Sustainability & efficiency

Multi-breed genomic prediction for methane production in Australian Angus beef cattle.

---

Abstract

The beef industry faces increasing pressure to reduce methane (CH4) emissions. This study evaluated whether using a combined multi-breed reference population composed of pure breeds improves genomic prediction accuracy of methane emissions for Angus breed compared with using an Angus-only reference population. Methane emissions were measured using the GreenFeed Emission Monitors (C-Lock Inc.), and trial-average methane production (g/day) was calculated for animals with a minimum of five spot samples collected between 2021 and 2023. In total, phenotypes were available for 2,296 feedlot beef cattle (1,350 Angus, 387 Hereford, 211 Shorthorn, 185 Charolais, and 163 Wagyu), and genotypes were available for 10,441 animals (4,139 Angus, 2,253 Hereford, 1,439 Shorthorn, 1,178 Charolais, and 1,432 Wagyu). The reference population for multi-breed scenario included data from all breeds, whereas the reference population for Angus-only scenario included data from Angus animals only. In each scenario, variance components and genomic estimated breeding values (GEBVs) were estimated using univariate genomic best linear unbiased prediction (GBLUP) models, with age and contemporary group fitted as fixed effects. For the multi-breed reference population, breed was also included as a fixed effect. Two genomic relationship metrices (GRMs) were constructed and evaluated for the multi-breed reference population: a shared GRM and a breed-adjusted GRM based on breed-specific allele frequences. A five-fold cross-validation was conducted to assess potential improvements in prediction accuracy gained by including multiple breeds (i.e. Angus plus other breeds) in the reference population. In each fold, 20% of the Angus animals were randomly masked, and variance components were estimated within each scenario. The results showed no substantial improvement in predictive ability for Angus breed when other breeds were included in the reference population. Prediction accuracies were similar across scenarios, regardless of reference population structure, with values of 0.68 ± 0.04 for the Angus-only reference population, 0.67 ± 0.04 for the multi-breed reference population using the shared GRM, and 0.68 ± 0.04 for the multi-breed reference population using the breed-adjusted GRM. However, the multi-breed reference population improved the estimated variance components compared with the Angus-only reference population. In conclusion, prediction accuracy within the Angus breed did not increase when using the larger multi-breed reference population due to strong within-breed structure; however, multi-breed genomic prediction may still be advantageous by capturing more proportion of additive genetic variance while reducing residual variance.

Keywords: 2026

How to Cite:

Clark, S., Madsen, M., Parna, N. & van der Werf, J., (2026) “Multi-breed genomic prediction for methane production in Australian Angus beef cattle.”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286591. doi: https://doi.org/10.31274/wcgalp.24059

Rights: 1