Summary and Implications
Data from a large pig breeding company which included number born alive records from company nucleus farms, company multiplier farms, customer-owned daughter nucleus farms, and customer-owned closed herd commercial farms were used to estimate breeding values and prediction accuracies of selection for purebred Landrace and Large White females. Prediction accuracies when breeding values were estimated using purebred information only and purebred and crossbred information were compared for three customer-owned commercial herds with pedigree information on more than 70% of their Landrace/Large White crossbred females. Change in accuracy and Pearson correlations with and without crossbred daughter records were calculated for active purebred sows. Estimates were given according to the number of additional daughter records obtained when crossbred records were included in breeding value estimation. Change in accuracy increased from 0.001 to 0.23 when 0 and >20 daughter records were included, respectively. Pearson correlations between accuracy estimates decreased from 1.00 to 0.71 when 0 and >20 daughter records were included, respectively. Results suggest that the inclusion of crossbred daughter records increases selection accuracy in purebred sows and that this could impact selection and mating decisions to improve genetic progress at the commercial level.
Iowa State University
Beam, Denise; Mabry, John W.; and Stalder, Kenneth J.
"Number Born Alive Accuracy Changes Based on the Number of Crossbred Records Included for Large White and Landrace Breeding Value Estimation from a Large Seedstock Supplier,"
Animal Industry Report:
AS 662, ASL R3114.
Available at: http://lib.dr.iastate.edu/ans_air/vol662/iss1/75