Selection and demography shape genomic variation in a ‘Sky Island’ species

Abstract

Over time, populations of species can expand, contract, and become isolated, creating subpopulations that can adapt to local conditions. Understanding how species adapt following these changes is of great interest, especially as the current climate crisis has caused range shifts for many species. Here, we characterize how Drosophila innubila came to inhabit and adapt to its current range: mountain forests in southwestern USA separated by large expanses of desert. Using population genomic data from more than 300 wild-caught individuals, we examine four distinct populations to determine their population history in these mountain-forests, looking for signatures of local adaptation to establish a genomic model for this spatially-distributed system with a well understood ecology. We find D. innubila spread northwards during the previous glaciation period (30-100 KYA), and has recently expanded even further (0.2-2 KYA). Surprisingly, D. innubila shows little evidence of population structure, though consistent with a recent migration, we find signatures of a population contraction following this migration, and signatures of recent local adaptation and selective sweeps in cuticle development and antifungal immunity. However, we find little support for recurrent selection in these genes suggesting recent local adaptation. In contrast, we find evidence of recurrent positive selection in the Toll-signaling system and the Toll-regulated antimicrobial peptides.

Recurrent evolution of high virulence in isolated populations of a DNA virus

“…Given that siRNA genes are not highly ubiquitously expressed in D. innubila ( Hill et al, 2019 ), that DiNV reduces host fecundity ( Unckless, 2011 ), and that a close relative of DiNV infects the host ovaries ( Palmer et al, 2018 ), DiNV could interact with chorion proteins during oogenesis, and piRNAs could be suppressing DiNV ( Lewis et al, 2018 ). Consistent with the arms race model, host genes we suspect are interacting with DiNV (such as the association study hits, AMPs, chorion genes, piRNA genes, and extracellular genes) show elevated levels of substitutions fixed by selection compared to background genes in D. innubila ( Figure 4 and Figure 4—figure supplement 1 , GLM p-value<0.05) ( Hill and Unckless, 2020 ). Finally, differentially expressed chorion genes, extracellular genes and AMPs have significantly more adaptive substitutions than non-differentially expressed genes in the same categories ( Figure 4 , blue dots, differentially expressed versus all other T-test: D. innubila t-value = 4.755, p-value=0.000671).…”

“…To distinguish between these possibilities and determine the timeframe of divergence, we used the site frequency spectrum of silent DiNV polymorphism to estimate effective population size backwards in time for all populations ( Liu and Fu, 2015 ). We found that the three populations (CH, HU and SR) expanded from a single viral particle (N e = 1) to millions of particles during the last glacial maximum (30–100 thousand years ago) when D. innubila settled its current range ( Figure 5—figure supplement 4 ; Hill and Unckless, 2020 ). This supports a single invasion event during a host-range change for each location.…”

“…To characterize the evolutionary dynamics of wild Drosophila innubila Nudivirus (DiNV) in its host ( D. innubila ), we sequenced wild-caught individuals from four populations with the expectation that some (~40% in previous samples) individuals would be infected ( Hill and Unckless, 2020 ). We considered strains to be infected with DiNV if they had at least 10x coverage for 95% of the genome.…”

“…This supports a single invasion event during a host-range change for each location. PR appears to expand between 1 and 10 thousand years ago, suggesting a much more recent bottleneck during the range expansion in PR ( Figure 5—figure supplement 4 ; Hill and Unckless, 2020 ).…”

“…In this study, we used previously collected and sequenced D. innubila ( Hill and Unckless, 2020 ). Briefly we collected these flies across the four mountainous locations in Arizona between the 22nd of August and the 11th of September 2017.…”