STABILITY OF EVALUATIONS FOR ACTIVE AI BULLS

Authors: POWELL, REX; SANDERS, ASHLEY; NORMAN, H

Submitted to: Journal of Dairy Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/2/2004
Publication Date: 7/25/2004
Citation: Powell, R.L., Sanders, A.H., Norman, H.D. 2004. Stability of genetic evaluations for active artificial insemination bulls. Journal of Dairy Science. 87(8):2614-2620.

Interpretive Summary: The majority of dairy cows in the U.S. are bred to bulls in active AI service. Accuracy and stability of genetic evaluations on those bulls are therefore of considerable importance in providing for the efficient production of dairy products. The latest (November 2003) evaluations were the standard for assessing earlier evaluations. Evaluations for Holstein bulls in active service generally declined in estimated merit, meaning that their earlier evaluations were too high. However, the declines were not of a size to raise serious concern about the utility of those evaluations. The initial active AI evaluation was particularly inflated but that bias diminished regularly for about three years, after which there was little change on average. Bulls sampled by the mainline AI organizations tended to drop less and there was not a great difference in stability among sampling organizations.

Technical Abstract: Genetic evaluations for Holstein bulls in active AI service since 1995 were examined for the changes to the November 2003 evaluation, accepted as the standard. Active AI bulls at each run showed mean declines for all three yield traits for all 31 runs examined. There was no evidence of a worsening situation over time. The early evaluations with active AI status showed much larger declines but this over-evaluation diminished and essentially disappeared after about 3 years. The bulls with first active AI evaluations since 1995 were the primary focus of the study. There was no apparent decline in evaluations for these bulls with the influx of second-crop daughters, attesting to the successful modification to the genetic evaluation system by expanding the genetic variance of short records. Mean decline, and the variation of those differences, was similar by bull sampling organization although the two largest had declines of 15 kg PTA milk less and 30 kg more that the overall mean. Bulls that changed from active to inactive AI status generally declined with that re-designation and declines were even larger compared to November 2003. Bulls with the code for standard sampling declined less than for bulls coded as other sampling, but the differences were much less than from previous reports, suggesting improvements in the industry. Bulls tended to decline with increases in reliability and therefore with time. Change in reliability under-predicted the variation of change in PTA, indicating that there are other important factors or that the assumptions for the calculation of the expected standard deviation of PTA change are not met.