Ancient polymorphisms contribute to genome-wide variation by long-term balancing selection and divergent sorting in Boechera stricta

Abstract

Background Genomic variation is widespread, and both neutral and selective processes can generate similar patterns in the genome. These processes are not mutually exclusive, so it is difficult to infer the evolutionary mechanisms that govern population and species divergence. Boechera stricta is a perennial relative of Arabidopsis thaliana native to largely undisturbed habitats with two geographic and ecologically divergent subspecies. Here, we delineate the evolutionary processes driving the genetic diversity and population differentiation in this species. Results Using whole-genome re-sequencing data from 517 B. stricta accessions, we identify four genetic groups that diverged around 30–180 thousand years ago, with long-term small effective population sizes and recent population expansion after the Last Glacial Maximum. We find three genomic regions with elevated nucleotide diversity, totaling about 10% of the genome. These three regions of elevated nucleotide diversity show excess of intermediate-frequency alleles, higher absolute divergence ( d XY ), and lower relative divergence ( F ST ) than genomic background, and significant enrichment in immune-related genes, reflecting long-term balancing selection. Scattered across the genome, we also find regions with both high F ST and d XY among the groups, termed F ST - islands. Population genetic signatures indicate that F ST - islands with elevated divergence, which have experienced directional selection, are derived from divergent sorting of ancient polymorphisms. Conclusions Our results suggest that long-term balancing selection on disease resistance genes may have maintained ancestral haplotypes across different geographical lineages, and unequal sorting of balanced polymorphisms may have generated genomic regions with elevated divergence. This study highlights the importance of ancestral balanced polymorphisms as crucial components of genome-wide variation. Electronic supplementary material The online version of this article (10.1186/s13059-019-1729-9) contains supplementary material, which is available to authorized users.

Evidence for widespread selection in shaping the genomic landscape during speciation of Populus

“…Recurrent natural selection, either purifying or positive selection, has therefore been reducing N e since at least the time when the ancestral P. tremula – P. davidiana population existed, and hence made ILS less likely to occur in regions of high functional density and low recombination rate (Charlesworth et al, 1993; Martin et al, 2019; Pease & Hahn, 2013). Unlike positive and purifying selection, we found that regions under balancing selection show increased ILS, which is in agreement with the expectation that balancing selection maintains ancient, trans‐species polymorphisms and hence results in a longer time for lineage sorting to occur (Charlesworth, 2006; Wang et al, 2019).…”

Reproductive barriers as a byproduct of gene network evolution

“…In addition, many models make an implicit assumption that two allopatric lineages only differ by fixed alleles, which does not capture the empirical diversity among individuals' gene expression (Kelly et al, 2017;Tyler et al, 2017;Gould et al, 2018;Mogil et al, 2018;Ryu et al, 2019) nor the observed importance of regulatory disruption and standing genetic variation in generating reproductive isolation (Hopkins and Rausher, 2011;Wittkopp and Kalay, 2012;Guerrero et al, 2016;Rougeux et al, 2019;Morgan et al, 2020). More importantly, substitutions originating from de-novo mutations fail to explain the recent evidence that alleles underlying reproductive barriers often predate speciation events and can evolve along parallel evolutionary trajectories (Kaeuffer et al, 2012;Sicard et al, 2015;Meier et al, 2017;Nelson and Cresko, 2018;Wang et al, 2019;Duranton et al, 2019;Marques et al, 2019).…”

Evidence for widespread selection in shaping the genomic landscape during speciation ofPopulus

“…Moreover, we found lower weightings of the ‘species topology’, higher ILS, and lower f d in the candidate balancing selection regions although the results were not significant, likely due to the small number of windows showing evidence for balancing selection (Figure 6C,D). We therefore infer that long-term balancing selection may not only influence the genomic landscape of diversity and divergence but may also play a role in shaping the genealogical relationship and barriers to introgression among closely related species (Charlesworth, 2006; Wang et al, 2019).…”