[(CO) VARIANCE COMPONENTS OF WEANING WEIGHT IN CUBAN ZEBU BERMEJO CATTLE

Abstract

The objective of this study was to estimate the (co) variance components of weaning weight in the Cuban bright reddish Zebu cattle a sample of 5120 records of male and female animals born between 1988 and 2007 was taken from three herds located in the provinces of Cienfuegos, Camagüey and Granma. Six different statistical models were used which had in common as fixed effects the contemporary groups (year-herd-season-sex) and the covariables, linear weaning age and mother’s age at linear and quadratic calving. Model 1 only included the random effect of the animal, model 2 also included the random effect of permanent environment, model 3 was equal to 1 but also including the random maternal genetic effect, regardless of the existence of covariance between the additive and maternal effects, model 4 was equal to 3 but considering the existence of covariance between additive and maternal effects, models 5 and 6 were the same as models 3 and 4, respectively, but adding the random effect of permanent environment. Weaning weight averaged 159.7 kg, with a standard deviation of 26.7 kg. Three criteria to compare models were used the log-likelihood (log L), Akaike information criterion (AIC) and Bayesian information criterion (BIC). Model 1 was that of worst fit considering the three criteria. Taking into account the criterion log L models 5 and 6 did not differ significantly between them and were the best compared to the rest. AIC and BIC criteria indicate the model 5 as the best. Heritability for the direct additive effect (0.32) in modelo1 was higher than that estimated for the other models (0.02 to 0.06). Heritabilities estimated for maternal effects in models of best fit reached values of 0.06 (model 5) and 0.04 (model 6), which although higher than those of the direct effect are also low. Permanent environmental effects determined between 10.9% (model 6) and 15.4% (model 3) of the total variability. The genetic correlation between direct and maternal effects reached values of 0.53 and 0.61 for models 4 and 6, respectively. The results indicate that there is little possibility of improving the weaning weight by selection in the present population.