tbea: tools for pre- and post-processing in Bayesian evolutionary analyses

Abstract

Abstract Estimating phylogenies in which branch lengths are expressed in units of absolute time is crucial for testing hypotheses in evolutionary biology. However, bioinformatic tools to pre- and post-process data from Bayesian divergence time estimation analyses are often not easily interoperable, and documenting methodological choices is not a generalized practice. The R package tbea is a tool-set to integrate biological, geological, and paleontological information to optimize the specification of models, their parameters, and prior distributions in divergence time estimation analyses. tbea implements statistical models to (i) better translate time information in dating sources into the specified calibration densities, (ii) improve comparisons between prior and posterior distributions for parameters of interest, (iii) carry out inference on origination times for a set of distributions, (iv) summarize different distributions into a single one, and (v) improve the reproducibility of divergence time estimation analyses allowing users to document methodological choices. We illustrate the functionalities of tbea by carrying out two worked examples, one on the phylogenetic relationships and divergence time estimation of South American Cynodontidae, and the other on the separation time of drainages east and west of the Andes in South America. It is expected that the tools available in tbea will be key when estimating events in time from sets of point estimates, and the combination of different posterior densities from the same parameter will be useful in justifying the selection of secondary calibration points, or discussing the timing of biogeographical events when multiple sources are available.