Introgression in the Drosophila subobscura--D. Madeirensis sister species: evidence of gene flow in nuclear genes despite mitochondrial differentiation

Abstract

Species hybridization, and thus the potential for gene flow, was once viewed as reproductive mistake. However, recent analysis based on large datasets and newly developed models suggest that gene exchange is not as rare as originally suspected. To investigate the history and speciation of the closely related species Drosophila subobscura, D. madeirensis, and D. guanche, we obtained polymorphism and divergence data for 26 regions throughout the genome, including the Y chromosome and mitochondrial DNA. We found that the D. subobscura X/autosome ratio of silent nucleotide diversity is significantly smaller than the 0.75 expected under neutrality. This pattern, if held genomewide, may reflect a faster accumulation of beneficial mutations on the X chromosome than on autosomes. We also detected evidence of gene flow in autosomal regions, while sex chromosomes remain distinct. This is consistent with the large X effect on hybrid male sterility seen in this system and the presence of two X chromosome inversions fixed between species. Overall, our data conform to chromosomal speciation models in which rearrangements are proposed to serve as gene flow barriers. Contrary to other observations in Drosophila, the mitochondrial genome appears resilient to gene flow in the presence of nuclear exchange.

Keywords: Adaptation; gene flow; inversions; population genetics; speciation.

Figure 1

Neighbor-joining trees reconstructed from total nucleotide variation. Bootstrap values were obtained after 1000 replicates and values higher than 85% are shown. Drosophila madeirensis, D. subobscura, and D. guanche are indicated with filled circles, open squares, and filled triangles, respectively. (A) RpS26, (B) CG8768, (C) Prosβ1, and (D) mtDNA.

Figure 2

Smoothed marginal posterior probability distributions for (A) time since population split, (B) effective population sizes, and (C) migration.

Figure 3

Maximum likelihood (ML) estimates of migration rates obtained with MIMAR for individual nuclear regions.