A distance-based method for inferring phylogenetic networks in the presence of incomplete lineage sorting

Abstract

Hybridization and incomplete lineage sorting (ILS) are two evolutionary processes that result in incongruence among gene trees and complicate the identification of the species evolutionary history. Although a wide array of methods have been developed for inference of species phylogeny in the presence of each of these two processes individually, methods that can account for both of them simultaneously have been introduced recently. However, these new methods are based on the optimization of certain criteria, such as parsimony and likelihood, and are thus computationally intensive. In this paper, we present a novel distance-based method for inferring phylogenetic networks in the presence of ILS that makes use of pairwise distances computed from multiple sampled loci across the genome. We show in simulation studies that the method infers accurate networks when the estimated pairwise distances have good accuracy. Furthermore, we devised a heuristic for post-processing the inferred network to remove potential false positive reticulation events. Themethod is computationally very efficient and is applicable to very large data sets.