Metodologias de análise dialélica como critério para a seleção de genitores de trigo

Abstract

By conducting a program of plant breeding, it’s questioned “what’s the best strategy for selecting genitors to obtain segregating populations with a greater number of superior recombinants”. Knowledge which will contribute to the improvement of this process is important. Thus, this study aimed to estimate the general combining abilities (GCA) and specific combining ability (SCA) through traditional and innovative methodologies in order to maximize the efficiency of parental selection and hybrid combinations of wheat. In the agricultural season of 2010, there were six hybridizations between wheat genotypes (CD 150, CD 113, CD 117, BRS Pardela, Fundacep Cristalino e IPR 85), based on the model complete diallel without reciprocals. In crop year 2011, the 15 populations F2 and six genitors were evaluated in experimental design of randomized complete block design with three replications. Each plot consisted of five rows of three meters in length, spaced 0.20 meters between rows and a density of 300 seeds per square meter. The fertilizer was 350 kg ha-1 NPK (5-20-20) and 60 kg ha-1 nitrogen applied at the beginning of tillering. The GCA and SCA were measured by four methods: a) method 2, model B of Griffing (1956) b) combining ability analysis of multivariate c) REML / BLUP and d) GGE Biplot. The methodologies diallel Griffing (1956), REML / BLUP and GGE Biplot corroborate each other, in most cases, the indication of the GCA and SCA of hybrid combinations evaluated. For most traits, estimates of GCA and SCA obtained by Griffing (1956) and REML / BLUP were positively associated with the average observed and predicted. The multivariate diallel analysis was effective in parents and crosses indicate promising, according to the general ability and specific combining, measured jointly for all the traits. The joint use of tested methodologies enabled a safer indication of the best parents and hybrid combinations. The results of the pioneered methodology by Griffing (1956) were complemented by RELM / BLUP, which gave the results in terms of genotypic values, and multivariate analysis, useful information that allowed the simultaneous selection of characters. Additionally, the graphical analysis using the methodology GGE Biplot generated summary information and allowed a better interpretation of the comparative performance of parental lines and hybrid combinations.