Genetic trends and multivariate interrelationships for grain quality of irrigated rice genotypes

Abstract

For genetic improvement programs, researches with multivariate approaches in rice are fundamental, to define genetic trends, clusters and correlations of agronomic characters that together help selection procedures. This work aimed to reveal the agronomic performance, trends and genetic interrelationships of grain quality based on multivariate models applied to elite lines of irrigated rice. The experiment took place in the 2017/2018 harvest, held at Estação Experimental Terras Baixas (ETB), of Embrapa Clima Temperado. The study used randomized blocks design, with three replications. There were fifteen F6 lines and four control cultivars. Evaluation of intrinsic physical quality attributes with the aid of S21 grain statistical analyzer, as well as grain yield and mill yield (whole and broken grains). Performance of analysis of variance, genetic parameters and Scott Knott cluster test, linear correlation, canonical correlations, cluster analysis via generalized Mahalanobis distance, using the Toucher method, in addition to BIPLOT principal component analysis. The results showed that PH18502 and PH18701 strains presented better agronomic performance for the studied characters, by univariate analysis. The linear and canonical correlations presented demonstrate potential in the direction of selection of multiple characters and point to the possibility of indirect selection among the relevant agronomic characters for the production chain of irrigated rice.