Abstract:

This study examined ways to correct for environmental effects and effectively utilize all the available performance records in Ethiopian smallholder dairy production systems. Modelling contemporary groups based on villages rather than the usual herd resulted in proper correction for environmental effects and gave higher (0.18 vs 0.12) heritability estimates for milk yield. Besides, in such production systems farmers tend to milk cows for several later lactations and performance records are a scarce resource. Thus, the use of a multiple-trait single-step genomic prediction model that considered 1st and 2nd plus later lactations as two distinct but correlated traits, gave the highest standard deviation of genomic enhanced breeding values (1.365 vs 0.840 kg/d) and better reliability of genomic prediction than univariate models.