Estudo da interação entre genótipos, locais e anos e do coeficiente de repetibilidade em etapas finais do melhoramento genético da cana-de-açúcar

Abstract

The last stages of a genetic breeding program are characterized by the conduction of multi-environmental trials (MET) aiming at the correct selection of promising genotypes in relation to characters of agronomic-industrial interest, such as stem production. Multi-environmental trials are characterized by the evaluation of a group of genotypes in different production environments in order to identify that have greater genetic stability and adaptability. Under these conditions, quantitative genetics studies, such as quantification of the coefficient of repeatability and interaction between genotypes and environments, have the potential to reveal information that can assist the breeders involved during the breeding stages. Considering three populations of sugarcane breeding, the present work aimed at quantifying the coefficient of repeatability to infer the representativeness of the data obtained for each genotype along the numbers of repetitions and / or cuts and their relation to quality experimental; and to develop a study of the triple interaction between genotypes, environments and harvest years through the generalized GGE analysis methodology for three factors context, with the aid of the Tucker3 decomposition. Among the results, the coefficient of repeatability showed a significant and negative correlation with coefficient of variation and that the coefficient of repeatability was not associated with production environments, especially by the classification presented by Prado et al. (2013), whose foundation is based on water availability and soil characteristics. For the analysis of genotypes, environments and local interaction, it was concluded that, when studying the sites, the grouping converged to the Koppen-Geiger climatic classification; and for the cut factor a contrast of selection between cane plant and harvests was detected.