Genome-wide association study (GWAS) and linkage disequilibrium
Reaction norms connect the environments to which a particular genotype is exposed and the phenotypes produced by that genotype in these environments. From the reaction norm values, two reaction norm-norm parameters (intercept and slope) were obtained for each trait using a joint regression analysis of the observed trait values across environments (Fig. 1c). In addition, with the identified environmental indices, the slope was obtained for each index using a joint regression analysis of the observed trait values across environments (Fig. S9). Treating the estimates of intercept and slope as a derived trait, an established mixed-model GWAS (Zhang et al., 2010, Yu et al., 2006) implemented in GAPIT (v.3) R package (http://zzlab.net/GAPIT) (Wang and Zhang, 2021) was used, separately for each trait, to identify genomic regions underlying the observed variation for slope and intercept across different genotypes.
The significance thresholds of association were calculated with Genetic type I error calculator (GEC) software (Li et al., 2012), and the calculated significance threshold was 7.77 × 10−7. To identify candidate genes, genes within the 100-kb upstream and downstream regions of significant SNPs were extracted, according to the linkage disequilibrium (LD) decay rate estimated by previous studies (Tang et al., 2021).