Modelos multicaracterísticos na avaliação genética de tamanho de leitegada em suínos

• Autor(a) principal: Ventura, Henrique Torres
• Data de Publicação: 2012
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://locus.ufv.br/handle/123456789/1816
• topic: Leitegada; Suínos; Modelo multicaracterístico; Avaliação genética; Litter traits; Pigs; Multi-trait; Genetic evaluation; CNPQ::CIENCIAS AGRARIAS::ZOOTECNIA::GENETICA E MELHORAMENTO DOS ANIMAIS DOMESTICOS

The aim of this study was to evaluate genetically litter traits in Landrace pigs, with the use of linear and Poisson bi-trait models via Bayesian methodology, linear and Poisson tri-trait models via Bayesian methodology and multi-trait and repeatability models via restricted maximum likelihood (REML). We obtained the following traits: number of piglets at first farrowing (NLN1), number of piglets born at second farrowing (NLN2), number of piglets at third farrowing (NLN3), number of piglets weaned at first farrowing (NLD1) number of piglets weaned at second farrowing (NLD2) and number of piglets weaned at third farrowing (NLD3). The bi-trait linear and bi-trait Poisson models analyzes were performed within parities, and two traits were used in each farrowing: number of piglets born and number of piglets weaned. The tri-trait linear and tri-trait Poisson models were analyzed across three orders of farrowing for the number of piglets born and for number of piglets weaned. Multi-trait models and repeatability model were analyzed across three orders of farrowing with the same traits. For bi-trait analysis within parities was observed that the genetic correlations between the number of piglets born and number of weaned piglets were positive and high in the first, second and third farrowing, which indicates that selection for increased number of piglets born result in genetic gain in the number of piglets weaned, which is a trait of greater economic importance. Based on DIC (deviance information criterion), linear model is the most appropriate to estimate the posterior distributions of heritabilities and genetic correlations for number of piglets born and number of piglets weaned in the first three orders of farrowing because it was the best fit of the data and less degree of complexity based on the criterion used. We conclude that number of piglets born and number of piglets weaned are of low heritability and are positively correlated between then in the three farrowing orders evaluated, and that the bitrait linear mixed model is the most appropriate to evaluate the animals in this population. For tri-trait across parities analysis was observed that the heritabilities in the first, second and third orders of farrowing were unequal both for the number of piglets born, and for the number of piglets weaned. The linear model was the best fit in the data for all traits evaluated considering DIC (deviance information criterion) as the criterion for comparing models. It was concluded that the traits evaluated in this study have low heritability and should be analyzed using a tri-trait linear mixed model. For the analyzes with multi-trait and repeatability models across three orders of farrowing, via REML, there was a trend of increased additive genetic variance and heritability along the orders of farrowing. It was found that the estimate of additive genetic variance for number of piglets born and number of piglets weaned at first farrowing was the most discrepant from the values of these estimates for all other orders of farrowing, indicating that probably the expression these characteristics at the first farrowing is controlled by genes or gene combinations other than those that regulate the expression of number of piglets born, and the number of piglets weaned in the second and third farrowing. It was concluded that the multi-trait is the most suitable for genetic evaluation of number of piglets born and number of piglets weaned in the first three orders of farrowing.