Genomics of parallel adaptation at two timescales in Drosophila

Li Zhao,David J Begun,Luciano Matzkin

doi:10.1371/journal.pgen.1007016

Li Zhao, David J Begun + Show 1 more

Open Access

https://doi.org/10.1371/journal.pgen.1007016

Copy DOI

Abstract

Two interesting unanswered questions are the extent to which both the broad patterns and genetic details of adaptive divergence are repeatable across species, and the timescales over which parallel adaptation may be observed. Drosophila melanogaster is a key model system for population and evolutionary genomics. Findings from genetics and genomics suggest that recent adaptation to latitudinal environmental variation (on the timescale of hundreds or thousands of years) associated with Out-of-Africa colonization plays an important role in maintaining biological variation in the species. Additionally, studies of interspecific differences between D. melanogaster and its sister species D. simulans have revealed that a substantial proportion of proteins and amino acid residues exhibit adaptive divergence on a roughly few million years long timescale. Here we use population genomic approaches to attack the problem of parallelism between D. melanogaster and a highly diverged conger, D. hydei, on two timescales. D. hydei, a member of the repleta group of Drosophila, is similar to D. melanogaster, in that it too appears to be a recently cosmopolitan species and recent colonizer of high latitude environments. We observed parallelism both for genes exhibiting latitudinal allele frequency differentiation within species and for genes exhibiting recurrent adaptive protein divergence between species. Greater parallelism was observed for long-term adaptive protein evolution and this parallelism includes not only the specific genes/proteins that exhibit adaptive evolution, but extends even to the magnitudes of the selective effects on interspecific protein differences. Thus, despite the roughly 50 million years of time separating D. melanogaster and D. hydei, and despite their considerably divergent biology, they exhibit substantial parallelism, suggesting the existence of a fundamental predictability of adaptive evolution in the genus.

Highlights

While parallel phenotypic evolution has long been recognized as one of the strongest pieces of evidence for adaptation [1], the general repeatability of adaptive evolution in natural populations remains poorly understood
Despite the fact that these species diverged from a common ancestor roughly 50 million years ago, the population genomics of latitudinal allele frequency differentiation shows that there is a substantial shared set of genes likely playing a role in the short term adaptive divergence of populations in both species
Analyses of longer-term adaptive protein divergence for the D. hydei-D. mojavensis and D. melanogaster-D. simulans clades reveal a striking level of parallel adaptation

Summary

Introduction

While parallel phenotypic evolution has long been recognized as one of the strongest pieces of evidence for adaptation [1], the general repeatability of adaptive evolution in natural populations remains poorly understood. Two lineages may have adapted to similar selection pressures through substitutions in largely non-overlapping genes which belong to the same pathway. This would represent convergence at the level of pathway but not the level of gene. In addition to major gaps in our descriptions of the frequency with which parallelism occurs at different levels of biological organization (from single nucleotides to complex phenotypes), we have little understanding of how lineage divergence in biological processes, ecology, or population genetics, may interact to influence the probability of parallelism at different levels of organization.

Methods

Results

Discussion

Conclusion