How Many Genes Should a Systematist Sample? Conflicting Insights from a Phylogenomic Matrix Characterized by Replicated Incongruence

The average size of molecular systematic data sets has grown steadily over the past 20 years. Combined phylogenetic matrices that include multiple genetic loci currently are the norm, and in many cases, rapid compilation of extremely large DNA data sets is feasible. Thus, a frequently asked question is “How many genes should a systematist sequence in order to generate a robust phylogenetic hypothesis?” This query generally has been addressed by computer simulation, where the amount of virtual DNA sequence data that can be generated is unlimited (e.g., Huelsenbeck and Hillis, 1993). Genomic data, however, provide systematists with a multitude of empirical molecular data for phylogenetic analysis, and several authors have taken advantage of this resource to examine the effects of increasing the number of genes to quantities that seemed impossible in the recent past (e.g., Cummings et al., 1995; Bapteste et al., 2002; Goremykin, 2004). In one noteworthy study, Rokas et al. (2003) compiled a large systematic matrix of 127,026 nucleotide positions from 106 genes for 7 species of Saccharomyces yeast and an outgroup (Candida albicans). Maximum likelihood (ML) and parsimony analyses of this large data set produced congruent, well-supported results with bootstrap scores of 100% for all clades (Fig. 1a). In spite of this overwhelming support, Rokas et al. (2003) noted that there was widespread topological conflict among gene trees. Separate analyses of individual genes produced various topologies that contradicted all nodes in the tree based on concatenation of 106 genes (Fig. 1a). Pairwise comparisons of gene trees showed extensive incongruence, and one conflicting clade, S. kudriavzevii + S. bayanus, was supported by a very large percentage of the gene trees (Fig. 1a). Replicated support for this anomalous clade was apparent in analyses of nucleotides, transversions, codons, and amino acids for a variety of systematic methods (Rokas et al., 2003; also see Holland et al., 2004, 2006; Phillips et al., 2004; Taylor and Piel, 2004; Collins et al., 2005; Gatesy et al., 2005; Ren et al., 2005; Hedtke et al., 2006). (a) The tree supported by the concatenation of 106 genes from Rokas et al. (2003). ML bootstrap scores are above internodes, and the percentage of ML gene trees that strictly supported a particular clade are indicated below internodes. The most common, conflicting clade, Skud+Sbay, also is shown. (b) Branch lengths for the optimal ML tree for the concatenation of 106 yeast genes; scale bar shows expected numbers of substitutions per site (length of outgroup branch is indicated). All phylogenetic analyses in this paper were branch and bound searches executed in PAUP* 4.0b10 (Swofford, 2002). All ML models were chosen by likelihood ratio tests as in Rokas et al. (2003) using PAUP* and ModelTest 3.06 (Posada and Crandall, 1998). Bootstrap analyses (Felsenstein, 1985) were as in Gatesy and Baker (2005). Scer = Saccharomyces cerevisiae; Spar = S. paradoxus; Smik = S. mikatae; Skud = S. kudriavzevii; Sbay = S. bayanus; Scas = S. castellii; Sklu = S. kluyveri; and Calb = Candida albicans. By examining correlations between bootstrap scores and possible confounding factors, however, Rokas et al. (2003) concluded that “… none of the factors known or predicted to cause phylogenetic error could systematically account for the observed incongruence, suggesting that there may be no good predictor of the phylogenetic informativeness of genes” (p. 802). Therefore, many randomly selected genes were necessary to overwhelm conflicting signals. In this case study, very large concatenated data sets of ∼ 20 genes were required to provide 95% bootstrap support for all nodes in the combined data tree, “substantially more genes than commonly used but a small fraction of any genome” (p. 799). Rokas et al. (2003) concluded that “These results have important implications for resolving branches of the tree of life” (p. 799) and “… important implications for many current practices in molecular phylogenetics” (p. 802), points that were reasserted in a commentary by Gee (2003). Specifically, if 20 or more genes generally are required to yield robust support, then most previous phylogenetic analyses are inadequate in terms of character sampling. This assertion is based on the assumption that the 8 taxa analyzed by Rokas et al. (2003) represent a typical systematic problem. Rokas et al. (2003) considered this issue, noting that “It is possible that the 8 yeast taxa we have analyzed represent a very difficult phylogenetic case, atypical of the situations found in other groups. However, the widespread occurrence of