Prediction of genetic gains in sweet potato genotypes by polycross

Sweet potato crops present a wide genetic diversity, but genetic parameters of populations should be estimated in breeding programs, mainly when the selection is focused on multiple traits. Thus, the objective of this study was to estimate genetic parameters and select clones via a selection index in a sweet potato population obtained from uncontrolled crosses. The experiment was carried out in Ilha Solteira, Sao Paulo, Brazil, from June to November 2020, to evaluate 144 clones and three controls, totaling 147 treatments. A randomized block design was used, with two replications and three plants per plot. The data were analyzed through mixed modeling (REML/BLUP), and the selection index proposed by Mulamba & Mock was applied using a selection pressure of 17.3% for the selection of superior clones. The direct selection enabled a higher prediction of gains for each trait, but the joint analysis maximized the selection gains for all traits of interest. Clones 38, 63, 79, 77, and 5 are recommended, among the selected clones, for further stages of sweet potato breeding focused on improving traits for production and table consumption. Palavras-chave: Ipomoea batatas (L.) Lam, Variabilidade genética, Índice de seleção, Parâmetros genéticos, Ganho com seleção