan’s method for combining phylogenetic trees. Systematic
Biology 44, 251–255.
Roberts, R.G. and Bobrow, M. (1998) Dystrophins in vertebrates and invertebrates. Human Molecular
Genetics 7, 589–595.
Roberts, R.G., Nicholson, L., Bone, Q. and Bobrow, M. (1995) Phylogenetic analysis of vertebrate and
invertebrate dystrophins and utrophins. Journal of Medical Genetics 32, 142–142.
Roberts, T.M. (2005) Major sperm protein. Current Biology 15, R153.
Scholl, E.H. and Bird, D.M. (2005) Resolving tylenchid evolutionary relationships through multiple gene
analysis derived from EST data. Molecular Phylogenetics and Evolution 36, 536–545.
Slowinski, J.B. and Page, R.D.M. (1999) How should species phylogenies be inferred from sequence data?
Systematic Biology 48, 814–825.
Stanton, J., Hugall, A. and Moritz, C. (1997) Nucleotide polymorphisms and an improved PCR-based
mtDNA diagnostic for parthenogenetic root-knot nematodes (Meloidogyne spp.). Fundamental and
Applied Nematology 20, 261–268.
Stein, L.D., Bao, Z., Blasiar, D. et al. (2003) The genome sequence of Caenorhabditis briggsae: a platform
for comparative genomics. Public Library of Science – Biology 1, 166–192.
Subbotin, S.A., Sturhan, D., Vovlas, N., Castillo, P., Tanyi Tambe, L., Moens, M. and Baldwin, J.G. (2006)
Application of secondary structure model of rRNA for phylogeny: D2-D3 expansion segments of the
LSU gene of plant-parasitic nematodes from the family Hoplolaimidae Filipjev, 1934. Molecular
Phylogenetics and Evolution 43, 881–890.
Sudhaus, W. and Kiontke, K. (2007) Comparison of the cryptic nematode species Caenorhabditis brenneri
sp n. and C. remanei (Nematoda : Rhabditidae) with the stem species pattern of the Caenorhabditis
elegans group. Zootaxa, 45–62.
Swofford, D.L. (2002) PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods). Sinauer
Associates, Sunderland, Massachusetts.
Tenente, G.C.M.V., De Ley, P., De Ley, I.T., Karssen, G. and Vanfleteren, J.R. (2004) Sequence analysis of
the D2/D3 region of the large subunit RDNA from different Meloidogyne isolates. Nematropica 34,
1–12.
Terry, M.D. and Whiting, M.F. (2005) Behavior of POY versus CLUSTALL across a broad parameter land-
scape. Cladistics 21, 272–281.
Thompson, J.D., Higgins, D.G. and Gibson, T.J. (1994) Clustal-W – improving the sensitivity of progressive
multiple sequence alignment through sequence weighting, position-specific gap penalties and weight
matrix choice. Nucleic Acids Research 22, 4673–4680.
Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. and Higgins, D.G. (1997) The CLUSTAL X win-
dows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools.
Nucleic Acids Research 25, 4876–4882.
Thorley, J.L. and Page, R.D M. (2000) RadCon: phylogenetic tree comparison and consensus. Bioinformatics
16, 486–487.

Perry_Chap-05.indd 137 6/15/2009 3:28:27 PM

 138 B. Adams et al.

Tigano, M.S., Carneiro, R., Jeyaprakash, A., Dickson, D.W. and Adams, B.J. (2005) Phylogeny of
Meloidogyne spp. based on 18S rDNA and the intergenic region of mitochondrial DNA sequences.
Nematology 7, 851–862.
Triantaphyllou, A.C. (1966) Polyploidy and reproductive patterns in the root-knot nematode Meloidogyne
hapla. Journal of Morphology 118, 403–414.
Triantaphyllou, A. (1985) Cytogenetics, cytotaxonomy and phylogeny of root-knot nematodes. In: Sasser,
J.N. and Carter, C.C. (eds) An Advanced Treatise on Meloidogyne, Volume 1. North Carolina State
University Graphics, North Carolina, Raleigh, North Carolina, pp. 113–126.
Trudgill, D.L. and Blok, V.C. (2001) Apomictic, polyphagous root-knot nematodes: exceptionally successful
and damaging biotrophic root pathogens. Annual Review of Phytopathology, 39, 53–77.
van der Beek, J.G., Folkertsma, R., Zijlstra, C., Van Koert, P.H.G., Poleij, L.M. and Bakker, J. (1998) Genetic
variation among parthenogenetic Meloidogyne species revealed by AFLPs and 2D-protein electro-
phoresis contrasted to morphology. Fundamental and Applied Nematology 21, 401–411.
Watson, J.D. (2008) Molecular Biology of the Gene. Benjamin-Cummings, New Jersey.
Wiens, J.J. (1998) Does adding characters with missing data increase or decrease phylogenetic accuracy?
Systematic Biology 47, 625–640.

Perry_Chap-05.indd 138 6/15/2009 3:28:27 PM