Complementary methods to identify environment-responsive QTLs
Different approaches have been proposed in the literature to dissect GxE into its genetic components (Malosetti et al., 2013; El-Soda et al., 2014). We used a mixed linear model with a random genetic effect accounting for the correlation structure of the MAGIC-MET design to identify the QEI. Extending the use of mixed linear models to MAGIC populations in the framework of MET analysis has been very rarely applied in crops. To our knowledge, only Verbyla et al., (2014) applied such approach in wheat and identified QEI for flowering time. Our model was adequate to account for the complex mating design of the MAGIC population by using the haplotype probabilities. Indeed, it allows estimating the QTL effect for each parental allelic class and for each environment at every SNP marker. Overall, 28 QEI were detected showing significant marker x environment interaction for ten traits.
Methods using plasticity as a trait per se are also attractive to identify environmentally sensitive QTLs. This strategy was applied in maize, sunflower, barley and soybean to detect the loci governing GxE (Lacaze et al., 2009; Gage et al., 2017; Kusmec et al., 2017; Mangin et al., 2017; Xavier et al., 2018). With different plasticity parameters, we identified a total of 63 plasticity QTLs and only 24% were also identified with the QEI models. Thus, both methods, using plasticity or mixed linear models, are complementary approaches to study the genetic component of GxE.