P2C2M.GMYC: An R package for assessing the utility of the Generalized Mixed Yule Coalescent model

Abstract

In clades that contain large numbers of undescribed species, DNA barcoding methods can serve as a first pass at identifying species limits. The General Mixed Yule Coalescent (GMYC) approach to species delimitation is accurate under certain conditions that are difficult to verify in the very clades for which the GMYC holds the greatest appeal. To circumvent this challenge, we have developed an R package for assessing how well the statistical model implemented in the GMYC fits empirical data. Our approach uses either a parametric bootstrap or a posterior predictive simulation to evaluate model fit. Computational requirements are modest for our package, and most analyses can be completed within minutes to an hour on a typical laptop, depending on whether a user selects a maximum likelihood or Bayesian framework. Results of simulation testing indicate that our approach is effective for assessing the utility of the GMYC model and suggest that it should be included in the analytical pipeline whenever researchers apply the GMYC to clades with unknown species boundaries.