Alternativas de avaliação genética de tourinhos nelore em provas de ganho em peso a pasto

• Ano de defesa: 2012
• Autor(a) principal: Breno de Oliveira Fragomeni
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Quadrados mínimos; Correlação; Inferência bayesiana; Seleção; Modelos mistos; Herdabilidade; Zootecnia
• Link de acesso: http://hdl.handle.net/1843/BUOS-8UBG7T

This work aimed to study different methods of genetic evaluation of Nelore young bulls coming from pasture performance tests with the objective of evaluating which method is the most adequate to proceed the selection. The traits analyzed were body weight fitted to the age of 550 days (W550), average daily gain (ADG) and a index that pondered the deviation of ADG by 0.6 and the deviation of P550 by 0.4. The information used came from 3,796 Nellore young bulls kept in exclusive pasture regimen. The animals were participants of 43 pasture performance tests attained by the group Provados a Pasto, that occurred between 1997 and 2008 in commercial farms in the state of Goiás. First of all, the least squares method and the mixed models were compared to evaluate and rank the animals. The heritability values were 0.73, 0.31 and 0.44 for P550, ADG and index, respectively. The young bulls was classified according to the residuals in the least squares methods and the predicted breeding values in the mixed models. Despite the values of the correlation obtained between both selection alternatives were high, there were differences in the number of animal selected in common when each method was used. The accuracy values in the mixed models were greater than those obtained with the least squares method. In this way, the mixed models have shown itself as a more appropriate method than the least squares, and its usage should be implemented in performance tests to predict breeding values with bigger accuracy. After this, by Gibbs sampler, variance components for the same traits were estimated, in a multitrait animal models. The posteriori means (higher posterior density at 95%) for the estimated heritability was 0.74 (to 0.55 from 0.93), 0.34 (to 0.22 from 0.47) and 0.45 (to 0.29 from 0.59) for P550, ADG and index respectively. The posteriori means (higher posterior density at 95%) for the genetic correlation was 0.74 (to 0.60 for 0.87) between W550 and ADG, 0.84 (to 0.75 from 0.92) between W550 e index and 0.97 (to 0.95 for 0.99) between ADG and index. The selection to increase W550 value has provided greater correlated responses in ADG and index than the direct selection for these traits. So, W550 could be considered the best selection criteria in pasture performance tests.