Prediction of genetic values across breeds in beef cattle

Records of birth (BW), weaning (WW), and yearling (YW) weight on F$\sb1$ calves produced in a topcross experiment involving Angus, Hereford, Polled Hereford, Charolais, Limousin, Simmental, Gelbvieh, Maine-Anjou, Chianina, Tarentaise, Shorthorn and Salers bulls mated to Hereford and Angus cows were used in this study Regressions (b) of crossbred calf performance on expected progeny differences (EPD) of the sires were estimated and used to adjust breed of sire means for genetic trend and sire sampling. Average (b) (kg/kg) of BW, WW, and YW of F$\sb1$ calves on EPD of the sire were 1.04 $\pm$.10,.88 $\pm$.11, and 1.40 $\pm$.11, respectively. The b were similar to the expected values of 1.0 except for YW. Adjusted breed of sire means tended to be regressed toward the average of all breeds, particularly for BW and WW. Correlations between the expression of genes from sites in purebred (P) and crossbred (C) progeny (r$\sb{\rm PC}$), and in Hereford (H)- and Angus (A)-cross F$\sb1$ calves (r$\sb{\rm HA}$) were estimated treating performance of F$\sb1$ calves in P and C, or in H and A as separate traits. A multivariate animal model with birth year-cow age-sex subclasses, random correlated direct and maternal additive genetic effects and maternal permanent environmental effects was used by breed of sire. Estimates of r$\sb{\rm PC}$ ranged from.88 to.97,.55 to 94, and.68 to.86 for BW, WW and YW, respectively. Estimates of r$\sb{\rm HA}$ ranged from.43 to.99,.56 to.95, and.50 to.98 for BW, WW and YW, respectively. Weighted averages of estimates of r$\sb{\rm PC}$ and r$\sb{\rm HA}$ across sire breeds were, respectively,.93 and.85 for BW,.77 and.73 for WW, and.76 and.86 for YW. Predictions of breeding values (PBV) of sires within and across breeds were obtained and the effect of accounting for heterogeneous variances across breed groups was evaluated. Two models differing in their covariance structure were compared. Model I assumed homogeneous variances. Model II accounted for heterogeneous variances. The PBV were obtained for when sires were to be mated to H, A or equally to H and A cows. Sires were ranked based on PBV for each model, and means of selected sires were calculated based on Model II. Effect of heterogeneous variances was most important for YW, and most important when sires are to be mated to H rather than A cows. Gains of 3 to 8 kg of YW may be achieved in the following calf crop if sires are chosen across breeds based on PBV from a model accounting for heterogeneous variances