Screening of Soybean (Glycine max L.) Genotypes through Multivariate Analysis

Kamrun Nahar Mili1, Bir Jahangir Shirazy2, Md. Mostofa Mahbub3


The physiological divergence was assessed in twenty-seven soybean genotypes by using principal component analysis, cluster mean analysis, principal coordinate analysis and canonical variate analysis to identify parental genotypes for the future breeding program in order to develop new high yielding varieties in a randomized complete block design with three replications. The genotypes under the experiment were grouped into five clusters. The highest number of genotypes found in cluster III. The highest intra-cluster distance was found in cluster II and while cluster V showed no intra-cluster distance values which revealed homogenous nature of the genotype within the cluster. The highest inter-cluster distance was found between cluster I and IV followed by I and V. Cluster II have early flowering genotypes whereas early maturity in cluster III and most of the desirable traits were found in cluster IV. Days to first flowering and pod length from cluster II, whereas pods per plant and yield per plant from cluster IV have the positive relative contribution to the entire divergence. According to principal component scores, LG-92P-1176 followed by KANH-33, AGS-79, MTD-452, GMOT-17, GC-82-332411, MTD-451 and BS-33 have the prominent influence towards varietal improvement. Selecting genotypes from distant clusters probably provide promising recombinants and better segregants for the future breeding platform.

Keywords: Cluster analysis, genotype, genetic variation, principal component analysis