Genotype selection by REML/BLUP methodology in a segregating population from an interspecific Passiflora spp. crossing

Abstract

The restricted maximum likelihood/best linear unbiased prediction (REML/BLUP) method involves, simultaneously, the estimation of variance components and the prediction of the genetic values, considering the optimum procedure to make inferences in genetic and non-phenotypic levels. By using the mixed models methodology, this study aimed to estimate genetics parameters and genotypic values in a segregating population comprising 118 individuals from Passiflora interspecific crossing. The experiment was arranged in randomized block design with two repetitions. The traits number of fruits, level of soluble solids and pulp mass presented large genetic variability, a fact that contributed to high heritability estimations and selective accuracy, and revealed excellent possibilities to the selection and breeding of Passiflora segregating populations. By comparing the average of the 30 selected genotypes with the general average of the population, it was possible to observe higher gain estimations for the number of fruits (319.15 %) and pulp mass (73.12 %), lower gains for fruit mass (21.50), soluble solids level (13.43), longitudinal fruit diameter (6.32) and transverse fruit diameter (4.99) as well as low reductions in shell thickness (−0.0036) and length of the androgynophore (0.0056), in relation to the first ordinated individuals. The assessed population presenting genetic variability and the superior individuals for each trait were identified, thus enabling the continuity of the selective process. The analyses by the REML/BLUP methodology and by the presented model proved adequate for gain prediction, with good genetic breeding perspectives.