Bayesian Phylogenetic Analysis For Diploid And Allotetraploid Species Networks

Abstract

Allopolyploid species are formed by genome doubling after hybridization between otherwise intersterile parental species. Allopolyploidy is a common speciation mechanism in land plants. Here we describe and evaluate a Bayesian approach to the phylogenetic analysis of species relationships when both ordinary speciation and allopolyploidy are present. The approach takes incomplete lineage sorting into account using the multi-species coalescent model, and extends this to deal with the extra complications due to allopolyploidy. The number of hybridizations is not assumed, which means that the number of parameters varies and a reversible-jump MCMC algorithm is needed to sample from the posterior. The main restriction is that only diploids and allotetraploids are considered. The model is implemented in the BEAST framework and is an extension of Jones et al. (2013). Simulations show that the topology of the network can be reliably inferred along with estimates of other parameters.