A generalized framework to expand incomplete phylogenies using non‐molecular phylogenetic information

Abstract

AbstractAimThe increasing availability of molecular information has lifted our understanding of species evolutionary relationships to unprecedent levels. However, current estimates of the world's biodiversity suggest that about a fifth of all extant species are yet to be described, and we still lack molecular information for many of the known species. Hence, evolutionary biologists will have to tackle phylogenetic uncertainty for a long time to come. This prospect has urged the development of software to expand phylogenies based on non‐molecular phylogenetic information, and while the available tools provide some valuable features, major drawbacks persist and some of the proposed solutions are hardly generalizable to any group of organisms.InnovationHere, we present a completely generalized and flexible framework to expand incomplete phylogenies. The framework is implemented in the R package “randtip”, a toolkit of functions that was designed to randomly bind phylogenetically uncertain taxa in backbone phylogenies through a fully customizable and automatic procedure that uses taxonomic ranks as a major source of phylogenetic information. Although randtip can generate fully operative phylogenies for any group of organisms using just a list of species and a backbone tree, we stress that the “blind” expansion of phylogenies using “quick‐and‐dirty” approaches often leads to suboptimal solutions. Thus, we discuss a variety of circumstances that may require customizing simulation parameters beyond default settings to optimally expand the trees, including a detailed step‐by‐step tutorial that was designed to provide guidelines to non‐specialist users.Main ConclusionsPhylogenetic uncertainty should be tackled with caution, assessing potential pitfalls and opportunities to optimize parameter space prior to launch any simulation. Used judiciously, our framework will help evolutionary biologists to efficiently expand incomplete phylogenies and thereby account for phylogenetic uncertainty in quantitative analyses.