Modeling plant phenotypic plasticity and its underlying genetic architecture (France, 2022-2025)

Phenotypic plasticity can contribute to crop adaptation to challenging environments. Plasticity indices are potentially useful to identify the genetic basis of crop phenotypic plasticity. Numerous methods exist to measure phenotypic plasticity. Among those, the ratio (also called the response coefficient; Bouslama and Schapaugh, 1984) or the relative distance plasticity index (RDPI; Valladares et al., 2006) rely on pairwise comparison of phenotypic means between environments (e.g. non-optimal versus optimal growth environments for instance). The coefficient of variation (CV) was calculated as the standard deviation of the trait for a genotype in all environments divided by the genotype phenotypic mean across all environments (Francis and Kannenberg, 1978). The Finlay-Wilkinson joint linear regression analysis corresponds to regression of the phenotypic mean of a trait for a given genotype against the averaged phenotypic means of all the genotypes in each studied environment (Finlay and Wilkinson, 1963). It allows calculation of the slope (also called linear plasticity) that represents the sensitivity of the trait across the environments for a genotype, and the mean square error from the regression (MSE, also called non-linear plasticity) that describes the deviation of the measured phenotypic value from the predicted phenotype in a given environment. The AMMI model uses an analysis of variance to characterize the genotype and environment main effects, and a principal component analysis to characterize their interactions (Gabriel, 1978). Parameters of AMMI analysis, such as the AMMI Zhang’s Euclidean distance (D) which is the distance of the interaction principal component point (IPC) from its origin in space (Zhang and Xiang, 1998; Xie et al., 2021), or the AMMI stability value (ASV) which considers the magnitude of the sum of squares of the first two IPC, have also been used as phenotypic plasticity indices (Purchase et al., 2000). This dataset contains plasticity indices (ratio, RDPI, CV, Slope, MSE, D, ASV) calculated for leaf area, shoot biomass and water use efficiency on 254 maize genotypes grown under well-watered and water stress treatments in an environmentally-controlled phenotyping platform, in four independent trials performed in different years and seasons (values for traits in each environment were published by Alvarez Prado et al., 2017). The scripts used for calculation of the plasticity indices, for estimation of the genetic variance, for identification of QTL associated with the plasticity indices using GWAS, to measure the contribution of the identified QTL to the genetic variance and to perform a step forward analysis to evaluate the efficiency of the analysis are provided. We also provide a Table comprehensively describing the equation used for calculation of the different plasticity indices. The ability QTL associated with plasticity indices to contribute to the genetic variance was compared to that of QTL associated with the phenotypic means.