Genomics of ecological adaptation in Canary Island Descurainia (Brassicaceae) and comparisons with other Brassicaceae.

Oceanic archipelagos provide striking examples of lineages that have radiated over pronounced ecological gradients. Accompanying this diversification, lineages have evolved adaptations allowing survival in extreme environments. Here, we investigate the genomic basis of ecological adaptation in Canary Island (Brassicaceae), an island relative of . The seven endemic species have diversified in situ along an elevational and ecological gradient, from low‐elevation scrub to high‐elevation sub‐alpine desert. We first generated a reference genome for phylogenetic analysis of which placed it as sister to . Ninety‐six gene families were found to be specific to and a further 1087 and 1469 gene families have expanded or contracted in size, respectively, along the branch. We then employed genome re‐sequencing to sample 14 genomes across the seven species of Canary Island and an outgroup. Phylogenomic analyses were consistent with previous reconstructions of Canary Island in resolving low‐ and high‐elevation clades. Using the branch‐site dN/dS method, we detected positive selection for 275 genes on the branch separating the low‐ and high‐elevation species and these positively selected genes (PSGs) were significantly enriched for functions related to reproduction and stress tolerance. Comparing PSGs to those in analyses of adaptation to elevation and/or latitude in other Brassicaceae, we found little evidence of widespread convergence and gene reuse, except for two examples, one of which was a significant overlap between and a species restricted to high latitudes. The study of Canary Island suggests that the transition to high‐elevation environments such as that found in the high mountains of the Canary Islands involves selection on genes related to reproduction and stress tolerance but that repeated evolution across different lineages that have evolved into similar habitats is limited, indicating substantially different molecular trajectories to adaptation.