Skip to main content
ARS Home » Research » Publications at this Location » Publication #99317


item Vanraden, Paul
item Smith, Lori

Submitted to: American Dairy Science Association Proceedings
Publication Type: Abstract Only
Publication Acceptance Date: 6/1/1999
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Inbreeding affects both the mean and variance component of a population. Additive relationship matrices can account for changes in variance, but changes in mean often are ignored. Removal of nonadditive effects makes additive predictions more precise, but nonadditive affects can also be predicted. Past effects of inbreeding were removed from national yield data aby regression, and future effects were calculated as the mean inbreeding would occur if each animal were randomly mated to its breed. A sample of 600 females born in 1995 represented each breed. In Holsteins, adjustments of data for past inbreeding increased the estimated genetic trend by 2.9%. Trend was higher because the model separated additive merit from inbreeding loss. Genetic trend was 5.9% lower after adjustments for both past and future inbreeding were applied, as compared to unadjusted trend. Trend was lower because popular animals cause more inbreeding in later generations than when first tested. Changes in rank were small for the 584 bulls being marketed because random progeny testing ensures that a bull's past sample of mates is similar to his expected future mates. The correlation of unadjusted breeding values with those adjusted for past inbreeding was .997. The correlation of unadjusted breeding values with those adjusted for both past and future inbreeding was .998. Inbreeding may be controlled more easily by mating than by selection. The value of a proposed mating is the progeny's inbreeding depression plus the parent's transmitting abilities adjusted for past inbreeding. Similarly, a bull's transmitting ability adjusted for past inbreeding plus the expected inbreeding depression within each nation could be used for global selection.