Fishing for significance in phylogenies: too many alternatives for the same outcome, or an appeal to Journal editors

Abstract

The ever increasing number of computer programs developed for phylogenetic research does not necessarily facilitate the construction of biologically relevant phylogenies. Regardless of the algorithm utilized by new software, the vast majority result in treelike graphs. We suggest that a new, more inclusive framework for phylogenetic studies needs to be developed, which includes trees as an alternative in the absence of conflicting signals in the sequence data set. Conflicts are caused by noisy phylogenetic signal deriving from hybridization, allopolyploidy and lateral gene transfer—biological processes that undermine the construction of simple dichotomic bifurcating graphs. A robust framework for determining biologically relevant phylogenetic relationships should include quality analysis of the phylogenetic signal, a thorough determination of homology, analyses for phylogenetic networks, and exploration of the data for character or tree conflicts.