AbstractLocal adaptation across species range is widespread. Yet, much has to be discovered on its environmental drivers, the underlying functional traits and their molecular determinants. Because elevation gradients display continuous environmental changes at a short geographical scale, they provide an exceptional opportunity to investigate these questions. Here, we used two common gardens to phenotype 1664 plants from 11 populations of annual teosintes. These populations were sampled across two elevation gradients in Mexico. Our results point to a syndrome of adaptation to altitude with the production of offspring that flowered earlier, produced less tillers, and larger, longer and heavier grains with increasing elevation. We genotyped these plants for 178 outlier single nucleotide polymorphisms (SNPs), which had been chosen because they displayed excess of allele differentiation and/or correlation with environmental variables in six populations with contrasted altitudes. A high proportion of outlier SNPs associated with the phenotypic variation of at least one trait. We tested phenotypic pairwise correlations between traits, and found that the higher the correlation, the greater the number of common associated SNPs. In addition, allele frequencies at 87 of the outlier SNPs correlated with an environmental component best summarized by altitudinal variation on a broad sample of 28 populations. Chromosomal inversions were enriched for both phenotypically-associated and environmentally-correlated SNPs. Altogether, our results are consistent with the set-up of an altitudinal syndrome promoted by local adaptation of teosinte populations in the face of gene flow. We showed that pleiotropy is pervasive and potentially has constrained the evolution of traits. Finally, we recovered variants underlying phenotypic variation at adaptive traits. Because elevation mimics climate change through space, these variants may be relevant for future maize breeding.Author summaryAcross their native range, species encounter a diversity of habitats promoting local adaptation of geographically distributed populations. While local adaptation is widespread, much has yet to be discovered about the conditions of its emergence, the targeted traits, their molecular determinants and the underlying ecological drivers. Here we employed a reverse ecology approach, combining phenotypes and genotypes, to mine the determinants of local adaptation of teosinte populations distributed along two steep altitudinal gradients in Mexico. Evaluation of 11 populations in two common gardens located at mid-elevation pointed to the set-up of an altitudinal syndrome, in spite of gene flow. We scanned genomes to identify loci with allele frequencies shifts along elevation. Interestingly, variation at these loci was commonly associated to variation of phenotypes. Because elevation mimics climate change through space, these variants may be relevant for future maize breeding.
Two are better than one: combining landscape genomics and common gardens for detecting local adaptation in forest trees
