Microbiome

Prediction of enteric methane emissions from faecal metagenome of dairy cattle

---

Abstract

The rumen microbiome is an indicator of enteric methane emissions (EME), but rumen sampling on commercial farms is impractical. We evaluated whether the faecal microbiome could serve as an alternative for predicting EME. Methane production (MeP; g/d), energy-corrected milk (ECM), methane intensity (MeI = MeP / energy-corrected milk [ECM]), dry matter intake DMI, and methane yield (MeY = MeP / dry matter intake [DMI]) were predicted in 46 dairy cows (26 animals for DMI and MeY). All animals fed similar diets across six experiments. Metagenomic features from 87 rumen and 120 faecal samples and animals' fixed effects were used as predictors in cross-validated random forest models. Ruminal fluid was collected via oesophageal probe, filtered through cheesecloth, and frozen at −80°C. Faecal samples were collected from fresh pats, subsampled, and frozen at −80°C. Microbial DNA was extracted using the ZymoBIOMICS DNA Miniprep Kit (Zymo Research). Sequencing libraries were prepared with the Native Barcoding Kit 96 V14 (Oxford Nanopore Technologies, ONT) and sequenced on a PromethION 24/2 (ONT). Basecalling was performed with FAST model of Dorado and reads ≥250 bp were retained. Samples were retained if they had N50 ≥ 2,000 and ≥1 Gb of Q10 reads. Taxonomic and functional annotation was performed using SqueezeMeta, obtaining features at the superkingdom and genus levels, KEGG Orthology and COG groups. Non-microbial reads were removed. Features detected across all rumen and faecal samples were retained. Abundances were normalised, scaled, and the first five principal components (PCs) computed. Approximately 8,000 metagenomic features (including PCs) were used per sample type. The predictability of methane production was compared using rumen metagenome, faecal metagenome, and animal non‑metagenomic variables. The prediction of MeP was moderate, with R² values of 0.52, 0.43, and 0.54 for the rumen metagenome, faecal metagenome, and non‑metagenomic predictors, respectively. For MeI, the fixed effects and rumen metagenome retained predictive power (R² = 0.53 and 0.12). For MeY, only the non‑metagenomic predictors showed predictive ability (R² = 0.27). Faecal metagenomes showed moderate to high predictive reliability for PCs of the rumen metagenome (R² up to 0.79). Several PCs from the rumen are associated with MeP: R² up to 0.33; methanogenic archaea explain approximately 10-15% of some PCs, while others are characterised by opposing loadings of Bacteria and Eukaryota, previously linked to methane emissions. These results confirm the merit of the rumen microbiome as a methane predictor and suggest that the faecal metagenome may serve as an indicator of rumen metagenome profiles associated with MeP. The predictive performance may reflect the improved resolution offered by long‑read sequencing. Larger, more diverse datasets, along with genetic parameter estimates, are needed to confirm these results and consider including faecal metagenomic information into methane‑reduction breeding strategies.

Keywords: 2026

How to Cite:

Sepulveda, B., Vander Jagt, C., Pryce, J., Chamberlain, A., Wang, J., Retegan, R., Marett, L., Williams, S., Jacobs, J. & Xiang, R., (2026) “Prediction of enteric methane emissions from faecal metagenome of dairy cattle”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2294612. doi: https://doi.org/10.31274/wcgalp.24323

Rights: 1