A Comparison of Linear Models for Estimating Co-Variance Components and Genetic Parameters in Holstein Dairy Cattle

The objective of this study was to compare four different statistical models for genetic evaluation of sometraits of Holstein-Friesian dairy cattle raised in Egypt. Data were collected from Alexandria Copenhagen Company; acommercial dairy herd located in Egypt on Cairo-Alexandria desert road, and represented 2846 first three lactationrecords pooled from cows having 60 sires and 428 dams. The studied traits were; days open (DO), 305-days milk yield(305-DMY), fat yield (FY) and protein yield (PY). Models were discriminated according to random effects fitted ineach model. The random effects were; direct additive genetic effects of animals, maternal additive genetic effects,permanent environmental effects, together with the covariance between direct and maternal genetic effects, andresiduals. Comparisons of statistical models were based on (AG) Log Likelihood values and estimates of geneticparameters of traits. Co-variance components and genetic parameters were estimated with VCE-6 software package.Heritabilities obtained from all models were ranged from (0.07 to 0.10), (0.24 to 0.32), (0.25 to 0.42) and (0.24 to 0.33)for DO, 305-DMY, FY and PY, respectively. Also, for all traits, the best-fitted model was characterized by the highestLog Likelihood value, the highest maternal heritability and the existence of direct-maternal genetic covariance.Estimated breeding values (EBVs) showed high variations with positive spearman’s rank correlations (≥ 0.83) amongmodels. This study showed that the inclusion of maternal effects with direct-maternal genetic covariances in thestatistical models for genetic evaluations would improve the current herd genetically.