摘要：This study presents the application of metabolomics to evaluate changes in the rumen metabolites of beef cattle fed with three different diet types: forage-rich, mixed and concentrate-rich. Rumen fluid samples were analysed by 1H-NMR spectroscopy and the resulting spectra were used to characterise and compare metabolomic profiles between diet types and assess the potential for NMR metabolite signals to be used as proxies of methane emissions (CH4 in g/kg DMI). The dataset available consisted of 128 measurements taken from 4 experiments with CH4 measurements taken in respiration chambers. Predictive modelling of CH4 was conducted by partial least squares (PLS) regression, fitting calibration models either using metabolite signals only as predictors or using metabolite signals as well as other diet and animal covariates (DMI, ME, weight, BW0.75, DMI/BW0.75). Cross-validated R2 were 0.57 and 0.70 for the two models respectively. The cattle offered the concentrate-rich diet showed increases in alanine, valerate, propionate, glucose, tyrosine, proline and isoleucine. Lower methane yield was associated with the concentrate-rich diet (p < 0.001). The results provided new insight into the relationship between rumen metabolites, CH4 production and diets, as well as showing that metabolites alone have an acceptable association with the variation in CH4 production from beef cattle.
