Genetic Diversity Analysis in Rice Maintainer Lines Using K-means Clustering for Yield and Yield Attributes

Abstract

The experiment was conducted during the kharif season (June–October, 2016) at the Regional Agricultural Research Station, Jagtial, Telangana, India to evaluate 40 maintainer lines for genetic diversity using K-means cluster analysis for grain yield and its yield attributes of rice. Multivariate analysis was performed in Randomized Block Design (RBD) with two replications to comprehend the nature and magnitude of genetic divergence among the 40 maintainer lines of rice. Based on the K-means cluster analysis, the optimal number of clusters formed was seven. The highest number of genotypes were grouped in cluster I (27), followed by cluster IV (7). Intra-cluster distances varied among the clusters with the lowest recorded in cluster I (8.64) followed by cluster III (33.64) and cluster IV (75.69), while inter-cluster distances were maximum between cluster II and VI (142.32) followed by cluster VI and VII (113.42) indicating greater genetic diversity between certain clusters. This diversity suggests potential for maximizing heterosis through hybridization between genotypes from the most divergent clusters. Cluster mean analysis revealed significant differences in traits among clusters with some clusters exhibiting consistently higher or lower values for specific characteristics. Kernel length, days to flowering, hulling percentage, and plant height are major contributors to genetic divergence. Certain maintainer lines, namely, B18, CMS46B, B42, and B88 from different clusters with suitable trait means, can be selected for use in crossing programs to create desirable variations in the breeding material for yield improvement.