Molecular Breeding Value Prediction of Pregnancy Rate in Holstein Dairy Cows Managed in a Heat-Stressed Environment Using Candidate Gene SNP

Reproductive performance in Holstein dairy cattle managed during summer in southern Sonora is a challenge because of high ambient temperature and relative humidity. Both of these factors contribute to heat stress which influence cow behavior. The physiological response of cows to heat stress is one component of a system-wide gene network. Within this environment, a superior cow’s ability to get pregnant early during postpartum is favorable as to reduce the trait days open and to increase productive life. Recently, many reproductive specialists have recommended using pregnancy rate as measure of reproductive success, after converting this trait into a quantitative value using a linear formula. In comparison to the traditional measure of days open, pregnancy rate calculation includes more easily cows that do not become pregnant; furthermore, the output variable indicates that larger values are more desirable, and therefore, more understandable by dairy producers. The objective herein was to predict pregnancy rate in lactating Holstein cows using molecular markers associated with fertility in Holstein cows under a heat-stressed environment. This study included 500 cows from three dairy herds located in the Yaqui Valley of Sonora. A blood sample was collected from every cow and spotted onto FTA cards. The DNA was extracted from each card and used to genotype 179 tag SNP within 43 genes in the prolactin and GH-IGF1 pathways. Five SNP within the genes IGFBP7, IGFBP2, PAPPA1, SSTR2 and STAT6 were associated with pregnancy rate using a mixed effects model. The genotype term was later included in this model to calculate allele substitution effects. Molecular breeding values of the individual cows were calculated by summing the additive genotype effect for each SNP that showed a significant independent association with pregnancy rate, and the average MBV was 0.46 ± 0.01%. Two statistical regression models were used to predict the variable pregnancy rate: a full model that included effects of days and number of lactations, contemporary group (e.g. farm management group), health status and MBV, and a reduced model that only included MBV. Coefficients of determination were 37.61% and 3.07% for full and reduced models, respectively (P<0.01). These results indicate that five SNP explained only a small proportion of the additive genetic variance for pregnancy rate. Additional research is needed to understand if these results are due to low heritability/repeatability of a fertility and (or) if these results are also influenced by heat stress.