incongruence at all taxonomic levels argues strongly against such a view. Rather, we believe that this group is a representative model for key issues that researchers in phylogenetics are confronting” (p. 802). Large matrices that combine information from 20 or more gene fragments are rare (e.g., Murphy et al., 2001; Bapteste et al., 2002; Gatesy et al., 2002; Goremykin, 2004); therefore, if the test case of Rokas et al. (2003) is representative, most published molecular systematic studies are, at best, preliminary efforts. Rokas et al. (2003) primarily used the nonparametric bootstrap (Felsenstein, 1985) to assess support and to search for correlates of incongruence in the yeast matrix. Recent reanalyses have utilized a variety of techniques to further characterize conflicting signals in the yeast data set. These approaches included Bayesian analysis (Taylor and Piel, 2004; Jeffroy et al., 2006), transversion coding (Phillips et al., 2004; Jeffroy et al., 2006), removal of rapidly evolving third codon positions (Collins et al., 2005; Jeffroy et al., 2006), partitioned Bremer support scores (Collins et al., 2005; Gatesy et al., 2005), consensus networks (Holland et al., 2004, 2006), isolation of genes with shifting base compositional biases (Collins et al., 2005), supertree bootstrapping (Burleigh et al., 2006), increased taxon sampling (Rokas and Carroll, 2005; Hedtke et al., 2006), and better fitting models of molecular evolution (Ren et al., 2005). Alternatively, several authors have suggested that reducing the number of taxa included in analysis can yield insights regarding the stability of phylogenetic hypotheses (Lanyon, 1985; Philippe and Douzery, 1994; Siddall, 1995; Brochu, 1997; Poe, 1998; Siddall and Whiting, 1999; Holland et al., 2003). Here, we use selected removal of taxa to explore patterns of incongruence in the yeast data set. In particular, we analyze different subsets of species to determine whether disagreements among gene trees are tempered or accentuated by altering taxonomic representation. In combination with documentation of branch lengths for individual gene trees, our subsampling results show that the set of species analyzed by Rokas et al. (2003) is not representative of most published systematic studies. We suggest that the yeast matrix does not provide a coherent, general recommendation for how many genes to sample in future molecular systematic studies. However, patterns of conflict for different subsets of species offer a very simple explanation for replication of the discrepant S. kudriavzevii + S. bayanus clade in many gene trees (Fig. 1a). Examination of the optimal topology for the concatenation of 106 genes showed a striking difference between branches that connected 5 closely related Saccharomyces species (S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, S. bayanus) and branches that led to S. castellii, S. kluyveri, and the outgroup C. albicans (Fig. 1b; Hedtke et al., 2006; Jeffroy et al., 2006). For the ML model utilized by Rokas et al. (2003), the branches that connected to S. castellii, S. kluyveri, and C. albicans ranged from 0.31 to 1.58 expected substitutions per site, whereas the branches that joined the remaining, closely related Saccharomyces species were from 0.03 to 0.08 substitutions per site. Only 15% of the inferred nucleotide substitutions occurred on branches that linked these 5 species (Fig. 1b). Consistent with an estimated Precambrian (∼ 723 Mya) divergence of Candida albicans from Saccharomyces cerevisiae (Hedges et al., 2004), the outgroup branch in the yeast tree was exceptionally long. Each site in the concatenated data set is expected to change 1.58 times on this branch according to the ML estimate (Fig. 1b). For the topology supported by the concatenation of 106 genes, 43% of the yeast genes had one branch that was > 2.00 expected substitutions per site, and 79% of the yeast genes had at least one branch that was > 1.00 substitution per site (also see Hedtke et al., 2006). For comparison, in an often cited discussion of long branches in a 28S rDNA tree of holometabolous insects, Huelsenbeck (1998) remarked that two branches in his analysis were “among the longest ever observed (approximately 1.0 substitution per site)” (p. 530). However, branches in many of the yeast gene trees dwarfed those in the insect rDNA tree and were up to 95 times longer (Fig. 2). From another perspective, the longest branches in the yeast data set exceeded those in a tree based on mitochondrial genomes from 5 animal phyla (Naylor and Brown, 1998) and also were much longer than branches in simulations designed to assess misplacement of long branches (e.g., Anderson and Swofford, 2004). Although it has been suggested that the set of species in the yeast data set represents a