Estimation & Prediction

Optimizing genomic reliability calculation for single-step evaluation of fertility traits in UK dairy cattle

---

Abstract

Current genomic evaluations in the United Kingdom are based on a two-step approach. Given the advantages of single-step (SS) methods, there is interest in implementing SS for multi- and across-breed evaluations in the UK dairy sector. This study focuses on approximating reliabilities in a single-step framework. Estimating genomic reliabilities is computationally intensive, especially for weekly evaluations involving hundreds of thousands to millions of genotypes, as is the case in the UK dairy industry. Our objective is to estimate genomic effective daughter contribution (EDC) gains following the Interbull working group guidelines on genomic reliability calculation (2024). The goal is to provide an approach that can easily be implemented to obtain final genomic reliabilities in a computationally efficient and timely manner. Two fertility traits were considered: calving interval (CI) and non-return rate at 56 days (NR56). The total number of records available was 4,725,264 for CI and 5,330,364 for NR56. Foreign information was incorporated for international bulls, with appropriate adjustments to avoid double counting with domestic data. In total, 145,950 CI and 112,875 NR56 de-regressed proofs were included. Genotypes were available for 881,402 animals and imputed to 79,051 SNPs using findhap (version 3). The pedigree comprised 9,538,578 animals. All data correspond to the August 2025 official UK national evaluations. Analyses were performed using MiX99 (version 25.0929) with a single-step SNP-BLUP (ssSNPBLUP) model assuming 10% polygenic effect. Pedigree reliabilities and EDC were estimated with APAX (version 25.0929). Direct genomic reliabilities for genotyped animals were then computed from adjusted EDC using SNP_BLUP_REL (version 0.990). To estimate genomic EDC gain, a reduced run was conducted in which phenotypic records were removed for bulls born after 2018 and cows born after 2020. Table 1 summarizes the average genomic reliability from the full run, the expected change in genomic reliability (based on GEBV differences between full and reduced runs), the average theoretical EDC (based on genomic reliability from the reduced run), and the adjustment factor (f) for validation bulls. The adjustment factor, which indicated whether genomic reliability from the reduced run is over- or underestimated, ranged from 0.703 to 1.271. This factor was used to calculate the average genomic EDC gain which can then be used in routine genomic reliability calculation together with pedigree reliabilities. It is important to note that computing direct genomic reliabilities for a single trait currently requires approximately 7 hours 30 minutes, whereas pedigree reliabilities require less than 10 minutes. Therefore, this strategy greatly improves the feasibility by incorporating variation in pedigree reliabilities while applying a fixed genomic EDC gain, which could be updated annually or biannually, making routine genomic reliability estimation computationally practical.

Keywords: 2026

How to Cite:

Id-Lahoucine, S., Mrode, R., Winters, M. & Coffey, M., (2026) “Optimizing genomic reliability calculation for single-step evaluation of fertility traits in UK dairy cattle”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286397. doi: https://doi.org/10.31274/wcgalp.23982

Rights: 1