Multivariate Principal Component Analysis and Clustering Methods for Assessing Genetic Diversity in Bread Wheat (Triticum aestivum L.) Genotypes

The present experiment was conducted during rabi season (November, 2019-May, 2020) at Agricultural Research Farm of Banaras Hindu University, Varanasi, Uttar Pradesh, India to assess the genetic diversity and phenotypic variation among 50 bread wheat accessions. The genotypes were grown in a randomized complete block design with three replications, and data was collected on 14 quantitative traits. The multivariate techniques, including Principal Component Analysis (PCA) and K-means clustering were employed to identify traits contributing the most to phenotypic variation and to classify the genotypes into distinct groups based on their characteristics. PCA identified 13 principal components, with the first five explaining 65.2% of the total variation. The first two principal components accounted for 37% of the total phenotypic variation, with grain yield plant^-1, thousand seed weight, harvest index, and canopy temperature as major contributors to PC1, while total biomass and biomass contributed primarily to PC2. K-means clustering grouped the genotypes into five distinct clusters, with clusters 2 and 5 having the highest number of genotypes (14 each). Cluster 1 exhibited the highest mean values for germination percentage, days to maturity, and vegetative index. Hierarchical clustering further confirmed the genetic diversity, delineating five distinct clusters using Ward’s method with Euclidean distance. This study provides valuable insights for breeders aiming to enhance traits such as yield and other morphological characteristics through heterosis and transgressive breeding.