Re-examination of service-sire conception rates in the United States

Abstract

Until recently sire conception rates (SCRs) in the United States had been published only for bulls from artificial-insemination (AI) organizations that paid dairy records processing centers a fee for editing the data and forwarding it to the national dairy database of the Council on Dairy Cattle Breeding (CDCB). In April 2015, the published list expanded because CDCB agreed to cover the fees; now SCRs are published for all AI bulls that exceed the minimum requirements for number of inseminations and herds. The number of bulls across all breeds with published SCRs in April 2015 was 2 799 compared with 1 939 in December 2014. Mean SCR is set to zero for all AI bulls within each breed, whether published or not. The mean of published AI Holstein bulls was 0.36 in December 2014 and 0.00 in April 2015. The standard deviation for Holsteins increased slightly from 2.04% in December 2014 to 2.08% in April 2015 because the additional bulls had lower SCRs than those included before. The SCRs generally increased with bull age as fixed effects for Holsteins in April 2015 were -0.58, -0.37, -0.21, 0.05, 0.06 and 0.63% for ages <1.3, 1.4-1.5, 1.6-1.7, 1.8–2.0, 2.1–4.5 and 4.6–5.5 years, respectively, but declined thereafter. Age effects for SCR are not intended to facilitate comparisons of bull ranking at a common age as is done for yield traits. An SCR is a phenotypic assessment of the bull’s prospective fertility and not a genetic evaluation. It represents a bull’s expected conception rate in the current environment and timeframe in contrast to a reflection of his lifetime success. Correlations between December 2014 and April 2015 SCRs were 0.96 for both Holsteins and Jerseys, 0.82 and 0.87 between April 2013 and April 2015 SCRs and 0.50 and 0.53 between April 2011 and April 2015 SCRs. The April 2015 Holstein and Jersey SCRs were based on a mean of 4 410 and 2 053 inseminations per bull with a mean reliability of 85 and 74%, respectively. Of particular interest was whether including an AI organization-year effect in the model was effective in improving the prediction of fertility. Results revealed that alternative models ignoring AI organization-year effect were still less effective in prediction of future conception rates, the same as in the past.