Résumé
Introduction :
Perennial ryegrass (Lolium perenne) is a grass species prevalent in natural grasslands across
Europe and also one of the main species bred for forage usage worldwide. In Europe, increased risk of
heat waves and severe drought could lead to production losses and high mortality in natural populations.
To address this problem, we need to develop breeding programmes delivering genotypes adapted to
future conditions. A !rst step required to design such breeding programmes is the identi!cation of genes
associated with climatic adaptation.
Materials and methods :
We used high-throughput genotyping (ca. 190,000 SNPs from genotyping-by-sequencing pool-Seq) and phenotyping (155 traits) to detect candidate loci under positive environmental induced selection. We investigated associations with bioclimatic indices, ecophysiological indices and soil properties. Our approach combined redundancy analysis (RDA; Forester et al., 2015) and mixed model predictions to detect candidate loci and identify environmental variables imposing selection. Additionally, we implemented a genome-wide association study (GWAS) to identify phenotypes linked to environmental adaptation in perennial ryegrass.
Results :
Te RDA analysis identi!ed 2,819 candidate SNPs. 396 of these candidates were highly associated with one or more environmental variables (mixed model, P<0.001). A high proportion of these candidates (176 SNPs) showed highest association with environmental variables linked to heat/water stress. Eleven out of 176 SNPs were also highly associated with nine di#erent traits in the GWAS analysis (P<0.001): acid detergent lignin content, acid detergent !bre content, spring canopy height, spring canopy height independent of heading earliness, autumn growth rate, growth vigour, autumn maximum height, water soluble carbohydrates content and soil coverage.
Conclusions :
Our results reveal that adaptation to heat stress in perennial ryegrass is associated with multiple genes. Most adaptive loci did not show association with phenotypic traits with agronomic value. But eleven alleles associated with adaptation to heat stress were also negatively associated with agronomic value, and therefore, should be avoided in breeding programmes. These alleles were associated with higher
lignin and !bre content, lower spring canopy height, lower autumn growth rate, lower water-soluble
carbohydrates content and lower soil coverage. This study will provide opportunities for the development
of new varieties of perennial ryegrass both adapted to future warmer climatic conditions and with high
agronomic value