Get reliable science

The demand for knowledge about the tree of life is steadily rising in ecology and other fields, but bioinformatic resources designed to meet these needs remain poorly developed. Ecologists pursuing phylogenetic insights into the organization of communities have come to rely on relatively conservative reference trees that, in general, are poorly resolved and documented. New methods for inferring very large trees by mining data from DNA sequence repositories will undoubtedly be useful in community phylogenetics but are not without limitations. Here we argue that the collective phylogenetic knowledge embodied in the literature of systematics is a valuable resource that can be tapped in assembling synthetic trees. Assembling a composite “literature-based” tree by the judicious grafting of clades is one way to achieve a synthesis of current knowledge, and could, under some circumstances, better represent “what we know” about phylogenetic relationships than results obtained from automated pipelines. We describe an approach and new software for storing and annotating trees from published studies and grafting clades together in a documented and repeatable manner. Using this agglomerative approach, we are in the process of assembling a literature-based tree for land plants, which presently contains 14?423 species from over 259 sources. For this strategy to be maximally effective, improvements in digital infrastructure are needed to capture and deploy advances in phylogenetic knowledge as they are published.