Résumé
As climate change drives rapid environmental shifts, crops face significant challenges due to their limited genetic diversity, which constrains their adaptive potential. Crop wild relatives —and potentially feral populations, which are wild populations derived from domesticated ones—are likely to be locally adapted to diverse environments and could serve as valuable reservoirs of genetic diversity for breeding and crop improvement. Exploring and conserving this genetic diversity is essential, as it preserves a broad spectrum of adaptations to varied environmental conditions. In this study, we investigate the genetic basis of local adaptation to climate and potential mismatches under future conditions in Brassica rapa (field mustard), a wild relative of turnip, Chinese cabbage, and pak choi, and a genome donor for Brassica napus (rapeseed). Seeds were collected through the BrasExplor project—a consortium coordinated by IGEPP (INRAE) with 11 partners from seven countries. A total of 62 spontaneous populations were sampled in Algeria, France, Italy, and Slovenia, revealing strong genetic structure between feral populations (from continental Europe) and wild populations (from Algeria and Sicily). We conducted genome–environment association analyses and applied the local score approach to identify highly significant genomic regions associated with climatic variables, including promising variants and candidate genes. However, broader genomic patterns suggest that adaptation to climate is not limited to a few loci but likely reflects a polygenic architecture involving multiple regions across the genome. While the inclusion of feral populations allows exploration of a broader climatic envelope and identifies variants potentially easier to introgress into crops, their reduced genetic diversity—likely due to domestication and rewilding bottlenecks—may limit our ability to capture the full spectrum of adaptive variation present in truly wild populations. This hypothesis was confirmed by repeating the analysis on a subset of wild populations, revealing a large number of additional candidate regions. We also conducted genomic offset (GO) analyses to estimate the potential mismatch between current allele frequencies and those expected under future climates. Several populations showed high GO values, suggesting nonoptimal allele frequencies in their future climate. Phenotypic data obtained in a common garden experiment indicated that GO may partially predict variation in fitness, but this relationship should be interpreted with caution. More specifically, fitness proxies (seed number and weight) may be biased in feral populations, and much of the adaptive signal specific to wild populations is not captured, limiting predictive power for those populations. These findings highlight the complex genetic architecture of climate adaptation in B. rapa and underscore both the potential and the limitations of genomic tools for guiding conservation and crop improvement strategies in a changing climate.