Seleção de genótipos superiores de batata-doce para a alimentação humana via modelos mistos

Abstract

Sweet potato is a high genetic variability specie that allows the selection of genotypes for different aptitudes, such as human and animal nutrition, ethanol production and industrial. In this sense, there are different studies aimed to increasing productivity and improving qualitative and nutritional characteristics. However, the genetic improvement of a given crop is a constant process, and due to the great genetic variability of sweet potatoes, it is possible to select superior genotyoes, especially by applying more efficient statistical methodologies. Thus, the aim of this study was to evaluate sweet potato genotypes, using mixed models, aiming the superior genotypes selection for human consumption, as well as predicting selection gains and evaluating the genotypic correlations among characters. The work was carried out at the Technology Development and Transfer Center (CDTT-UFLA) from April to October 2019. A design derived from a partially balanced triple lattice was used, with three replications. As treatments were evaluated 92 sweet potato genotypes selected in the first selection cycle, plus eight commercial genotypes, Brazlândia Roxa, Princesa, Uruguaiana, BRS Amélia and Beauregard, and the pre-evaluated genotypes UFVJM-57, UFVJM-58 and UFVJM-61. The treatments were ranked by the mixed linear model via REML/BLUP, highlighting the 15 best genotypes in each agronomic trait, and, subsequently, the best ones together, considering the quantitative and qualitative characteristics. High heritabilities were found for the characteristics of part area productivity (65.89%), total root productivity (93.74%), commercial root productivity (70.54%), total dry mass productivity (74.51%) and shoot dry matter productivity (52.46%). There are considerable gains with the selection for aerial part productivity, total root productivity and commercial root productivity, in the order of 41.84, 18.23 and 12.25 t ha-1, respectively. The genotypes 2018-19-464, 2018-36-807, 2018-72-1409, 2018-19-443, 2018-72-1376, 2018-54-1137, 2018-72-1432, 2018-72-1428, 2018-72-1418, 2018-19-455, 2018-54-1114 and 2018-65-1249 are indicated for selection because they have good characteristics linked to the productivity and the quality of roots. The genotypes 2018-28-556 and 2018-31-666 showed dual aptitude (human and animal feed), because in addition to good characteristics of productivity and quality of roots, they also showed good productivity of aerial part and dry mass of aerial part. Among the quantitative characteristics, the main genetic correlations are observed between interdependent characteristics, highlighting the association between aerial part productivity and aerial part dry mass productivity. The commercial standard, on the other hand, is an efficient way of evaluating the productivity and quality of roots, since it correlates positively with the general format and the productivity of commercial roots.