PCA-BASED MULTIVARIATE APPROACH FOR SEGMENTATION OF VARIANCE IN INDIAN MUSTARD (BRASSICA JUNCEA [L] CZERN & COSS)

Abstract

The presented study assessed 27 Indian mustard (Brassica juncea [L.] Czern & Coss) genotypes for 13 quantitative and one biochemical trait by a PCA-dependent multivariate analysis, which split the total divergence into 14 accountable principal components among them. The first five PCs, which showed an Eigenvalue of more than one, contributed significantly to about 76.35% of the total divergence. Interpretation of the PCA-Biplot declared that the genotypes, namely, two and 11, mostly subsidized the overall variance, which is about 20% and 17%, respectively. The study of the bar plot of contribution percentage presented relevance with the PCA-biplot study, which again indicated the significance of genotypes two and 11 in total variance. The biplot analysis revealed that traits, NSB and SYP, contribute appreciably to the variation of the genotypes placed in the second coordinate, showing a detrimental interference with PC 1 and positive interference with PC 2. Likewise, in the third coordinate, traits, such as LMS, influenced the variance of the genotypes of that coordinate. The percentage contribution study for features in the first and second PCs revealed that characteristics, such as, LMS, PH, NSMS, BYP, NSS, and SYP participated prominently to accelerate the total variance. This research work can be a groundwork for further crop improvement featuring the studied materials.