Genotype × environment assessment for grain quality traits in rice

Physicochemical properties of 10 rice advanced breeding lines (BC2F7) across 3 environments in Malaysia were evaluated. Highly significant G × E interactions for all measured quality traits were detected using ANOVA, and the additive main effect and multiplicative interaction (AMMI) statistical model was applied to analyze them. The results showed that the grain quality parameters had large genotype by environment (G × E) interactions. Differences among genotypes and environments accounted for 16-73% and 0.5-56% of the total sum of squares, respectively, while the G × E interaction accounted for 15-52% of the total sum of squares. The first and second AMMI axes captured 67-96% and 4-33% of the total variation due to G × E interaction, respectively. The biplots of genotypes/environments means and scores on first Principal Component Axis (AMMI-1 biplot) for all traits accounted for most of the total treatment sum of squares. Genotypes G7 (in terms of head rice percentage and amylose content) and G16 (in terms of head rice percentage, protein and amylose content) were detected as winning genotypes in mega-environments, according to the AMMI-1 model. The best genotype in one environment was not always best in other test environments. However, most genotypes showed higher quality parameters in Bumbung Lima and Sungai Besar than in Gurun.