Twinning is a multifactorial trait influenced by both genetic and environmental factors that can negatively impact animal welfare and economic sustainability on commercial dairy operations. To date, using genetic selection as a tool for reducing twinning rates on commercial dairies has been proposed, but not yet implemented. In response to this market need, Zoetis (Kalamazoo, MI, USA) has developed a genomic prediction for twin pregnancies, and included it in a comprehensive multitrait selection index. The objectives of this study were to (1) describe a genetic evaluation for twinning in Holstein cattle, (2) demonstrate the efficacy of the predictions, (3) propose strategies to reduce twin pregnancies using this information. Data were retrieved from commercial dairies and provided directly by producers upon obtaining their permission. The twin pregnancies trait (TWIN) was defined as a pregnancy resulting in birth or abortion of twin calves, classified as a binary (0,1) event, and analysed using a threshold animal model. Predictions for a subset of cows were compared to their on-farm twin records. The heritability for twin pregnancies was 0.088, and genomic predicted transmitting abilities ((g)PTAs) ranged from −7.45–20.79. Genetic correlations between TWIN and other traits were low, meaning that improvement for TWIN will not negatively impact improvement for other traits. TWIN was effectively demonstrated to identify cows most and least likely to experience a twin pregnancy in a given lactation, regardless of reproductive protocol used. Effective inclusion of the prediction in a multitrait selection index offers producers a comprehensive tool to inform selection and management decisions. When combined with sound management practices, this presents a compelling opportunity for dairy producers to proactively reduce the incidence of twin pregnancies on commercial dairy operations.

Abortion in dairy cattle causes great economic losses due to reduced animal health, increase in culling rates, reduction in calf production, and milk yield, among others. Although the etiology of abortions can be of various origins, previous research has shown a genetic component. The objectives of this study were to (1) describe the development of the genomic prediction for cow abortions in lactating Holstein dairy cattle based on producer-recorded data and ssGBLUP methodology and (2) evaluate the efficacy of genomic predictions for cow abortions in commercial herds of US Holstein cows using data from herds that do not contribute phenotypic information to the evaluation. We hypothesized that cows with greater genomic predictions for cow abortions (Z_Abort STA) would have a reduced incidence of abortion. Phenotypic data on abortions, pedigree, and genotypes were collected directly from commercial dairy producers upon obtaining their permission. Abortion was defined as the loss of a confirmed pregnancy after 42 and prior to 260 days of gestation, treated as a binary outcome (0, 1), and analyzed using a threshold model. Data from a different subset of animals were used to test the efficacy of the prediction. The additive genetic variance for the cow abortion trait (Z_Abort) was 0.1235 and heritability was 0.0773. For all animals with genotypes (n = 1,662,251), mean reliability was 42%, and genomic predicted transmitting abilities (gPTAs) ranged from −8.8 to 12.4. Z_Abort had a positive correlation with cow and calf health traits and reproductive traits, and a negative correlation with production traits. Z_Abort effectively identified cows with a greater or lesser risk of abortion (16.6% vs. 11.0% for the worst and best genomics groups, respectively; p < 0.0001). The inclusion of cow abortion genomic predictions in a multi-trait selection index would allow dairy producers and consultants to reduce the incidence of abortion and to select high-producing, healthier, and more profitable cows.