Skip to main content
ARS Home » Research » Publications at this Location » Publication #116984
Title: IMPACT OF PATERNITY ERRORS IN COW IDENTIFICATION ON GENETIC EVALUATIONS ANDINTERNATIONAL COMPARISONS

Author
item BANOS, G - INTERBULL CENTRE
item Wiggans, George
item Powell, Rex

Submitted to: Journal of Dairy Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/6/2001
Publication Date: N/A
Citation: N/A

Interpretive Summary: Errors in identification of cow sires will cause errors in estimates of genetic merit of bulls. Bull evaluations will be more similar than is correct because good (or poor) bulls are assigned daughters sired by more nearly average bulls. The impact of these cow paternity errors on genetic evaluations and international comparisons of dairy cattle for milk, fat and protein yield was investigated. Undiscovered pedigree errors exist in all national databases but were assumed absent except for the 11% introduced in the US Holstein database. Those cows had their sire identification randomly replaced by another sire. These errors reduced estimated annual genetic improvement by 11-15%. Variation among bull estimates decreased by 8-9%. International comparisons of bulls were based on national evaluations from the US, Canada, New Zealand, and the Netherlands. Genetic correlation estimates (a measure of how similarly bulls are expected to rank across countries) between the US and the other countries decreased by .04 to .06. The resulting bias towards domestic bull selection and inability to identify the truly superior animals available internationally could decrease sire selection differential by about 50 pounds of milk on the US scale. Losses for the other three countries were lower because the genetic correlations between them and the US were not unity. Results clearly show the importance of accurate pedigree data.

Technical Abstract: The impact of cow paternity errors on genetic evaluations and international comparisons of dairy cattle for milk, fat and protein yield was investigated. Eleven percent of all AI Holstein cows in the US database had their sire identification randomly replaced by another sire. US genetic evaluations based on these modified pedigree data were computed and compared with official national evaluations. Estimated breeding values from the analysis with 11% paternity error rate were biased, especially in later generations. Estimated genetic trends decreased by 11-15%. Sire standard deviation estimates also decreased by 8-9%. International comparisons of bulls were based on national evaluations from the US, Canada, New Zealand, and the Netherlands. Genetic correlation estimates between the US and the other countries decreased by .04 to .06 by US pedigree errors. The resulting bias towards domestic bull selection and inability to identify the truly superior animals available internationally could decrease sire selection differential by .07-.09 standard deviation units in the US. Losses for the other three countries were lower, ranging from .02 to .05 standard deviation units, because the genetic correlations with the US were not unity. Undiscovered pedigree errors exist in national databases in all countries but were assumed absent except for those introduced. Results clearly show the importance of accurate pedigree data.