We chatted about this Marquez, R. et al. on leapfrog distribution of poison-dart frogs associated with divergence and gene flow. The Zoom discussion was enthusiastic and insightful.
This paper attempted to understand the leapfrog distribution of a species complex of Phyllobates poison-dart frogs. Phenotypically similar (bright yellow) P. bicolor populations demonstrates spatial differentiation and were separated by stripy P. aurotaenia populations.
It turns out that each P. bicolor was more closely-related to their (geographically) nearby P. aurotaenia populations than with the P. bicolor population far away.
With EEMS and Spacemix, the authors found evidence of gene flow between P. bicolor and nearby P. aurotaenia, although the estimates of admixture proportion were very low. Therefore, the authors concluded that the relatedness of P. bicolor and their nearby P. aurotaenia reflected in the mt and nuclear DNA phylogenies was caused by gene flow between these parapatric diverged lineages.
Conventional ancestral state reconstruction would conclude that the yellow coloration arose independently in bicolor northern and southern populations. Alternatively, if we consider gene flow in the divergence context, the northern versus southern bicolor had a common origin, but geographical isolation differentiated the two, and recent introgression between bicolor and aurotaenia inserted aurotaenia between northern and sourthern bicolor. This entails that majority of the genomic regions would be homogenized (between bicolor and nearby aurotaenia) by gene flow so that the best supported species trees reflect the introgression history as opposed to the divergence history. However, the low estimated admixture among populations casted doubt on such reasoning.
This paper raised an important awareness that we should integrate gene flow in phylogenetic inference. We found this paper very interesting and valuable. However, additional lines of evidence of gene flow should be presented to make a compelling argument, especially when the estimates of admixture proportions were very low. We liked the EEMS analysis which agrees with the geography, however it was still a bit circular in this context—more gene flow between genetically similar populations that would cluster together on the phylogeny. ABBA-BABA like tests should be used to parse out introgression versus the alternatives.
Another scenario could generate similar pattern is that the genetic regions underlying bright aposematic yellow adaptively introgressed amongst the biocolor popultions, leaving most of the differentiated genome behind. So the genome-wide phylogenetic relationship would not reveal the introgression (restricted to small fraction of the genome). This scenario would be more likely if the genetic architecture underlying yellow coloration is simple (few genetic regions of large phenotypic effect). We are excited that the authors are doing follow-up work on genetic basis of frog coloration that would shed light on this discussion (but please try to kill less frogs).
Reference:
Marquez, R. et al. BioRxiv. Divergence, gene flow and the origin of leapfrog geographic distributions: The history of color pattern variation in Phyllobates poison-dart frogs
Comments