Submitted to: Journal of Animal Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: March 28, 1996
Publication Date: N/A
Interpretive Summary: Information resulting from genetic evaluation of beef cattle is an integral part of marketing breeding stock and genetic improvement of beef cattle. The accuracy of genetic evaluation rests on statistical assumptions about the data and the analytical models used. In this work, we test the adequacy of procedures currently used in national cattle evaluation to correct weaning weight records for age of dam effects. These procedures are found to be biased. The observed bias materially changes the genetic evaluation for maternal weaning weight. This compromises accurate selection of breeding stock and reduces the potential rate of genetic improvement of maternal weaning weight. Alternative methods for removing this source of bias from national cattle evaluation are recommended.
Technical Abstract: One condition for unbiased prediction of random effects is correct specification of the linear model. The difference in preweaning growth of calves nursing two-year-old dams compared with contemporaries nursing older dams is accentuated in the Miles City Line 1 Hereford herd relative to the differences implied in national cattle evaluation. Three analyses of 205-day weight records for 4998 calves were conducted. Fixed effects in analysis 1 were contemporary groups (year, sex of calf subclasses), age of dam (2,3,4,5-10, & 11+ years of age) and linear regressions on inbreeding of calf and dam. Analysis 2 was conducted using the same model after adjustment of the 205-day weight records using factors from the American Hereford Association's Total Performance Records program. Analysis 3 was similar to analysis 2, except age of dam was omitted from the model. Predictions of direct and maternal additive genetic effects on 205-day weight for sires and maternal grandsires were extracted from each analysis. As differences from the 5-10 year old age effect, estimated age of dam effects from analysis 1 were 45, 19, 6, & 19 kg for 2,3,4, & 11+ age of dam classes, respectively. Comparable estimates from analysis 2 were 20, 6, 1, & 14 kg. The rank correlations of expected progeny differences for direct and maternal 205-day weight from analyses 1 & 3 were .97 & .72, respectively. These data clearly indicate bias of industry standard age of dam correction factors in adjusting 205-d weight of Line 1 Hereford cattle raised at Fort Keogh Livestock and Range Research Laboratory. Also shown is potential for biased genetic evaluation when adjustment factors to correct for dif- ferences in age of dam are not appropriate to the targeted population.