Deciphering genotype × environment interaction for grain yield in durum wheat: an integration of analytical and empirical approaches for increased yield stability and adaptability

Abstract

The development of stable and high-yielding wheat cultivars offers a sustainable solution to the challenge of food security and self-sufficiency in developing countries. The main goals of this study were to evaluate the effects of genotype, environment and genotype by environment (G×E) interaction on grain yield in durum wheat genotypes, to identify high-yielding and stable genotypes, and to identify climatic variables that significantly affect the G×E interaction. Twenty-one durum wheat breeding lines originating from ICARDA and CIMMYT, along with four national durum wheat cultivars, were evaluated using a randomized complete block design with three replications across seven locations (differing in winter temperature and rainfall) and three cropping seasons (2020–23). Four statistical models, including (i) additive main effects and multiplicative interaction (AMMI) (ii) genotype plus G×E (GGE) biplots, (iii) factorial regression (FR) and (iv) partial least squares (PLS) regression for investigating the G×E interaction for grain yield and identifying the climatic variables that significantly affect the G×E interaction, were applied. The combined analysis of variance indicated that the effects due to genotype, environment and the G×E interaction were highly significant (P < 0.01). The environment was the main source of variation and accounted for 94.2 % of the total grain yield variation, while the G×E interaction contributed 4.7 %, and the genotype contributed 0.5 %. The combined and yearly data analysis by the “which-won-where” pattern of the GGE biplot showed consistent results across years for environmental grouping, resulting in four mega-environments in durum wheat yield trials. These results suggested that the use of these genotypes could be recommended for deployment in their respective mega-environments. Both AMMI and GGE biplots approved selecting breeding lines G19, G12, G22 and G23 as high-yield and stable genotypes across diverse environments for further breeding programs and genotype recommendation. Based on the FR model, climatic variables related to monthly rainfall and temperature explained 69.5 % of the G×E interaction variation. Using the PLS biplot, the environments were separated based on temperature and rainfall, and the genotypes with the most sensitivity (i.e., G4, G9, G24, G25) or insensitivity (i.e., G23, G21 and G14) to climatic variables were identified. These findings provide relevant information for future durum wheat breeding programs that consider improved productivity and yield stability in durum wheat under climate change conditions.