typical phylogenetic problem (Rokas et al., 2003), the extraordinarily long branch lengths in most yeast gene trees demonstrate that this is not the case (e.g., Fig. 2). The 8 yeast genes with the longest ML branch lengths for the tree supported by the concatenation of 106 genes. The scale bar represents 1.00 expected substitution per site; branches that connect the 5 most closely related Saccharomyces species are tiny at this scale. The length of the longest branch in each yeast gene tree is indicated. Note that some topologies show more than one branch that is > 1.00 expected substitution per site. Abbreviations for yeast species are as in Figure 1. For the yeast data set, gene trees that included all 8 species showed many conflicts among the 5 most closely related Saccharomyces species (Fig. 1a). In ML analyses, 11% of gene trees conflicted with the S. cerevisiae + S. paradoxus clade, 28% conflicted with the S. cerevisiae + S. paradoxus + S. mikatae clade, and 45% conflicted with the S. cerevisiae + S. paradoxus + S. mikatae + S. kudriavzevii clade. Given the moderate lengths of branches that linked S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, and S. bayanus (Fig. 1b), we were surprised by the widespread discrepancies among genes at this level. To further explore differences among gene trees, we reanalyzed the 5 closely related species of Saccharomyces in isolation from their distant relatives, S. castellii, S. kluyveri, and C. albicans. Given the diversity of gene trees for all 8 taxa, we expected to find many conflicting topologies but were shocked by complete congruence among the 106 gene trees in ML analyses (Fig. 3). There are 15 possible bifurcating trees (unrooted) for a data set of 5 taxa; assuming an equal probability for each topology a priori, the chance of recovering the same tree 106 straight times is astronomically low (P = 3.24 × 10− 124). Ironically, a systematic data set that has been presented as a prime example of pervasive, inexplicable conflict among genes (Rokas et al., 2003) can be transformed, with the removal of 3 species, into a remarkably congruent data set that shows 100% agreement among 106 genes. For the set of 5 closely related Saccharomyces species, 20 genes were not necessary to resolve relationships; basically any gene will do (Fig. 3c). Subsets of only 600 randomly resampled nucleotides from the yeast data set were sufficient to support the 2 nodes in the 5 taxon tree in > 95% of and only nucleotides each clade > of the (Fig. Fig. These are very small of to most systematic studies. congruence among 106 gene trees for the 5 closely related Saccharomyces In ML analyses, all genes supported the same topology (a) that included and randomly nucleotides were required to these clades in analyses of the 5 closely related Saccharomyces species of only 600 nucleotides supported each clade > 95% of the = The 106 congruent ML gene trees for and Sbay are in ML bootstrap scores and the number of nucleotides for each gene are indicated. are as in Figure 1. For all genes, ML models for 5 species were chosen as in Rokas et al. (2003) using PAUP* and ModelTest of randomly from the yeast data set (b) were in PAUP* and included for each sample size and nucleotide ML searches were branch and In ML analyses of S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, and S. bayanus, there were no topological discrepancies among gene trees (Fig. but more related taxa (S. castellii, S. kluyveri, and C. were gene trees showed extensive conflicts regarding among the 5 closely related Saccharomyces we noted such conflicts for the of 8 species (Fig. 1a). of the 7 Saccharomyces species, the outgroup C. also widespread incongruence among genes (Rokas et al., 2003), as analyses of species (Fig. supported by the concatenation of 106 genes in ML analyses of Note the very long outgroup expected substitutions per site for the outgroup branches are indicated. In the topology with the conflicting clade was ML bootstrap scores are above internodes, and the percentage of ML gene trees that strictly supported a particular clade are indicated below internodes. For the and trees, numbers in the percentage of times that the clade was for the tree, the number in is the percentage of gene trees that supported the clade. are as in Figure 1. For each gene and for the concatenation of 106 genes, ML models for each set of species were chosen as in Rokas et al. (2003) using PAUP* and ModelTest For the species set of 5 closely related Saccharomyces species and C. the in length between the outgroup branch and branches was of the expected character change was to the outgroup branch substitutions per and the concatenation of 106 genes supported a of S. kudriavzevii + S. bayanus with an ML bootstrap of (Fig. This was with