Comparative Analysis of Statistical Models for Evaluating Genotype × Environment Interaction in Rainfed Safflower

The main objective of this study was to compare the several statistical models, i.e., joint regression analysis (JRA), additive main effect and multiplicative interaction (AMMI) and genotype and genotype × environment (GE) interaction (GGE) biplot, for analyzing of GE interaction for grain yield of rainfed safflower multi-environment trials. The effectiveness of each model was compared for identifying the best performing genotypes across environments, identifying the best genotypes for mega-environment differentiation and evaluating the yield and stability performance. Grain yield data of 13 cold-tolerant safflower breeding lines along with a check cultivar grown in three rainfed research stations for two cropping seasons were used. Environment (E) main effect accounted for 57.1% of total variation, compared to 8.8 and 34.1% for G and GE interaction effects, respectively. Spearman's rank correlation analysis indicated that the three methods (GGE biplot, AMMI analysis and JRA) were significantly correlated (P < 0.01) in ranking of genotypes for static (biological) stability, suggesting that they can be used interchangeably. All three methods identified genotypes G4 and G9 as the most stable genotypes with low-yielding performance, and the breeding line G3 as high-yielding stable genotype across environments. Based on the results, the Maragheh was an ideal test location with a demonstrated high efficiency in selecting new cultivars with a wide adaptability. The main conclusions were the similarity between the dominant genotypes in the three models. The GGE biplot was more versatile and flexible and provided a better understanding of GE interaction than the other methods. Positive increase in yield and yield stability is attributable predominately to genetic improvement in safflower breeding lines. The breeding line 415/338 could serve as a good genetic source for both high yielding and stability in safflower breeding programs for highland cold rainfed areas of Iran. © 2017 NAAS (National Academy of Agricultural Sciences).