Estimating the phylogeny and divergence times of primates using a supermatrix approach

Abstract

Background: The primates are among the most broadly studied mammalian orders, with the published literature containing extensive analyses of their behavior, physiology, genetics and ecology. The importance of this group in medical and biological research is well appreciated, and explains the numerous molecular phylogenies that have been proposed for most primate families and genera. Composite estimates for the entire order have been infrequently attempted, with the last phylogenetic reconstruction spanning the full range of primate evolutionary relationships having been conducted over a decade ago.

Results: To estimate the structure and tempo of primate evolutionary history, we employed Bayesian phylogenetic methods to analyze data supermatrices comprising 7 mitochondrial genes (6,138 nucleotides) from 219 species across 67 genera and 3 nuclear genes (2,157 nucleotides) from 26 genera. Many taxa were only partially represented, with an average of 3.95 and 5.43 mitochondrial genes per species and per genus, respectively, and 2.23 nuclear genes per genus. Our analyses of mitochondrial DNA place Tarsiiformes as the sister group of Strepsirrhini. Within Haplorrhini, we find support for the primary divergence of Pitheciidae in Platyrrhini, and our results suggest a sister grouping of African and non-African colobines within Colobinae and of Cercopithecini and Papionini within Cercopthecinae. Date estimates for nodes within each family and genus are presented, with estimates for key splits including: Strepsirrhini-Haplorrhini 64 million years ago (MYA), Lemuriformes-Lorisiformes 52 MYA, Platyrrhini-Catarrhini 43 MYA and Cercopithecoidea-Hominoidea 29 MYA.

Conclusion: We present an up-to-date, comprehensive estimate of the structure and tempo of primate evolutionary history. Although considerable gaps remain in our knowledge of the primate phylogeny, increased data sampling, particularly from nuclear loci, will be able to provide further resolution.

Figure 1

Mitochondrial tree of primate genera. Maximum-clade-credibility tree of the Order Primates, inferred from a genus-level mitochondrial DNA supermatrix using the Bayesian phylogenetic software BEAST. Nodes are labelled with a/b, where a represents the Bayesian posterior probability expressed as a percentage and b represents the percentage of 1,000 maximum-likelihood bootstrap replicates that support the node. Asterisks indicate 100% support; nodes with 100% support in both Bayesian and maximum-likelihood frameworks are labelled with single asterisks. The tree is drawn to a timescale, with node heights representing mean posterior estimates.

Figure 2

Mitochondrial tree of strepsirrhine species. Maximum-clade-credibility subtree of Strepsirrhini, inferred from a species-level mitochondrial DNA supermatrix using the Bayesian phylogenetic software BEAST. Nodes are labelled with a/b, where a represents the Bayesian posterior probability expressed as a percentage and b represents the percentage of 1,000 maximum-likelihood bootstrap replicates that support the node. Asterisks indicate 100% support; nodes with 100% support in both Bayesian and maximum-likelihood frameworks are labelled with single asterisks. The tree is drawn to a timescale, with node heights representing mean posterior estimates.

Figure 3

Mitochondrial tree of cercopithecoid species. Maximum-clade-credibility subtree of Cercopithecoidea, inferred from a species-level mitochondrial DNA supermatrix using the Bayesian phylogenetic software BEAST. Nodes are labelled with a/b, where a represents the Bayesian posterior probability expressed as a percentage and b represents the percentage of 1,000 maximum-likelihood bootstrap replicates that support the node. Asterisks indicate 100% support; nodes with 100% support in both Bayesian and maximum-likelihood frameworks are labelled with single asterisks. The tree is drawn to a timescale, with node heights representing mean posterior estimates.

Figure 4

Mitochondrial tree of platyrrhine species. Maximum-clade-credibility subtree of Haplorrhini, inferred from a species-level mitochondrial DNA supermatrix using the Bayesian phylogenetic software BEAST. Nodes are labelled with a/b, where a represents the Bayesian posterior probability expressed as a percentage and b represents the percentage of 1,000 maximum-likelihood bootstrap replicates that support the node. Asterisks indicate 100% support; nodes with 100% support in both Bayesian and maximum-likelihood frameworks are labelled with single asterisks. The tree is drawn to a timescale, with node heights representing mean posterior estimates.

Figure 5

Mitochondrial tree of hominoid species. Maximum-clade-credibility subtree of Hominoidea, inferred from a species-level mitochondrial DNA supermatrix using the Bayesian phylogenetic software BEAST. Nodes are labelled with a/b, where a represents the Bayesian posterior probability expressed as a percentage and b represents the percentage of 1,000 maximum-likelihood bootstrap replicates that support the node. Asterisks indicate 100% support; nodes with 100% support in both Bayesian and maximum-likelihood frameworks are labelled with single asterisks. The tree is drawn to a timescale, with node heights representing mean posterior estimates.

Figure 6

Nuclear tree of haplorrhine genera. Maximum-clade-credibility tree of Haplorrhini, inferred from a genus-level nuclear DNA supermatrix using the Bayesian phylogenetic software BEAST. Nodes are labelled with a/b, where a represents the Bayesian posterior probability expressed as a percentage and b represents the percentage of 1,000 maximum-likelihood bootstrap replicates that support the node. Asterisks indicate 100% support; nodes with 100% support in both Bayesian and maximum-likelihood frameworks are labelled with single asterisks. The tree is drawn to an arbitrary timescale, obtained using a fixed substitution rate of 1.0 substitution/site/time-unit. Node heights represent mean posterior estimates.