PREDICTION OF BREEDING VALUES FOR TWINNING RATE AND OVULATION RATE WITH A MULTIPLE TRAIT, REPEATED RECORDS ANIMAL MODEL

A genetic correlation near unity between ovulation rate in heifers and later twinning frequency led to consideration of using measures of ovulation rate in heifers for each estrous cycle, beginning at puberty, to increase accuracy of selection for twinning rate. An initial evaluation with a multiple trait animal model for predicting breeding values included six genetic groups: 1) selected Scandinavian bulls, 2) transfers from other populations at the Research Center to a twinning project, 3) early-purchased Holsteins, 4) late-purchased Holsteins, 5) early purchases-other breeds. and 6) late purchases-other breeds. For ovulation and twin measures, heritabilities of .l00 and .070 and repeatabilities of .120 and .092 were assumed. Assumed phenotypic correlation between ovulation and twin measures of .08 was accounted for by genetic correlation of .89 and permanent environmental correlation of .19. The number of animals evaluated was 1,745; 6,912 estrous cycles were measured for ovulation rate on 840 heifers and 1,929 parturitions were observed for Occurrence of twinning on 851 cows, of which 346 had ovulation rate measured as heifers. The remaining 400 animals were foundation animals that created relationships among those with records or were sires of animals with records. The Scandinavian genetic group effect was substantially greater than that of the others. Joint evaluations were compared to evaluations using only twinning measures. For animals with twin evaluations based only on parents but with ovulation rates measured, the multiple trait evaluation increased accuracy of evaluation from .62 (twin information only) to 3 1 . With one parturition, multiple-trait evaluation increased accuracy from .84 to .92. Correlation of multiple trait evaluation of ovulation rate with single trait evaluation was .71 for cows with no parturitions and .87 for cows with one parturition