Improving pedigree prediction, sire, and cow evaluation by adjusting production data for past reproductive performance
Two BLUP models were used to evaluate young Holstein bulls for production with and without adjusting records for days open. Model I included the mean, HYS, groups, sires of herdmates (fixed) and young bulls within groups (random). In model II, the same effects were used considering all bulls and sires random with adding the relationship among bulls. Data were 24,997 progenies of 484 young bulls and 119,013 herdmates progenies of 291 sires. The maximum differences among adjusted and unadjusted young bulls' predictors were 674 pounds of milk and 21 pounds of fat, model I, and 622 pounds of milk and 19 pounds of fat, model II. The rank correlations among all bulls' adjusted and unadjusted predictors were ≥0.98. However, the rank correlations for the best 5% young bulls were 0.84 for milk and 0.92 for fat, model I, and 0.75 for milk and 0.82 for fat, model II. For the poorest 5% young bulls, the rank correlations were 0.76 for milk and 0.91 for fat, model I, and 0.82 for milk and 0.91 for fat, model II;Third BLUP model was used to evaluate sires and cows for production using multiple lactation data with and without adjusting records for present days open, previous days open, and previous days dry. This model included the mean, HYS, groups, sires within groups, and cows nested within sires. The relationship among sires was used. Data were 315,502 cows with 678,502 lactations, daughters of 522 sires. The maximum differences among adjusted and unadjusted predictors were 507 pounds of milk and 16 pounds of fat for sires, and 2485 pounds of milk and 91 pounds of fat for dams. The rank correlations among adjusted and unadjusted predictors were ≥ 0.97 for sires and dams. But, the rank correlations for the best 5% of sires were 0.77 for milk and 0.91 for fat; they were 0.94 for milk and 0.95 for fat, for the poorest 5% of sires. Rank correlations were 0.7 for milk and 0.59 for fat of the top 20 dams and 0.85 for milk and 0.72 for fat of the bottom 20 dams;Adjusting pedigree information for past reproductive performance improved the regression coefficients for sires and maternal grandsires in predicting young bulls' proofs. However, the regression coefficients for dams were very small (0.02 to 0.03) with or without adjustment.