Comparison of models to estimate genetic parameters of growth traits in dual purpose crossbred cattle.

(Co)variance components and genetic parameters for growth traits were estimated in two dual purpose herds affiliated to the GANADOBLE foundation. The herds are located in two tropical environments in Zulia State, Venezuela. Animals were from a crisscrossing program between the Holstein and Brahman breeds mainly. Method of analysis was by Mixed Models, implemented under restricted Maximum Likelihood (REML) on an animal model, by using the derivative free algorithm and the MTDFREML package. Four models were proposed for the analyses under univariate mode and applied to the traits birth weight (PN), and adjusted weights at 244, 365 and 558 days of age (P244), (P365) y (P548) respectively. A total of 11587, 4412, 5774 y 3720 records were used for PN, P244, P365 and P548 respectively. Models differed for each trait only in the random part, as well as, in the assumptions made on them: model I was a simple animal model including the direct additive genetic effect; model II included the maternal additive effect as well, which was assumed to be independent of the first; in model III was admitted a covariance between the additive genetic effects proportional to the numerator relationship matrix; model IV was similar to model III but included the permanent maternal effect. The fixed part of the model included the effects: sex of animal, breed predominance, herd-year-season of birth and parity number of the dam, the later only for PN, P244 and P365. The likelihood ratio test (PRV) was performed based on the logarithm of the likelihood function from results of models II vs. I, III vs. II and IV vs. III in order to asses the significance of the additive maternal, the covariance between the additive effects and the maternal permanent effect. The PRV showed the necessity of incorporating the additive maternal along with the direct additive genetic effects on the prediction of the breeding values for PN, P244 and P365 (P < 0.01). The covariance between the additive genetic effects was only detected on P365, originating a -0.63 +/- 0.15 correlation between them. Genetic parameter estimates for PN, P244 and P365 were, direct heritability: 0.23 +/- 0.04; 0.17 +/- 0.05 y 0.23 +/- 0.06, maternal heritability: 0.07 +/- 0.02; 0.10 +/- 0.03 and 0.24 +/- 0.06. These estimations are comparable to most of the reported under tropical environments. None of the traits showed a permanent effect (P > 0.05). For P548 estimated of direct heritability was 0.34 +/- 0.05. It can be concluded that the additive maternal effect remains until the year of age affecting the growth of this the population and should be included in the predictions of the breeding values.