Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: February 17, 2006
Publication Date: July 9, 2006
Citation: Van Raden, P.M., Tooker, M.E., Cole, J.B., Wiggans, G.R., Megonigal Jr, J.H. 2006. Genetic evaluations for mixed breed populations [abstract]. Journal of Dairy Science. 89(Suppl. 1):98(abstr. 40).
New programs to include data from crossbred animals were tested for genetic evaluation of US dairy cattle. An all-breed animal model including an adjustment for general heterosis was applied to usable records back to 1960. Yield records were adjusted to 36 months of age instead of mature equivalent. Total numbers of cows with records were >20 million for yield traits, for productive life (PL), and for daughter pregnancy rate (DPR), and >10 million for somatic cell score (SCS). About 1% of recent cows are first generation (F1) crossbreds, and the percentage is growing. Unknown parent groups were separate by breed, pedigree path (dams of cows, sires of cows, and parents of sires), US or foreign origin, and birth year. Groups were formed when they included at least 500 animals within time periods and at least 2000 animals across all years. Heterogeneous variance adjustments were estimated separately within herd, year, and sire breed. Convergence of the animal model was fairly rapid, indicating sufficient connections among the breeds and crossbred groups. Genetic differences among breeds were similar to previous estimates, and genetic rankings were very highly correlated with previous within-breed evaluations. Correlations for high reliability bulls exceeded .999 for Holsteins and exceeded .98 within other breeds. Changes were largest in the smallest breeds. Predicted transmitting abilities (PTA) changed more for bulls with fewer daughters, and for cows. Additional group-mates of another breed can add accuracy but also could cause bias if managed differently or not modeled properly. Evaluations of crossbred animals are more accurate in all-breed than within-breed analyses because the relationship matrix can link to reliable sire PTA for breeds in both the maternal and paternal ancestry. Joint evaluation of all breeds and crossbred animals does not greatly change rankings for animals with mostly purebred group-mates and relatives, but provides routine calculation of breed differences. Evaluations may be converted back to original within-breed bases for display to the public.