Using Comparative Population Genomics to Find the Genes Driving Convergent Local Adaptation

Local adaptation is common in many species but we still know little about its genetic basis: how many genes are involved and how repeatable is evolution at the genomic level? Genome-wide association studies (GWAS) provide a window into evolution and a means to answer this question. However, a major challenge in using GWAS arises when population structure is oriented along the same axis as variation in the selection pressures that drive adaptation. When this happens, failure to correct for population structure will result in many false positives, but methods that correct for population structure may remove much of the true positive signal of local adaptation. Comparisons of GWAS results among species can resolve this problem, because drift is unlikely to generate a strong association in a given gene in more than one species, so correction for population structure is unnecessary. Here we use capture arrays to sequence a large proportion of the gene space in individuals from hundreds of populations in two conifers: lodgepole pine and interior spruce. We develop novel methods for comparing signatures of local adaptation among species, first identifying top candidates for climate adaptation within species and then testing whether orthologs in the alternate species also have strong signatures of association. We identify a large number of genes with signatures of local adaptation in both species, which are enriched for duplicates and genes with conserved patterns of differential expression. We also explore the robustness of our comparative method for applications to other species.