Supplemental files
Supplemental Figure 1: Selection of 7 environmental covariates for the factorial regression model. Three periods – each of 20 days – were defined from planting to the end of flowering on the 4th truss. The period from 20 to 60 days after planting (DAP) covered vegetative growth and flowering on the 4th truss and the measured climatic variables averaged during this period. The different environmental covariates are described
Supplemental Figure 2: Boxplot distribution of the traits across environments. The colors of the boxplot are according to the groups defined by clustering of the environments
Supplemental Figure 3: Heritability in the MAGIC-MET design. For each trait, heritability was computed at every environment and plotted with heritability of the full design \(\mathbf{H}^{\mathbf{2}}\)(in green)
Supplemental Figure 4: Proportion of the sum of square attributed to the different factors in the factorial regression model. For each trait, the orange and green stacked bars represent the proportion of the SSq explained by the Genotype and Environment factors in model (4). The remaining colors represent the effect part of the GxE that could be explained by the different environmental covariates. Only significant covariates were highlighted within the bars.
Supplemental Figure 5: Reaction norms from the Finlay-Wilkinson regression model (A) and the factorial regression model (B). In figure 5 A, the blue and orange lines represent the positive and negative reaction norms. In Figure 5 B, the green and purple lines represent the positive and negative reaction norms
Supplemental Figure 6: Histogram distribution of mean and all plasticity parameters for each trait
Supplemental Figure 7: Physical positions of the MAGIC-MET QTLs for diam, leaf, height, fset, nflw, nfr, firm and SSC. The outer circle with gray font represents the known and cloned QTL/gene for each trait. The following circle with black bars represents the different domestication/improvement sweep regions identified in (Zhu et al. 2018). The other circles plot the CI of QTLs identified on mean (green), plasticity (orange) or with QEI analysis (purple)
Supplemental Figure 8: Number of the MAGIC-MET QTLs identified within or outside the domesticated/improved regions. Only the MAGIC-MET QTLs within short CI (lower than 2Mbp) were considered. The response specific category included QEI and plasticity specific QTLs; the common category correspond to QTLs that were commonly identified on mean, plasticity and QEI or at least two of them
Supplemental Figure 9: Zoom plot on Chromosome 11 region from 53 -57 Mbp. Each color represents a different QTL located in this region and the top black bars are the Sweep regions SW254 and SW255
Supplemental Figure 10: Correlation between the genotypic sensitivities to environmental covariates from the factorial regression model and slopes from the Finlay-Wilkinson regression model
Supplemental Figure 11: Venn diagram of the number of QTL specific or commonly detected with mean, PP or using the QEI models.
Supplemental Table 1: Description of the MAGIC-MET design with the 12 environments and their respective names
Supplemental Table 2: Description of the phenotypic traits evaluated in the MAGIC-MET design
Supplemental Table 3: Estimates of the variance components from model (2)
Supplemental Table 4: Results of QTL and QEI analysis in the MAGIC-MET design
Supplemental Table 5: Genetic location of the MAGIC-MET QTLs overlapping with the Sweep (domestication/improvement) regions.
Supplemental Table 6: QTLs identified for fruit size, fruit width and fruit length in the MAGIC population
Supplemental Table 7: Selected candidate genes for all the mean and plasticity QTLs located within 2Mbp CI region