The efficiency of fixed regression model and estimation of genetic parameters for body weight of Mehraban sheep using random regression models

Weight records of Mehraban sheep from birth to 365 days, were analyzed using fixed and random regression models. Fixed and random (direct and maternal additive genetic, direct and maternal permanent environmental) effects were modeled with orthogonal Legendre polynomials of age at recording. Models were compared by log likelihood function (LogL) and using log likelihood ratio test (LRT). Results showed that fixed regression model is not a suitable model for growth trait data analysis. Results from random regression models showed that a model with 3, 3, 4, 4 orders of fit for direct additive genetic, maternal additive genetic, animal permanent environment and maternal permanent environment covariance function, respectively are suitable. Direct heritability estimates increased after birth and were reached to its maximum in 270 days of age. Maternal heritability estimates increased after birth and reached to their maximum in 120 days of age and decreased gradually after that to 365 day. Correlation tended to decrease with increasing the period between records. Animal additive genetic correlation and maternal additive genetic between birth weights with other ages were low, and between other ages were moderate to high.