Interspecific Gene Flow and Mitochondrial Genome Capture during the Radiation of Jamaican Anolis Lizards (Squamata; Iguanidae)

Abstract Gene flow and reticulation are increasingly recognized as important processes in the diversification of many taxonomic groups. With the increasing ease of collecting genomic data and the development of multispecies coalescent network approaches, such reticulations can be accounted for when inferring phylogeny and diversification. Caribbean Anolis lizards are a classic example of an adaptive radiation in which species have independently radiated on the islands of the Greater Antilles into the same ecomorph classes. Within the Jamaican radiation at least one species, Anolis opalinus, has been documented to be polyphyletic in its mitochondrial DNA, which could be the result of an ancient reticulation event or incomplete lineage sorting (ILS). Here, we generate mtDNA and genotyping-by-sequencing (GBS) data and implement gene tree, species tree, and multispecies coalescent network methods to infer the diversification of this group. Our mtDNA gene tree recovers the same relationships previously inferred for this group, which is strikingly different from the species tree inferred from our GBS data. Posterior predictive simulations suggest that our genomic data violate commonly adopted assumptions of the multispecies coalescent model (MSCM), so we use network approaches to infer phylogenetic relationships. The inferred network topology contains a reticulation event but does not explain the mtDNA polyphyly observed in this group; however, coalescent simulations suggest that the observed mtDNA topology is likely the result of past introgression. How common a signature of gene flow and reticulation is across the radiation of Anolis is unknown; however, the reticulation events that we demonstrate here may have allowed for adaptive evolution, as has been suggested in other, more recent, adaptive radiations. [Adaptive radiation; hybridization; introgression; multispecies network coalescent; posterior predictive simulation.]