Regressão aleatória na avaliação da produção de leite de búfalas Murrah utilizando inferência bayesiana

• Ano de defesa: 2013
• Autor(a) principal: Brito, Lais Costa
• Orientador(a): Euclydes, Ricardo Frederico
• Banca de defesa: Silva, Fabyano Fonseca e , Lopes, Paulo Sávio
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Mestrado em Zootecnia
• Departamento: Genética e Melhoramento de Animais Domésticos; Nutrição e Alimentação Animal; Pastagens e Forragicul
• País: BR
• Palavras-chave em Português: Avaliação genética; Caprinos leiteiros; Curva de lactação
• Palavras-chave em Inglês: Genetic evaluation; Dairy goats; Lactation curve
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::ZOOTECNIA::GENETICA E MELHORAMENTO DOS ANIMAIS DOMESTICOS
• Link de acesso: http://locus.ufv.br/handle/123456789/5767

Considering the importance of raising buffaloes to the national dairy farming, the objective of this study was to model variations in test-day milk yield (TDMY) during the first lactation of Murrah buffaloes by random regression models (RRM), using orthogonal Legendre polynomials (POL) and linear B-splines (BSP). A total of 15,161 test-day records from 1,158 first lactations of Murrah buffaloes breed from herds belonging to São Paulo state, Brazil, with calving recorded from 1988 to 2005, were analyzed. For TDMY, 38 weekly classes of lactation days were considered and analyzed using a single-trait RRM, including the additive genetic, permanent environmental and residual random effects. In addition, contemporary group, linear and quadratic effects of the covariate age of the buffalo at birth, number of milkings and mean lactation curve of the population, modeled using a fourth-order orthogonal Legendre polynomials, were included as fixed effects. Covariance functions of additive genetic and permanent environmental regression coefficients were equally modeled by POL (3th to 6th order) or linear BPL (4 to 6 knots). Residual variances were modeled considering 9 classes uniformly spaced. The (co)variance components of parameters were estimated by Bayesian inference, using the GIBBS3F90 software. The convergence of Gibbs chains was verified by graphic inspection and using BOA package from the R program. The post-gibbs analyses were performed using the POSTGIBBS1F90 software. The selection of the models was based on deviance information criterion (DIC). The average TDMY was 6.31±2.13 kg. The sixth order POL and six-knot BSP (knots at 8, 29, 57, 120, 239 and 267 days), with highest degree of complexity (51 parameters) were considered as the most suitable to describe the TDMY variation. The genetic parameters presented similar tendencies for both models, excepted for genetic variances and heritability, which showed low estimates at first lactation weeks for model BSP6. It was possible realize that, despite the low correlation between the predicted breeding value for 270-day milk yield and lactation persistency, the best models classified the bulls similarly. The results indicate that linear B-spline function with six knots was the most appropriate RRM to genetic evaluation of TDMY of Murrah buffaloes, given that parcimonious models with higher computational advantages to model variations of TDMY are most sought, and given the better control of heritability estimates of the edges of the lactation curve.