Bayesian Weighted Assembly of Phylogenetic Trees Under the Multispecies Coalescent

• Activity Number: 180 - Statistical Methods in Phylogenetics and Metagenomics
• Type: Contributed
• Date/Time: Tuesday, August 4, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Andrew Richards* and Laura Kubatko
• Companies: Ohio State University and Ohio State University
• Keywords: Phylogenetics; Coalescent; Bayesian; Supertrees

Abstract

The inference of the evolutionary relationships between species is a central question in biology. This inference is difficult to perform as the abundance of data and high dimensionality of the problem leads to long computational times. We have developed a new fast and accurate algorithm for inference in the presence of incomplete lineage sorting (ILS), commonly modeled by the multispecies coalescent. This paper builds upon Chifman and Kubatko (2015), who derived the likelihood of a four-taxon species tree under the coalescent process. We use Bayes’s Rule to derive posterior probabilities of rooted three-taxon species trees, and use them in the Triplet MaxCut algorithm of Sevillya et al. (2016) to estimate the final n-taxon tree. We show results on simulated data and multiple real data sets. We also demonstrate that directly estimating the rooted four-taxon species tree using an uninformative prior can lead to a maximum posterior tree that differs systematically from the true tree under simulated conditions.