2.6 | Genome-Wide Association Studies (GWAS)
GWAS were performed on the two different phenotypic traits (DLA-R and DLA-S) estimated for each isolate in (Dumartinet et al., 2019) using two different statistical models: the multi-locus mixed linear model (MLMM, (Segura et al., 2012)) and the settlement of MLM under progressively exclusive relationship (SUPER) (Wen et al., 2018) both implemented in GAPIT software (Lipka et al., 2012). The first model accounts for the linkage disequilibrium between loci associated with traits while the second model is known to have higher statistical power than regular mixed linear models. All GWAS were corrected for population genetics by estimating a distance matrix between all isolates, and all single-marker p-values were corrected using the Benjamini–Hochberg FDR procedure (Benjamini & Hochberg, 1995). The above-mentioned local score procedure was also applied to GWAS p-values to increase GWAS resolution to detect clusters of loci with low p-values, as in Bonhomme et al. 2019.