Rice yield is a complex polygenic trait influenced by intricate interactions among component characters. In this study, integrative quantitative genetics and multivariate models were used to dissect yield architecture and trait networks, and identify the promising genotypes among 21 rice genotypes that were tested in a randomized block design with three replicates. Analysis of variance revealed highly significant genotypic differences (p < 0.001) for most agronomic, yield, and grain-quality traits. Genetic variability analysis revealed high genotypic and phenotypic coefficients of variation, together with high heritability and genetic advance, for grain density per panicle, number of grains per panicle, biomass, flag leaf area, tillering ability, harvest index, and grain yield, indicating considerable genetic potential for genetic improvement. Pearson’s correlation and path coefficient analyses revealed that biomass and number of grains per panicle were the major determinants of grain yield per hill, with biomass exhibiting the strongest positive association (r = 0.678) and the largest direct effect. Principal component analysis indicated that the first two principal components explained 51.4% of the total phenotypic variation, with yield components, biomass, tillering traits, and grain-quality attributes contributing most strongly to genotype differentiation. Hierarchical cluster analysis grouped the genotypes into four distinct clusters, revealing substantial phenotypic divergence and valuable parental combinations for hybridization. The multi-trait selection index identified VAR16 as the most promising genotype, followed by VAR1, VAR17, VAR18, and VAR12, owing to their desirable combination of high grain yield and superior grain quality. Overall, this study offers a robust foundation for ideotype breeding and parental selection to enhance rice productivity and grain quality under subtropical conditions.
Manikala et al. (Fri,) studied this question.