Seleção para aptidão milho verde em diferentes germoplasmas de milhos especiais

Abstract

Corn (Zea mays L.) holds a prominent position in global agricultural production and is widely cultivated in Brazil due to its adaptability to different soil and climatic conditions. Among the various forms of corn exploitation, green corn and supersweet corn represent promising market niches, especially for small-scale family farmers. The present study aimed to evaluate the potential of half-sibling progenies of green and supersweet corn for the production of commercial ears for in natura consumption through recurrent selection and multivariate techniques. In the green corn experiment, 88 half-sibling progenies and two commercial controls, BRS3046® and Quindim®, were evaluated. In the supersweet corn experiment, 178 half-sibling progenies and the commercial controls Isla® and Confeito® were assessed. Both experiments were conducted using a randomized block design with three replications. Nine variables related to agronomic adaptation and ear yield potential were evaluated. Genetic parameters, broad-sense heritability, expected selection gain, and the improved population mean were estimated. Genetic divergence among genotypes was assessed using Mahalanobis’ generalized squared distance, and genotypes were grouped based on genetic dissimilarity using the UPGMA and Tocher Optimization methods. Additionally, the evaluated variables underwent principal component analysis (PCA). Multivariate statistical analyses allowed for the characterization of genetic variability and the identification of superior genotypes. The results indicated the presence of favorable genetic variability among the progenies, enabling the selection of genotypes with greater aptitude for in natura green and supersweet corn production. The UPGMA and Tocher Optimization grouping methods were efficient in forming groups and showed similar patterns, despite some differences in genotype allocation. The Tocher method provided a more detailed organization, with a greater number of distinct groups for supersweet corn (nine groups and six isolated progenies) compared to green corn (five groups and five isolated genotypes). Supersweet corn germplasm was more sensitive to the grouping methods, revealing greater genetic variability. Principal component analysis (PCA) reduced the set of variables, with the first three components explaining 75.21% (green corn) and 76.82% (supersweet corn) of phenotypic variance. In both germplasms, the variables related to ear yield were the most influential in forming PC1. PCA enabled the identification of the best-performing genotypes, leading to the selection of 26 half-sibling progenies of green corn and 53 progenies of supersweet corn. The estimation of expected selection gain also showed satisfactory results, especially for variables related to the productive potential of green corn ears. The results highlight highly favorable conditions for the special corn breeding program at UEPG, potentially enabling the future development of open-pollinated varieties with greater aptitude and productive potential for in natura green corn.