Comparação entre modelos de análise genômica utilizando dados simulados e dados reais em ovinos

• Ano de defesa: 2015
• Autor(a) principal: Pires, Michele Porto [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Ovino; Genômica; Melhoramento genetico; Métodos estatísticos; Genetica animal; Genomics
• Link de acesso: http://hdl.handle.net/11449/135987; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/15-02-2016/000858097.pdf

In Chapter 2 the objective of the work was to evaluate the quality of the genetic merit of predictions between BLUP and GBLUP methods in the presence of uncertainty in the data. In chapter 3 the objective was to compare the BLUP methods, GBLUP and BAYESR considering whether or not to genotypeenvironment interaction in multi-racial sheep. The data used for the realization of Chapter 2 were simulated with three levels of inaccuracy in phenotypic data, 0%, 25% and 50% in three features with magnitude of heritability of 0.02, 0.15 and 0.30, respectively, the BLUP and GBLUP methodologies were applied this data and their performances were compared. To compare the methodologies the mean square error parameters, prediction error variance, bias, accuracy and selection efficiency were used. Better results were obtained by BLUP method over the GBLUP method. The GBLUP method showed no advantage over the BLUP method, except when the magnitude of heritability was 0.02 and there was no inaccuracy in phenotypic data. For chapter 3 we were used 4,288 phenotypes and genotypes dual-core herds of information (Information Nucleus Flock) coming from the Australian Sheep Cooperative Research Center program (CRC). To access the accuracy of the estimates, cross-validation was used. Methods BLUP, GBLUP and BAYESR, were used to estimate breeding values under two models, univariate, which does not consider the genotype-environment interaction and the bivariate model with the interaction effect is considered. The accuracies of the univariate models ranged from 0.17 to 0.27 for the environment 1 and 0.19 to 0.20 for the environment 2. The accuracies of bivariate models ranging from 0.14 to 0.23 for the environment 1 to the environment and the accuracies were 0.19. There was no benefit from the application of genomic selection on BLUP method in environment 2 for gastrointestinal resistance feature. On the other hand, for the application of 1 atmosphere genomic ...