Genus]_[species]; Presenting phylogenies to facilitate synthesis.

Abstract

Each published phylogeny is a potential contribution to the synthesis of the Tree of Life and countless downstream projects. Steps are needed for fully synthesizable science, but only a minority of studies achieve these. We here review the range of phylogenetic presentation and note aspects that hinder further analysis. We provide simple suggestions on publication that would greatly enhance utilizability, and propose a formal grammar for phylogeny terminal format. We suggest that each published phylogeny should be accompanied by at minimum the single preferred result in machine readable tree (e.g. Newick) form in the supplement, a simple task fulfilled by fewer than half of studies. Further, the tree should be clear from the file name and extension; the orientation (rooted or unrooted) should match the figures; terminals labels should include genus and species IDs; underscores should separate strings within-field (instead of white spaces); and if other informational fields are added these should be separated by a unique delimiting character (we suggest multiple underscores or the vertical pipe character, |) and ordered consistently. These requirements are largely independent of phylogenetic study aims, while we note other requirements for synthesis (e.g. removal of species repeats and uninformative terminals) that are not necessarily the responsibility of authors. Machine readable trees show greater variation in terminal formatting than typical phylogeny images (owing presumably to greater scrutiny of the latter), and thus are complex and laborious to parse. Since the majority of existing studies have provided only images, we additionally review typical variation in plotting style, information that will be necessary for developing the automated phylogeny transcription tools needed for their eventual inclusion in the Tree of Life.