Estimation & Prediction

Developing multi-breed multi-country beef cattle single-step genomic evaluations for feed efficiency

---

Abstract

Novel sustainability traits, such as feed efficiency, are difficult and expensive to measure, leading to small numbers of phenotypes collected. As a result, building a large national reference population is challenging. International collaborations can help expand the reference population by pooling data from multiple countries and performing joint evaluations. In beef cattle, data are collected on multiple breeds and crossbreds of different breed compositions. Taking advantage of information collected across different breeds and crosses is important to improve the accuracy of genomic evaluations, especially when national datasets are limited and in the case of small local populations. However, current international beef cattle evaluations led by Interbeef at the Interbull Centre are pedigree-based, performed separately on each breed, and focus on popular pure breeds and conventional traits. Therefore, the aim of this study was to develop a multi-breed multi-country beef cattle single-step genomic evaluation for feed efficiency. The dataset included ~13,000 animals phenotyped for feed efficiency from 5 breeding organizations across 3 countries (Italy, Great Britain and Ireland), and more than 15 breeds and crossbreds. The pedigree included ~142,000 animals, of which ~20,000 were genotyped. Genetic correlations across countries were estimated with Gibbs sampling using both pedigree and genomic data. A multi-breed multi-country single-step Single Nucleotide Polymorphism Best Linear Unbiased Prediction (ssSNPBLUP) genomic evaluation was implemented, modelling the trait of each organization as a different, but correlated trait. For comparison, a pseudo-national single-step genomic model was applied using only nationally submitted phenotypes and genotypes, mimicking national genomic evaluations. Accuracy, increases in population accuracy, level, and dispersion bias of genomic estimated breeding values (GEBVs) were estimated using the Linear Regression method for animals born from 2021 onwards, and compared between the multi-country and pseudo-national models.Overall, moving from a national to a multi-country reference population increased the accuracy for countries with limited data, such as Italy and Great Britain, while the accuracy did not further improve for Ireland, where the national reference population was already large enough for accurate genomic evaluations. Results for level and dispersion bias varied across validation groups. In conclusion, multi-breed multi-country single-step genomic evaluations are feasible and can improve the accuracy of genomic prediction. This is particularly important for countries with small (local) populations and for difficult-to-measure traits, where national evaluations may not be feasible. These results can contribute to the development of genomic evaluations for novel sustainability traits in cattle, and enable countries with limited national data to have access to GEBVs with acceptable or high reliability.

Keywords: 2026

How to Cite:

Aivazidou, S., Vandenplas, J., Macedo, F., Evans, R., Cassandro, M., Matthews, K., Roozen, T., Burke, M., Veerkamp, R. & Bonifazi, R., (2026) “Developing multi-breed multi-country beef cattle single-step genomic evaluations for feed efficiency”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285436. doi: https://doi.org/10.31274/wcgalp.23706

Rights: 1