Gene Tree Parsimony in the Presence of Gene Duplication, Loss, and Incomplete Lineage Sorting

Inferring species trees from multi-locus data needs to account for gene tree discordance due to various biological processes, including incomplete lineage sorting (ILS) and gene duplication and loss (GDL). Gene tree parsimony (GTP) is a popular approach for estimating species trees that seeks to minimize the number of evolutionary events required to reconcile the species tree with gene trees. Minimizing gene duplication and loss (MGD and MGDL) are GTP approaches that typically make the simplifying assumption that population-related effects such as incomplete lineage sorting (ILS) are negligible. However, this assumption is problematic for denser phylogenies, where ILS is more prominent. Here, we extend the existing GTP methods to account for both GDL and ILS by minimizing a weighted sum of the GDL and “deep” coalescence events required for a given collection of gene trees. We provide a graph-theoretic characterization and present a dynamic programming algorithm for this problem. Through an extensive evaluation study on a collection of simulated and empirical datasets, we compared our proposed GTP approaches with the leading methods in the field.