the S. cerevisiae + S. paradoxus + S. mikatae + S. kudriavzevii clade that had 100% bootstrap support in the analysis of 106 genes for 8 species (Fig. 1a). Thus, a systematic data set of 8 species, in 20 genes were considered sufficient for robust phylogenetic support (Rokas et al., 2003), can be transformed, with the of 2 species, into a data set of 106 genes that a > concatenated genes not provide bootstrap support at all nodes for this set of species (Fig. analyses of taxon (Fig. Fig. show that the number of genes required to yield bootstrap scores is on the of a systematic problem and suggest that long branches (Fig. much of the incongruence among genes in the yeast data set (Taylor and Piel, 2004; Hedtke et al., 2006; Jeffroy et al., 2006). In ML analyses of the 5 closely related Saccharomyces species, all gene trees had the same 7 branches (Fig. of the to 5 of these 7 branches will yield the S. kudriavzevii + S. bayanus clade (Fig. the related C. albicans was to a matrix that included S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, and S. bayanus, all ML gene trees were with the topology for these 5 Saccharomyces species (Fig. but was among the 7 branches (Fig. The most was on the branch (S. bayanus; For the other genes, the was the of the and the of gene trees of supported the S. kudriavzevii + S. bayanus clade and an equal a probability of recovering the S. kudriavzevii + S. bayanus clade or the S. cerevisiae + S. paradoxus + S. mikatae + S. kudriavzevii clade, it be for one of these to be supported in of gene trees, as was the case probability of but patterns were observed in ML gene trees for other of taxa, in the 5 closely related Saccharomyces species were with S. or S. S. kudriavzevii + S. bayanus was the most common, conflicting clade (Fig. Rokas et al. (2003) the same of support for the conflicting S. kudriavzevii + S. bayanus in gene trees for the 7 Saccharomyces species, the outgroup C. albicans. (a) ML tree supported by concatenation of 106 genes for the 5 closely related Saccharomyces species, with possible shown. of the 7 yield the conflicting clade. (b) of the in 106 ML gene trees for species and of the in 106 ML gene trees for all 8 yeast on branches the number of times a particular branch was by Calb (b) or by and Sklu For 3 genes, there were 2 optimal on branches in (b) that in of the 2 trees, the was to that are as in Figure 1. the possible bifurcating topologies for all 8 species, the S. kudriavzevii + S. bayanus is found in only of all trees. However, ML analyses of 8 species showed that S. kudriavzevii + S. bayanus was times in 106 gene trees (Fig. previous and Bayesian results showed this same of incongruence whether nucleotides, transversions, codons, or amino acids were analyzed (Rokas et al., 2003; Phillips et al., 2004; Taylor and Piel, 2004; Collins et al., 2005; Gatesy et al., 2005; Ren et al., 2005; et al., 2006; Holland et al., 2004, 2006). of the S. kudriavzevii + S. bayanus clade in ∼ of our ML gene trees strongly suggested an all ML gene trees for 8 species were with in the tree for the 5 closely related Saccharomyces species (Fig. but different of the 3 long branch taxa (Fig. Fig. in many gene trees that supported the S. kudriavzevii + S. bayanus clade (Fig. and Fig. in the analyses of taxa (Fig. support for the conflicting S. kudriavzevii + S. bayanus was to of the topology for S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, and S. bayanus (Fig. by the very related S. castellii, S. kluyveri, and C. albicans (Fig. to Fig. for discussion of distant see Huelsenbeck et al., 2002; Holland et al., 2003; Anderson and Swofford, 2004; 2005; et al., 2005; and 2005; et al., 2006). ML topologies for yeast genes. All gene trees are with the topology for the 5 closely related Saccharomyces species (Fig. branches that connect these 5 species are as in according to Figure The genes in (a) all supported the clade support for is in these genes were was not and the topology by the concatenation of 106 genes was optimal Gatesy and 2005). The gene trees in (b) support of these genes not support of the 5 closely related Saccharomyces species but were with the tree for these species (Fig. Rokas et al. (2003) suggested that their analyses of 106 genes from 8 species had important implications for resolving the tree of In particular, that 20 or more genes be required to robust support for phylogenetic This assertion is based on two There are no good for the of different genes, and the 8 species in their data set represent a typical phylogenetic problem. Recent reanalyses of the yeast data set have of these Phillips et al. noted that differences in signals in the yeast matrix. Collins et al. found that in base were most at third codon positions (also see 2006). Collins et al. resampled