Principal component and cluster analysis in grain appearance, milling and cooking quality traits in rice (Oryza sativa L.)

Principal components of rice grain quality traits

Abstract

ice grain quality characteristics such as grain appearance, milling, cooking, and eating parameters are quite complex phenotypic attributes. The rice grain quality is governed principally by the genetic makeup of the rice variety, the biochemical composition of rice grain, and the bio-physical environment where it is produced or grown. In addition to this grain quality is also affected by the moisture conditions of rice gains at the time of processing, the equipment by which rice is processed to a finished product, and to some extent how the product is stored. It is necessary to improve grain quality traits in rice to meet the consumers' demand for premium quality rice, reduce loss at the time of milling, increase production efficiency by increasing total head rice yield, and give more profit to producers and traders of the rice supply chain.  This investigation aimed to evaluate the extent of diversity present in the population comprising one hundred-five genotypes for fifteen-grain quality attributes through principal component analysis and cluster analysis. The principal component analysis identified five principal components, (PCs) accounting for 73.10% of the sum total of variations perceived between the rice genotypes of the population. From the biplot analysis of PC1 and PC2, as the two axes of a 2-dimensional coordinate plain, it was found that the grain quality traits, kernel L/B ratio, water uptake, gel consistency, head rice recovery, kernel length after cooking, thousand-grain weight, milling percentage, and kernel length have a greater contribution to the total diversity of the rice population. By using UPGMA cluster analysis, all 105 genotypes of the population were classified into five discrete groups or clusters. The observed inter-cluster distance was largest between the clusters of cluster I and cluster III (149.22), the next largest inter-cluster distance was between cluster II and cluster IV (128.35), and that of cluster III and cluster IV was (109.61) which indicated significant genetic divergence among the genotypes. A larger inter-cluster distance between the clusters suggests the possibility of harnessing heterosis for the quality traits under study as they possess higher mean values for quality attributes with corresponding greater diversity.