Inferência bayesiana e modelos mistos na cultura de milho : aplicações em dialelos, adaptabilidade e estabilidade e ganhos por seleção

• Autor(a) principal: Rodovalho, Marcos de Araujo
• Data de Publicação: 2012
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Estadual de Maringá (RI-UEM)
• Texto Completo: http://repositorio.uem.br:8080/jspui/handle/1/1151
• topic: Milho; Melhoramento genético; Genética quantitativa; Teoria Bayesiana; Agricultura; Brasil.; Corn; Genetical enhancement; Quantitative genetics; Bayesian theory; Agriculture; Brazil.; Ciências Agrárias; Agronomia

This study aimed to apply the Bayesian inference and mixed models in diallel, adaptability and stability and selection gains in corn. The Bayesian method has been implemented through the Chains of Independence (Independence Chain) algorithm, with the informative prior information for the components of family and residual variance (Inverse Gamma prior distribution).Estimates of genetic parameters of capacity expansion and yield of popcorn in an essay of 169 half-sib families, through the Bayesian methodology as a predictor of genetic merit, demonstrated that the study population, has significant genetic variability for cycles subsequent improvement and that the Bayesian methodology can assist in the selection of families popcorn process. For estimates of general and specific combining ability of the yield of ten lines of maize using diallel indicated that strains 6, 5, 9 and 10 were promising for use in the breeding program. The hybrid combinations 2x6, 1x8 and 1x5 were the ones that stood out most to yield, with the highest estimates of specific combining ability . For adaptability and stability of 43 corn simple hybrids in trials network, the methodology of mixed linear models. According to the genotypic values, the ten best hybrids, considering the interaction of genotype and environment effects medium were: 14, 32, 13, 4, 34, 6, 33, 22, 10 and 37 . It is concluded that the methodology of linear mixed models is helpful and suitable for genotypic evaluation of corn hybrids in multisite alternative.