Estimativas de parâmetros genéticos para produção de leite e persistência da lactação em vacas Gir, aplicando modelos de regressão aleatória
Carregando...
Data
2008-09-01
Autores
Herrera, Luis Gabriel González [UNESP]
El Faro, Lenira
Albuquerque, Lucia Galvão de [UNESP]
Humberto, Tonhati [UNESP]
Machado, Carlos Henrique Cavallari
Título da Revista
ISSN da Revista
Título de Volume
Editor
Sociedade Brasileira de Zootecnia
Resumo
Com o objetivo de estimar parâmetros genéticos para a produção de leite no dia do controle (PLDC), foram usadas as 2.440 primeiras lactações de vacas da raça Gir leiteira, com partos registrados entre 1990 e 2005. As PLDC foram consideradas em dez classes mensais e analisadas por meio de modelo de regressão aleatória (MRA) utilizando-se como efeitos aleatórios o genético-aditivo, o de ambiente permanente e o residual e, como efeitos fixos, o grupo de contemporâneos (GC), a co-variável idade da vaca ao parto (efeito linear e quadrático) e a curva média de lactação da população. Os efeitos genético-aditivos e de ambiente permanente foram modelados utilizando-se as funções de Wilmink (WIL) e Ali e Schaeffer (AS). As variâncias residuais foram modeladas utilizando-se 1, 4, 6 ou 10 classes. Os grupos de contemporâneos foram definidos como rebanho-ano-estação do controle contendo no mínimo três animais. Os testes indicaram que o modelo com quatro classes de variâncias usando a função paramétrica AS foi o que melhor se ajustou aos dados. As estimativas de herdabilidade variaram de 0,21 a 0,33 para a função AS e de 0,17 a 0,30 para WIL e foram maiores na primeira metade da lactação. As correlações genéticas entre as PLDC foram positivas e elevadas entre os controles adjacentes e diminuiram quando a distância entre os controles aumentou. Para o melhor modelo, foram estimados os valores genéticos para a produção de leite acumulada até os 305 dias e, para períodos parciais da lactação, foram obtidas como médias dos valores genéticos preditos naquele período. Os valores genéticos foram comparados, por meio da correlação de posto, ao valor genético predito para a produção acumulada até os 305 dias, pelo método tradicional. As correlações entre os valores genéticos indicaram que podem ocorrer divergências na classificação dos animais pelos critérios estudados.
Data consisting of 2,440 first lactations of Gyr cows calving between 1990 and 2005 were used to estimate genetic parameters of monthly test-day milk yield (TDMY). Records were analyzed by random regression models (MRA) that included the additive genetic and permanent environmental random effects and the contemporary group, age of cow at calving (linear and quadratic components) and the average trend of the population as fixed effects. Random trajectories were fitted by Wilmink's (WIL) and Ali & Schaeffer's (AS) parametric functions. Residual variances were fitted by step functions with 1, 4, 6 or 10 classes. The contemporary group was defined by herd-year-season of test-day and included at least three animals. Models were compared by Akaike's and Schwarz's Bayesian (BIC) information criterion. The AS function used for modeling the additive genetic and permanent environmental effects with heterogeneous residual variances adjusted with a step function with four classes was the best fitted model. Heritability estimates ranged from 0.21 to 0.33 for the AS function and from 0.17 to 0.30 for WIL function and were larger in the first half of the lactation period. Genetic correlations between TDMY were high and positive for adjacent test-days and decreased as days between records increased. Predicted breeding values for total 305-day milk yield (MRA305) and specific periods of lactation (obtained by the mean of all breeding values in the periods) using the AS function were compared with that predicted by a standard model using accumulated 305-day milk yield (PTA305) by rank correlation. The magnitude of correlations suggested differences may be observed in ranking animals by using the different criteria which were compared in this study.
Data consisting of 2,440 first lactations of Gyr cows calving between 1990 and 2005 were used to estimate genetic parameters of monthly test-day milk yield (TDMY). Records were analyzed by random regression models (MRA) that included the additive genetic and permanent environmental random effects and the contemporary group, age of cow at calving (linear and quadratic components) and the average trend of the population as fixed effects. Random trajectories were fitted by Wilmink's (WIL) and Ali & Schaeffer's (AS) parametric functions. Residual variances were fitted by step functions with 1, 4, 6 or 10 classes. The contemporary group was defined by herd-year-season of test-day and included at least three animals. Models were compared by Akaike's and Schwarz's Bayesian (BIC) information criterion. The AS function used for modeling the additive genetic and permanent environmental effects with heterogeneous residual variances adjusted with a step function with four classes was the best fitted model. Heritability estimates ranged from 0.21 to 0.33 for the AS function and from 0.17 to 0.30 for WIL function and were larger in the first half of the lactation period. Genetic correlations between TDMY were high and positive for adjacent test-days and decreased as days between records increased. Predicted breeding values for total 305-day milk yield (MRA305) and specific periods of lactation (obtained by the mean of all breeding values in the periods) using the AS function were compared with that predicted by a standard model using accumulated 305-day milk yield (PTA305) by rank correlation. The magnitude of correlations suggested differences may be observed in ranking animals by using the different criteria which were compared in this study.
Descrição
Palavras-chave
Covariance functions, Genetic parameter, parametric function, random regression, Funções de covariância, função paramétrica, Parâmetros genéticos, Regressão aleatória
Como citar
Revista Brasileira de Zootecnia. Sociedade Brasileira de Zootecnia, v. 37, n. 9, p. 1584-1594, 2008.