Phenomics

Phenotypic and genetic prediction of methane emissions from milk mid-infrared spectra in Norwegian Red dairy cattle

---

Abstract

Milk mid-infrared (MIR) spectroscopy is routinely used in dairy recording and has been proposed as a low-cost indicator for difficult-to-measure traits such as enteric methane (CH₄) emissions. However, evidence for the robustness and transferability of MIR-based CH₄ prediction across cows and herds remains limited, and MIR spectra are themselves heritable phenotypes that may share genetic covariance with methane production.The objective of this study was to evaluate both the phenotypic robustness and genetic relevance of MIR-based prediction of CH₄ emissions using commercial dairy cattle data.The dataset comprised 13,602 test-day MIR spectra from 1,506 Norwegian Red cows, merged with 693,830 daily CH₄ records collected using GreenFeed systems across 14 commercial herds. Methane phenotypes were matched to MIR records using a ±5-day window. Partial least squares regression was used to predict daily CH₄ emissions from MIR spectra, and prediction performance was evaluated using cow-blocked and leave-one-herd-out cross-validation schemes.Phenotypic prediction accuracy was moderate (R2 = 0.31) under cow-blocked cross-validation but declined substantially under leave-one-herd-out validation (R2 = 0.04), demonstrating limited transferability of MIR-based CH₄ prediction across herds. Genetic analyses on each of the 899 MIR wavenumbers showed that MIR wavenumbers were moderately heritable and exhibited genetic correlations with CH₄ production ranging from −0.41 to 0.50, indicating that specific spectral regions capture genetically relevant information related to methane emissions.These results show that while MIR-based phenotypic prediction of CH₄ across herds is limited, MIR spectra contain genetic signal relevant for methane emissions. This highlights the importance of genetic validation of MIR-based methane indicators before their use in dairy cattle breeding programs.

Keywords: 2026

How to Cite:

Bakke, K., Difford, G., Gjuvsland, A., Gorjanc, G., Heringstad, B. & Jenko, J., (2026) “Phenotypic and genetic prediction of methane emissions from milk mid-infrared spectra in Norwegian Red dairy cattle”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285357. doi: https://doi.org/10.31274/wcgalp.23655

Rights: 1