Research Article

Large number of taxa used to estimate a rooted species tree with the ABC method from an unrooted gene tree

Published: January 31, 2020
Genet. Mol. Res. 19(1): GMR18433 DOI: 10.4238/gmr18433


Various approaches use gene trees to infer species trees produced from incomplete lineage sorting. Generally, one of these approaches is used to deduce the rooted species tree from a rooted gene tree, or another method can be used to determine the unrooted species tree from an unrooted gene tree. Typically, this unrooted species is then rooted through at least one outgroup. However, in theory, the unrooted gene tree can be used consistently and directly to infer the rooted species tree without using an outgroup. We used an unrooted gene tree with the assumption of a multispecies coalescent model to infer the rooted species tree by using the approximate Bayesian computation (ABC). In certain cases, this could be useful, especially when it is hard to locate a fitting outgroup neglected by gene trees. To address the challenges of increasing the taxa number, an ABC was used to gauge the rooted species tree of a large number of taxa, using an unrooted gene tree to develop the rooted species tree. This is the first ABC application that can handle very large numbers of taxa. Based on the results, the Robinson-Foulds (RF) distance is generally equal to 2 when the species tree is in imbalance. When the shape of the species tree is balanced, the RF distance is normally equal to 0. Out of all shapes of species trees, the most recent one is the most appropriate for every clade.

About the Authors