Globally Relaxed Selection and Local Adaptation inBoechera stricta

Abstract

The strength of selection varies among populations and across the genome, but the determinants of efficacy of selection remain unclear. In this study, we used whole-genome sequencing data from 467 Boechera stricta accessions to quantify the strength of selection and characterize the pattern of local adaptation. We found low genetic diversity on 0-fold degenerate sites and conserved non-coding sites, indicating functional constraints on these regions. The estimated distribution of fitness effects and the proportion of fixed substitutions suggest relaxed negative and positive selection in B. stricta. Among the four population groups, the NOR and WES groups have smaller effective population size (Ne), higher proportions of effectively neutral sites, and lower rates of adaptive evolution compared to UTA and COL groups, reflecting the effect of Ne on the efficacy of natural selection. We also found weaker selection on GC-biased sites compared to GC-conservative (unbiased) sites, suggested that GC-biased gene conversion (gBGC) has affected the strength of selection in B. stricta. We found mixed evidence for the role of recombination rate on the efficacy of selection. Positive and negative selection was stronger in high-recombination regions compared to low-recombination regions in COL but not other groups. By scanning the genome, we found different subsets of selected genes suggesting differential adaptation among B. stricta groups. These results show that differences in effective population size, nucleotide composition and recombination rate are important determinants of the efficacy of selection. This study enriches our understanding of the roles of natural selection and local adaptation shaping genomic variation.

Increased positive selection in highly recombining genes does not necessarily reflect an evolutionary advantage of recombination

“…Even without gBGC, using the same framework, previous theoretical studies demonstrated that the nearly universal mutation bias towards AT necessarily induces a fixation bias towards GC in selectively constrained sequence, also because of beneficial back-mutations (Latrille and Lartillot, 2022; Kaj et al, 2023). This calls for a re-interpretation of results from several studies that interpreted the fact that GC alleles segregate at higher frequency at non-synonymous sites as an evidence for the negative impact of gBGC on adaptation (Hämälä and Tiffin, 2020; Liang et al, 2022). For the same reasons that increased positive selection in highly recombining genes is not an evidence for a beneficial effect of recombination, a fixation bias towards GC at non-synonymous sites is not an evidence for the presence of gBGC.…”

“…Increased positive selection in highly recombining genes does not necessarily reflect an evolutionary advantage of recombination higher frequency at non-synonymous sites as an evidence for the negative impact of gBGC on adaptation (Hämälä and Tiffin, 2020;Liang et al, 2022). For the same reasons that increased positive selection in highly recombining genes does not necessarily reflect a beneficial effect of recombination, a fixation bias towards GC at non-synonymous sites does not necessarily reflect the presence of gBGC.…”

Increased Positive Selection in Highly Recombining Genes Does not Necessarily Reflect an Evolutionary Advantage of Recombination

“…Even without gBGC, this work is consistent with previous theoretical studies showing that the nearly universal mutation bias towards AT necessarily induces a fixation bias towards GC in selectively constrained sequence, also because of beneficial back-mutations ( Latrille and Lartillot 2022 ; Kaj et al 2023 ). This calls for a re-interpretation of results from several studies that interpreted the fact that GC alleles segregate at higher frequency at nonsynonymous sites as an evidence for the negative impact of gBGC on adaptation ( Hämälä and Tiffin 2020 ; Liang et al 2022 ). For the same reasons that increased positive selection in highly recombining genes does not necessarily reflect a beneficial effect of recombination, a fixation bias towards GC at nonsynonymous sites does not necessarily reflect the presence of gBGC.…”

A Hyb‐Seq phylogeny of Boechera and related genera using a combination of Angiosperms353 and Brassicaceae‐specific bait sets

“…Boechera stricta (previously known as Arabis drummondii ) is sister to all other members of the erect fruit clade. It is widespread in the mountains and boreal regions of western North America and has been a primary focus of research within Boechera (Song et al, 2006; Windsor et al, 2006; Schranz et al, 2007, 2009; Manzaneda et al, 2010; Lee and Mitchell‐Olds, 2013; Lee et al, 2014; Naithani et al, 2014; Anderson and Gezon, 2015; Anderson et al, 2011, 2015; Salmela et al, 2016; Heo et al, 2018; Lee et al, 2018; Rojek et al, 2018; Wagner and Mitchell‐Olds, 2018; Bemmels and Anderson, 2019; Olsen et al, 2019; Wang et al, 2019; Hamann et al, 2021; Lin et al, 2021; Liang et al, 2022). It is one of the most distinctive species in the genus, easily separated from all other sexual diploids by its unique pubescence.…”