From the Research: Myths Worth Dispelling
As DNA sequences have become more readily available, it has become increasingly desirable to infer species phylogenies from multigene data sets. Much recent work has centered around the recognition that substantial incongruence in single-gene phylogenies necessitates the development of statistical procedures to estimate species phylogenies that appropriately model the process of evolution at the level of the individual genes. One process that gives rise to variation in the histories of individual genes is incomplete lineage sorting, which is commonly modeled by the coalescent, and thus much current work is focused on proper estimation of species phylogenies under the coalescent model. A second common source of discord in single-gene phylogenies is hybridization, a process that is ubiquitous in many groups of plants and animals. Although methods to incorporate hybridization into phylogenetic estimation have also been developed, only a handful of methods that address both coalescence and hybridization have been proposed. Here, I propose an extension of an existing model that incorporates both of these processes simultaneously by utilizing gene trees for inference in a likelihood framework. The model allows examination of the evidence for hybridization in the presence of incomplete lineage sorting due to deep coalescence via model selection using standard information criteria (e.g., Akaike information criterion and Bayesian information criterion). The potential of the method is evaluated using simulated data.