Bayesian approach to estimate genetic parameters and selection of sweet potato half-sib progenies

Abstract The selection of progenies in breeding programs can generate great advances when associated with Bayesian inference because it allows the incorporation of a priori knowledge. Selection at self plant level is advantageous when evaluating half-sib progenies. However, it becomes very difficult for sweet potato cultivation are expected to support the improvement of sweet potato populations. In this context, BLUPIS (best linear unbiased prediction individual simulated) becomes an good alternative technique. Therefore, this study aimed to use the a priori knowledge obtained in previous sweet potato experiments through Bayesian inference to estimate genetic parameters and gains from selection and afterwards to choose the better half-sib progenies considering the BLUPIS. Sixteen progenies were evaluated for root and branch yield, root shape, and resistance to soil insects. The data were analyzed using Bayesian theory, considering data from 12 previous experiments to obtain the informative a priori. All variables tested, as total root yield, commercial root yield, branch green mass yield, average weight of commercial roots, root shape e resistance to soil insects showed high values for coefficient of heritability. Expressive gains are expected to support the improvement of sweet potato populations. This applied methodology, will be allowed breeders to re-design and select the most promising progenies during all breeding process for improve well determined and specific traits in sweet potato.