Pinniped Diphyly and Bat Triphyly: More Homology Errors Drive Conflicts in the Mammalian Tree

Abstract

Homology is perhaps the most central concept of phylogenetic biology. At difficult to resolve polytomies that are deep in the Tree of Life, a few homology errors in phylogenomic data can drive spurious phylogenetic results.Feijoo and Parada (2017)assembled three phylogenomic data sets for mammals and reported methodological discrepancies and unexpected results that contradict the monophyly of well-established clades in Pinnipedia and Yangochiroptera. Examination ofFeijoo and Parada's (2017)data sets reveals extensive homology errors (paralogous sequences, alignments of different exons to each other) and cross-contamination of sequences from different species. These problems predictably result in distorted estimates of gene trees, species trees, bootstrap support, and branch lengths. Correction of these errors resulted in robust support for conventional relationships in Pinnipedia and Yangochiroptera. Phylogenomic data sets are not immune to the problems of homology errors in sequence alignments. Rather, sequence alignments underlie all inferences in molecular phylogenetics and evolution and should be spot-checked for obvious errors via manual inspection of alignments and gene trees.