Estimação de componentes de co-variância para pesos corporais do nascimento aos 365 dias de idade de bovinos Guzerá empregando-se modelos de regressão aleatória
Carregando...
Data
2009-01-01
Autores
Pelicioni, Luciele Cristina [UNESP]
Albuquerque, Lucia Galvão de [UNESP]
Queiroz, Sandra Aidar de [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Sociedade Brasileira de Zootecnia
Resumo
Um total de 19.770 pesos corporais de bovinos Guzerá, do nascimento aos 365 dias de idade, pertencentes ao banco de dados da Associação Brasileira dos Criadores de Zebu (ABCZ) foi analisado com os objetivos de comparar diferentes estruturas de variâncias residuais, considerando 1, 18, 28 e 53 classes residuais e funções de variância de ordens quadrática a quíntica; e estimar funções de co-variância de diferentes ordens para os efeitos genético aditivo direto, genético materno, de ambiente permanente de animal e de mãe e parâmetros genéticos para os pesos corporais usando modelos de regressão aleatória. Os efeitos aleatórios foram modelados por regressões polinomiais em escala de Legendre com ordens variando de linear a quártica. Os modelos foram comparados pelo teste de razão de verossimilhança e pelos critérios de Informação de Akaike e de Informação Bayesiano de Schwarz. O modelo com 18 classes heterogêneas foi o que melhor se ajustou às variâncias residuais, de acordo com os testes estatísticos, porém, o modelo com função de variância de quinta ordem também mostrou-se apropriado. Os valores de herdabilidade direta estimados foram maiores que os encontrados na literatura, variando de 0,04 a 0,53, mas seguiram a mesma tendência dos estimados pelas análises unicaracterísticas. A seleção para peso em qualquer idade melhoraria o peso em todas as idades no intervalo estudado.
A total of 19,770 body weight records of Guzera cattle, measured from birth to 365 days of age and supplied by the Brazilian Zebu Breeders Association, was analyzed with the following objectives of: 1) to compare different residual variances through step functions with 1, 18, 28 and 53 classes and through variance functions with orders ranging from two to five using ordinary polynomials and 2) to estimate covariance functions considering different orders for direct additive genetic effects, animal permanent environmental and maternal permanent environmental effects, using random regression models. The random effects included were modeled by regression on Legendre Polynomials with orders ranging from linear to quartic. The models were compared through the likelihood ratio test, Akaike's information criterion and the Schwarz's Bayesian information criterion. The model with 18 heterogeneous classes was the one that best fitted the residual variances, according to the statistical tests; however, the model with variance function of 5th order also showed to be appropriate. The direct heritability estimates were higher than those found in literature, ranging from 0.04 to 0.53, showing similar trends when compared to those estimated using univariate models. Selection on body weight in any age should improve the body weight in all ages in the studied interval.
A total of 19,770 body weight records of Guzera cattle, measured from birth to 365 days of age and supplied by the Brazilian Zebu Breeders Association, was analyzed with the following objectives of: 1) to compare different residual variances through step functions with 1, 18, 28 and 53 classes and through variance functions with orders ranging from two to five using ordinary polynomials and 2) to estimate covariance functions considering different orders for direct additive genetic effects, animal permanent environmental and maternal permanent environmental effects, using random regression models. The random effects included were modeled by regression on Legendre Polynomials with orders ranging from linear to quartic. The models were compared through the likelihood ratio test, Akaike's information criterion and the Schwarz's Bayesian information criterion. The model with 18 heterogeneous classes was the one that best fitted the residual variances, according to the statistical tests; however, the model with variance function of 5th order also showed to be appropriate. The direct heritability estimates were higher than those found in literature, ranging from 0.04 to 0.53, showing similar trends when compared to those estimated using univariate models. Selection on body weight in any age should improve the body weight in all ages in the studied interval.
Descrição
Palavras-chave
covariance function, Orthogonal polynomials, residual variance classes, classes de variâncias residuais, Função de covariância, Polinômios ortogonais
Como citar
Revista Brasileira de Zootecnia. Sociedade Brasileira de Zootecnia, v. 38, n. 1, p. 50-60, 2009.