Modelos de regressão aleatória com diferentes estruturas de variância residual para descrever o tamanho da leitegada
Carregando...
Data
2011-12-01
Autores
Cavalcante-Neto, Aderbal
Tholon, Patrícia
Lui, Jeffrey Frederico [UNESP]
Lara, Maria Aparecida Cassiano
Fonseca, Carlos
Ribeiro, Maria Norma
Sarmento, José Lindenberg Rocha
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal do Ceará (UFC)
Resumo
Objetivou-se comparar modelos de regressão aleatória com diferentes estruturas de variância residual, a fim de se buscar a melhor modelagem para a característica tamanho da leitegada ao nascer (TLN). Utilizaram-se 1.701 registros de TLN, que foram analisados por meio de modelo animal, unicaracterística, de regressão aleatória. As regressões fixa e aleatórias foram representadas por funções contínuas sobre a ordem de parto, ajustadas por polinômios ortogonais de Legendre de ordem 3. Para averiguar a melhor modelagem para a variância residual, considerou-se a heterogeneidade de variância por meio de 1 a 7 classes de variância residual. O modelo geral de análise incluiu grupo de contemporâneo como efeito fixo; os coeficientes de regressão fixa para modelar a trajetória média da população; os coeficientes de regressão aleatória do efeito genético aditivo-direto, do comum-de-leitegada e do de ambiente permanente de animal; e o efeito aleatório residual. O teste da razão de verossimilhança, o critério de informação de Akaike e o critério de informação bayesiano de Schwarz apontaram o modelo que considerou homogeneidade de variância como o que proporcionou melhor ajuste aos dados utilizados. As herdabilidades obtidas foram próximas a zero (0,002 a 0,006). O efeito de ambiente permanente foi crescente da 1ª (0,06) à 5ª (0,28) ordem, mas decrescente desse ponto até a 7ª ordem (0,18). O comum-de-leitegada apresentou valores baixos (0,01 a 0,02). A utilização de homogeneidade de variância residual foi mais adequada para modelar as variâncias associadas à característica tamanho da leitegada ao nascer nesse conjunto de dado.
The objective of this work was to compare random regression models with different residual variance structures, so as to obtain the best modeling for the trait litter size at birth (LSB) in swine. One thousand, seven hundred and one records of LSB were analyzed. LSB was analyzed by means of a random-regression, single-characteristic animal model. The fixed and random regressions were represented by continuous functions over the farrowing order, adjusted by third-order Legendre's orthogonal polynomials. To obtain the best modeling for the residual variance, variance heterogeneity was assumed by means of 1 to 7 classes of residual variance. The general analysis model included a contemporary group; the fixed regression coefficients for modeling the population's average trajectory; the random regression coefficients of the direct additive genetic effects both of the litter and of the animal's permanent environment; and the residual random effect. The likelihood-ratio test, Akaike's information criterion, and Schwarz's Bayesian information criterion appointed the model that considered variance homogeneity as being the one that provided the best adjustment to the data used. Overall, the heritabilities obtained were close to zero (0.002 to 0.006). Regarding the permanent environment proportion, different magnitudes were observed for the farrowing order: increasing from the 1st (0.06) to the 5th (0.28) orders and decreasing from there to the 7th order (0.18). The common litter effect presented low values (from 0.01 to 0.02). The use of residual variance homogeneity was more suitable for modeling variances associated to the trait litter size at birth in this data set.
The objective of this work was to compare random regression models with different residual variance structures, so as to obtain the best modeling for the trait litter size at birth (LSB) in swine. One thousand, seven hundred and one records of LSB were analyzed. LSB was analyzed by means of a random-regression, single-characteristic animal model. The fixed and random regressions were represented by continuous functions over the farrowing order, adjusted by third-order Legendre's orthogonal polynomials. To obtain the best modeling for the residual variance, variance heterogeneity was assumed by means of 1 to 7 classes of residual variance. The general analysis model included a contemporary group; the fixed regression coefficients for modeling the population's average trajectory; the random regression coefficients of the direct additive genetic effects both of the litter and of the animal's permanent environment; and the residual random effect. The likelihood-ratio test, Akaike's information criterion, and Schwarz's Bayesian information criterion appointed the model that considered variance homogeneity as being the one that provided the best adjustment to the data used. Overall, the heritabilities obtained were close to zero (0.002 to 0.006). Regarding the permanent environment proportion, different magnitudes were observed for the farrowing order: increasing from the 1st (0.06) to the 5th (0.28) orders and decreasing from there to the 7th order (0.18). The common litter effect presented low values (from 0.01 to 0.02). The use of residual variance homogeneity was more suitable for modeling variances associated to the trait litter size at birth in this data set.
Descrição
Palavras-chave
Função de covariância, Parâmetro genético, Fêmea suína, Covariance functions, Genetic parameter, Sow
Como citar
Revista Ciência Agronômica. Universidade Federal do Ceará (UFC), v. 42, n. 4, p. 1043-1050, 2011.