Implicações da interação genótipos por ambientes na seleção e otimização de recursos no melhoramento do tomateiro

• Ano de defesa: 2021
• Autor(a) principal: Pinto, Paulo Henrique Crosara
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Genética e Melhoramento de Plantas; UFLA; brasil; Departamento de Biologia
• Palavras-chave em Português: Ensaios multiambientais; Tomateiro - Melhoramento genético; Predição de híbridos não testados; Biplot; Mega-ambientes; Multi-environment trials; Tomato - Genetic improvement; Prediction of untested hybrids; Mega-environments; Melhoramento Vegetal
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/48441

Multi-environment trials (MET) play an important role to support decisions about the selection and recommendation of genotypes and represent a large part of the expenses in a tomato breeding program, which makes it relevant to evaluate the impact of the genotype by environment interaction (GxE) and the search for the optimization and judicious application of resources through predictive methods. Thus, the objective of this work was to evaluate the GxE interaction from MET in tomato and to evaluate the interrelationship of environments and the prediction of the performance of hybrids in untested environments for resource optimization purposes. For this, we used MET data carried out in the agricultural years of 2017, 2018, 2019 and 2020 in the main tomato producing regions of Brazil by the research team of the company Syngenta. Fifty-two different hybrids were evaluated in a total of 39 experiments regarding the traits fruit yield (kg/plant), fruit classification (3A, 2A and 1A) and average fruit weight (kg). The analyses were carried out using the mixed model approach, with the prediction of untested hybrids using the compound symmetry genetic covariance structure chosen using the Schwarz information criterion (BIC) and study of the representativeness and formation of mega-environments via GGEBiplot method. There was an effect of GxE interaction for all characters measured. The GxE was predominantly non-crossover type for most traits. The accuracy of hybrid predictions in untested environments was moderate for yield, in which the mean genetic correlation across environments was greater than 0.72. The Holambra environment showed to be representative for all traits under study.