Phenotypic plasticity
Three different parameters were used to quantify phenotypic plasticity in the MAGIC-MET design. For each trait, the slope and VAR from the Finlay-Wilkinson regression model and the genotypic sensitivity to the most important environmental covariate (SCv) from the factorial regression model were extracted. A large genetic variability was observed for plasticity of all traits (Supplemental Figure 5 and Supplemental Figure 6). Besides, significant correlations were found between the mean phenotypes and plasticity parameters (Figure 2) for most of the traits. The best average-performing genotypes were usually the most responsive to environmental variation as highlighted by the positive correlation between the genotypic means and slope from the Finlay-Wilkinson regression model. The majority of the MAGIC lines responded in the same direction to the environmental quality and only a few genotypes (none in the case of height) showed negative reaction norms; however, more divergent shapes of reaction norms were observed from the factorial regression model (Supplemental Figure 5).