Multivariate analyses for identification of high-yielding lines of rice based on yield and its component characters

The present study assessed 64 advanced breeding lines of rice and four checks to evaluate the selection criteria for yield improvement. Analysis of variance (ANOVA) for 15 yield-contributing attributes revealed significant genotypic differences, indicating considerable genetic variability. The genotypes NLR 3128, NLR 20002, NLR 20016, NLR 33891 and NLR 34417 performed better. Traits such as number of panicles per plant (NPP), panicle weight (PW), number of filled grains per panicle (NFGP) and number of chaffy grains per panicle (NCGP), total number of grains per panicle (TGP), straw yield (SY) and grain yield per plant (GYP) recorded high estimates of phenotypic coefficient of variation (PCV) and genotypic coefficient of variation (GCV), heritability and genetic advance suggesting their potential for genetic improvement. Grain yield exhibited significant positive correlations with the number of tillers per plant (NTP), panicle length (PL), test weight (TW), SY and harvest index (HI), along with their direct effects, indicating the importance of these traits in selection. A composite selection index incorporating NTP, PL, 1000-grain weight and HI showed maximum efficiency. A genetic diversity study via D2 statistics grouped genotypes into 12 clusters; of these, 42 genotypes were in Cluster I, while Clusters X, XI and XII were solitary, Clusters IX and XI exhibited maximum inter-cluster divergence. The unweighted pair group method with arithmetic mean (UPGMA)-based hierarchical clustering approach classified 64 genotypes into nine clusters. Principal component analysis (PCA) revealed that the first six principal components (PCs) explain about 91.5 % of total variance. These findings highlight the presence of significant genetic diversity among genotypes for identifying key traits for targeted breeding strategies to improve rice yield potential.