Principal component analysis for selection of superior maize genotypes

Abstract

Constant advances in studies on the behavior of maize genotypes and their interactions with the environment are of great importance for the best performance of the plant. This study verifies effects and causes of agronomic variables of maize hybrids on grain yields and performs the indirect selection of superior genotypes by principal component analysis (PCA). Two hundred and thirty maize genotypes were used, with two hundred and twenty-         -nine topcross hybrids (consisting of crossings of two hundred and twenty-nine partially inbred genotypes with a tester) and one check in a randomized block design with two repetitions. The genotypes were evaluated during the 2016 and 2016/2017 crops considering the agronomic variables plant height, ear insertion height, ear position, lodging, breakage, and grain yield. Data were submitted to analysis of variance and means were compared by the Scott-Knott test (p<0.05) with subsequent multivariate exploratory analysis by PCA. In the principal component analysis, components explained 52.07% and 55.69% of the variance contained in the original variables for the 2016 and 2016/2017 crops, respectively. The variable that was most significant in both crops was ear insertion height, allowing the indirect selection of more productive genotypes. Indirect selection of the most productive genotypes was also conducted through variables that contributed significantly in the principal component analysis. Thus, the use of multivariate exploratory analysis is efficient in the characterization and selection of maize genotypes evaluated in different crop seasons.