genes with base only genes were required to bootstrap for supported by the concatenated data set. By genes by large in base were necessary to yield the same of support (Collins et al., 2005). In a on Bayesian support Taylor and found the branch length were for trees that were with the tree Fig. Fig. (p. a that was In the that there are no good of phylogenetic for particular genes does not to for this data set. Taylor and Ren et al. Hedtke et al. Jeffroy et al. and Holland et al. noted the of exceptionally long branches and that a of divergence and branch length (e.g., Fig. were of conflict among genes in the yeast data set. Here, we these and concluded that the 8 species in the yeast data set do not represent a phylogenetic problem. The tree based on the concatenated matrix of 106 genes showed in branch lengths (Fig. 1b), but individual gene trees had some branches that were up to expected substitutions per site (Fig. 2). This of nucleotide substitution does not represent a typical phylogenetic many systematists that this of divergence is a very difficult problem (Felsenstein, and 1998; and Siddall, 2001; Holland et al., 2003; Anderson and Swofford, 2004; 2005; et al., 2005). S. castellii, S. kluyveri, and C. albicans are very distant from each other and from the 5 most closely related Saccharomyces species (Fig. Fig. Fig. and Fig. Therefore, it was not that there were conflicts among gene trees in and ML analyses (Taylor and Piel, 2004; Hedtke et al., 2006). In the 3 taxa, also were by the in base (Collins et al., 2005), for all conflicts among genes in ML of extensive incongruence, Rokas et al. (2003) found that 20 randomly genes from the yeast matrix were required for a supported tree of 8 species, but this has no this 106 gene it is that some systematic are much more difficult to to For the 5 closely related Saccharomyces species, one gene be sufficient (Fig. 3c). ML analyses of individual genes produced the same tree 106 straight and sets of 600 randomly nucleotides supported this topology (Fig. By in ML analyses of these 5 species C. 106 concatenated genes were the optimal ML tree (Fig. contradicted the tree for all 8 yeast species, a topology that was to show an of support (Fig. Fig. 1b; 2003; Rokas et al., 2003). the of genes that is required to resolve will be on the of the phylogenetic problem at and 2005; Hedtke et al., 2006), as as a particular of (e.g., et 2001; Siddall, 2002; and 2003; et al., 2004; Taylor and Piel, 2004; Jeffroy et al., 2006). For phylogenetic where divergence among taxa is not and are a gene provide bootstrap support (e.g., Fig. 3). However, in this 2 or more genes may be that linked nucleotides do not provide for phylogenetic For where exceptionally long branches are 106 genes not be (e.g., Fig. with branch lengths increased taxonomic sampling (e.g., and Hillis, from the (e.g., Brochu, or a set of more evolving genes (e.g., et al., with base (e.g., Collins et al., may be can be but in the the amount of character data to at a well-supported phylogenetic can only be by data to data and then the reanalyses a very simple explanation for support of the conflicting S. kudriavzevii + S. bayanus clade in many yeast gene trees (Fig. 1a). of a topology for 5 (Fig. by 3 distant taxa (Fig. can account for this the topology for S. cerevisiae, S. paradoxus, S. mikatae, S. kudriavzevii, and S. bayanus was (Fig. of the (Fig. the discrepant S. kudriavzevii + S. bayanus clade. In the most case that we data set that included of 8 species in the yeast the was over the in the of ML gene trees of 106 = and in the concatenated analysis of 106 genes (Fig. This represents a for in > genes from taxa may be and where congruence among individual gene trees has been used to assess support (e.g., Rokas et al., 2003; Holland et al., 2004, 2006; et al., 2006). many authors have that large of genes can provide but bootstrap support of model inadequate taxon or (e.g., Philippe and Douzery, 1994; and Brown, 1998; Holland et al., 2004, 2006; Phillips et al., 2004; et al., 2004; et al., 2004; Hedtke et al., 2006; Jeffroy et al., 2006). Here, we an of conflict in a consensus from analyses of > genes to the as many gene trees the of S. kudriavzevii + S. bayanus over the S. cerevisiae + S. paradoxus + S. mikatae + S. kudriavzevii clade (Fig. In to concatenation of genes, it be expected that partitioned phylogenetic analyses of individual genes should be to supported but this is not the case (Fig. and a simple compilation of many genes for very taxa (Rokas et al., 2003; Rokas and Carroll, be as a general for 2003). We C. and two for on different of the Gatesy was supported by and was supported by the and in at the of and by was supported by the of animal mitochondrial Rokas published multiple sequence and that the