Parâmetros genéticos de macieiras estimados por meio de modelos mistos (REML/BLUP)

Abstract

One of the main goals in the genetic breeding of apple (Malus domestica Borkh) in Brazil is the selection of genotypes with a low chilling requirement. This necessity is crucial to expand the productive area into subtropical regions. Such genotypes must possess broad adaptation allowing them to produce regularly under the conditions of irregularity and high thermal amplitude typical of the dormancy period. To achieve this, the creation and maintenance of Active Germplasm Banks with ample genetic diversity are of great importance, enabling the selection of genotypes combining a low chilling requirement with high yield and quality. The objective of this study was to perform the estimation of genetic parameters using the REML/BLUP mixed model methodology based on phenotypic traits of apple trees from the breeding programme of the Instituto de Desenvolvimento Rural do Paraná, in the Palmas municipality, Paraná. Sixteen genotypes were assessed, 15 of which were previously selected as promising, with the ‘Eva’ cultivar serving as a comparison. The plants were trained under a central leader system, with four plants of each genotype, grafted onto Marubakaido/M-9 or M-7 rootstocks. The evaluated traits included: phenological stages of budding commencement, full bloom, and mature fruit; canopy volume; number of spurs and short shoots (brindilas); yield (kg plant-1); and post-harvest quality assessments (soluble solids, titratable acidity, and pulp firmness). Data from four harvests (2021/22 to 2024/25) were submitted to principal component analysis (PCA) and to the estimation of variance components, coefficients of variation, and broad-sense heritability via REML/BLUP. Most of the evaluated genotypes were early for the phenological stages of budding, flowering, and fruit ripening. Broad-sense heritability was of low magnitude for all assessed characteristics, with the exception of titratable acidity, which showed high magnitude. Accuracy presented adequate values for all assessed traits, with the exception of the number of spurs (brindilas) and total soluble solids. Upon evaluating the BLUEs (Best Linear Unbiased Estimates), the M-7 rootstock demonstrated the best adjusted means. The BLUPs (Best Linear Unbiased Predictions) indicated that the PR2.60 genotype, cultivated on the Marubakaido/M-9 rootstock, was the most notable performer for the firmness and Soluble Solids (SS) variables. The hierarchical clustering analysis demonstrated similarity between the 32 genotype/rootstock combinations, forming only two groups. In the PCA, the first two components accounted for 67% of the total variation, with pulp firmness being responsible for 44.2%. The selection of superior genotypes was calculated using two selection indices, with the Smith-Hazel index being the most suitable for use. The mixed models methodology (REML/BLUP) allowed for the indication of the best genotype and rootstock combinations for obtaining new cultivars that are more adapted and stable in the assessment environment.