Unguided species delimitation using DNA sequence data from multiple Loci.

Z Yang,B Rannala

doi:10.1093/molbev/msu279

Abstract

A method was developed for simultaneous Bayesian inference of species delimitation and species phylogeny using the multispecies coalescent model. The method eliminates the need for a user-specified guide tree in species delimitation and incorporates phylogenetic uncertainty in a Bayesian framework. The nearest-neighbor interchange algorithm was adapted to propose changes to the species tree, with the gene trees for multiple loci altered in the proposal to avoid conflicts with the newly proposed species tree. We also modify our previous scheme for specifying priors for species delimitation models to construct joint priors for models of species delimitation and species phylogeny. As in our earlier method, the modified algorithm integrates over gene trees, taking account of the uncertainty of gene tree topology and branch lengths given the sequence data. We conducted a simulation study to examine the statistical properties of the method using six populations (two sequences each) and a true number of three species, with values of divergence times and ancestral population sizes that are realistic for recently diverged species. The results suggest that the method tends to be conservative with high posterior probabilities being a confident indicator of species status. Simulation results also indicate that the power of the method to delimit species increases with an increase of the divergence times in the species tree, and with an increased number of gene loci. Reanalyses of two data sets of cavefish and coast horned lizards suggest considerable phylogenetic uncertainty even though the data are informative about species delimitation. We discuss the impact of the prior on models of species delimitation and species phylogeny and of the prior on population size parameters (θ) on Bayesian species delimitation.

Highlights

Genetic sequence data have gained importance in delimiting species in recent years and several inference methods have been proposed for this purpose
Species concepts are often linked to particular mechanisms of achieving and maintaining genetic isolation between incipient species but a useful delimitation method should not be wedded to any particular mechanism of isolation; instead it should be based on detecting the ultimate outcome of speciation—genetic isolation on an evolutionary timescale
From the posterior distribution of the models, we can calculate the posterior distribution of various summaries, such as the posterior probability of each species delimitation, the posterior probability of each delimited species, and the posterior probability for the number of delimited species

Summary

Introduction

Genetic sequence data have gained importance in delimiting species in recent years and several inference methods have been proposed for this purpose (reviewed in Fujita et al 2012; Yang 2014). A general approach to investigating the evolutionary and genetic structure of a group of related organisms using multilocus genetic sequence data can be envisioned as follows: 1) Assign individuals to populations whose members currently interbreed, 2) determine whether the populations are genetically isolated on an evolutionary timescale and are putative species, and 3) determine the phylogenetic history relating the delimited species (Wiens 2007; Yang and Rannala 2010). Those three steps (population assignment, species delimitation, and phylogenetic inference) are usually carried out as separate procedures. A joint analysis should be optimal in maximizing the power and reliability of the inferences

Results

Discussion

Conclusion