GENETIC EVALUATION OF FERTILITY TRAITS OF FRIESIAN X BUNAJI DAIRY COWS USING MULTIVARIATE MILK COMPOSITION BASED MODELS

The study investigated the improvement in genetic evaluation of fertility traits of dairy cows using single trait models and multiple-trait models that combined information on milk composition and milk yield traits. Data on fertility and milk production records of 61 Friesian x Bunaji cows were used for this study. Four fertility traits; days to first insemination (DFI), days open (DO), number of insemination per conception (NIC) and non-return rate after 56 days of insemination (NRR56). The models prediction ability and stability were assessed by the coefficients of determination adjusted (R 2 -adj), root mean square error of prediction (RMSEP) and P-value of the resulted models. Based on the evaluation criteria, the models that combined milk composition traits and one or more milk yield traits showed better model stability and predictive ability than the single-trait models for all the fertility traits evaluated. In addition, the single-trait models underestimated the prediction ability of the models. The estimates of h 2 were very low for the fertility traits (0.014 to 0.035) and moderate for milk yield and milk composition traits (0.315-0.415). Moreover, there was a moderate correlation between most of the milk yield and fertility traits. These estimates of parameters indicated that the accuracy of the prediction of fertility traits would increase using multiple-trait model that include milk yield and milk composition traits. These results suggested that genetic evaluation of fertility traits would be improved using multiple-trait models that combine milk yield and composition traits