Use of international clinical mastitis data as independent trait in the US evaluation system

Abstract

Since its first publication in Holsteins (HOL) in April 2018, and later inclusion of Jerseys (JER) in April 2020, US bull evaluations for mastitis resistance (MAS) have been exchanged with Interbull participating countries. Foreign phenotypes for MAS have been used since then to enhance the domestic reference population. Prior to April 2021, countries had the choice of exchanging pure clinical mastitis, somatic cell score, or a combination of methodologies where clinical/subclinical mastitis or a multi-trait approach using multiple sources of information. Hence, only IDs of bulls coming from certain countries with similar trait definitions and if the country of most daughters does not send only SCS were being used in the US. Effective in the April 2021 routine run, Interbull introduced a new trait named SNP training for clinical mastitis (STCM) to better estimate SNP effects specifically for clinical mastitis. The new edits in the US were validated in January 2021 in a full test run. In the April run, genotyped bulls with an international evaluation from the other participating countries used STCM results, whereas the previous trait (called hereafter as the MAS), that combines mastitis from some countries and correlated SCS from others continued to be used for non-genotyped animals. In order to evaluate the impact on the US evaluations, this study aimed to compare PTA and REL between December 2020 routine run (2012r) and January 2021 test run (2101t) and April routine run (2104r), for both HOL and JER breeds, but now taking into account the aforementioned criteria. Descriptive statistics, Pearson and Spearman correlations (rg) as well as regression coefficients (b1) by predicting MAS on STCM to measure potential biases, were calculated. The 2104r PTA means were slightly smaller for HOL and relatively smaller for JER in all scenarios. Pearson and Spearman correlations were always higher than 0.90 in all scenarios for both breeds, no matter the evaluation set of comparison. These results demonstrate a practically null impact on the US evaluations. By comparing 2012r and 2101t, the b1 values were, in general, close to 1 (range 0.98-1.06). On the other hand, a bit more bias can be seen by comparing 2012r with 2104r and 2101t with 2104r. These results may have been partially due to >30,000 corrected phenotypes received in the U.S. April evaluation. Our results suggest that minor impact is expected, and genetic progress improvement is enabled by the implementation of international STCM in the next routine run (April 2021). On the other hand, identifying which countries or individual bulls had direct MAS or only correlated SCS was previously difficult to automate correctly at the national level. Therefore, it is of most importance that Interbull continues to keep track of STCM countries, so several benefits such as gains in REL for bulls with many STCM daughters, would be achieved.