
Kadambot Siddique
DirectorThe UWA Institute of Agriculture, The University of Western Australia, AUSTRALIA

Submitted
2025/03/06First Online
2026/04/25doi
10.18805/LR-5488Cite article - P. Shanthi, M. Umadevi, N. Manivannan, M. Gunasekaran, D. Sassikumar. (2026). Principal Component Analysis with Association Analysis and Clustering of Genotypes by K-means Clustering in Greengram [ Vigna radiata (L.) Wilczek]. Legume Research. 49(10): 1714–1720. doi: 10.18805/LR-5488
Background: India is recognized as the prime global producer of greengram, which is grown in nearly across the states. This legume is a remarkable source of protein, fiber and iron. The identification of superior genotypes is primarily dependent on genetic diversity. Analytical techniques, including correlation, principal component analysis (PCA) and cluster analysis, are essential for clarifying the diversity found among various genotypes.
Methods: The study was carried out over two consecutive years, 2020 and 2021, during the Kharif season at the National Pulses Research Centre (NPRC) located in Vamban, Pudukkottai, Tamil Nadu, utilizing a randomized block design.
Result: The association analysis among the five yield-contributing traits revealed that the two traits, namely plant height and total number of pods per plant, exhibited a significant positive association. Additionally, 100 seed weight demonstrated a significant positive correlation with plant height and showing a significant negative correlation with both 50 per cent flowering and the number of primary branches per plant. The principal components associated with the dataset, namely PC1 (Dim.1) and PC2 (Dim.2), are crucial in elucidating the variability within the dataset. PC1, with an eigen value of 2.3, accounted for a substantial portion of the total variance (36.27%), indicating a strong reflection of plant height. Conversely, PC2, with an eigen value of 1.75, contributed an additional 29.78 per cent to the total variance, highlighting a significant loading related to grain yield per hectare, which aligns with the observed positive correlations. A K-means cluster analysis was conducted; these genotypes were grouped into four distinct clusters. The majority of the genotypes originated from diverse parental crosses and was allocated to different clusters, suggesting a lack of correlation between genetic and geographical diversity. The high-yielding genotypes identified in cluster three, including VGG 20-010, VGG 20-066, VGG 20-069, VGG 20-068, VGG 20-071, VGG 20-088, VGG 20-091, VGG 20-092, VGG 20-230, VGG 20-232, VGG 20-233, VGG 20-234 and VBN(Gg)2 were significantly contributed to the overall diversity in terms of the number of primary branches and pods per plant, indicating their potential utility in future breeding programs.
Distinguished scholars and subject experts committed to editorial excellence, scientific integrity and rigorous peer review.
View Full Board