Comparison of five methods for the estimation of methane production from vented in vitro systems.

BACKGROUND There are several methods for estimating methane production (MP) from feedstuffs in vented in vitro systems. One method (A; 'gold standard') measures methane proportions in the incubation bottle's headspace (HS) and in the vented gas collected in gas bags. Four other methods (B, C, D and E) measure methane proportion in a single gas sample from the HS. Method B assumes the same methane proportion in the vented gas as in the HS, method C assumes constant methane to carbon dioxide ratio, method D has been developed based on empirical data, and method E assumes constant individual venting volumes. This study aimed to compare the MP predictions from these methods to that of the gold standard method under different incubation scenarios, to validate these methods based on their concordance with a gold-standard method. RESULTS Methods C, D and E had greater concordance (0.85, 0.88 and 0.81), lower root-mean-square error (RMSE; 0.80, 0.72 and 0.85) and lower mean bias (0.20, 0.35, -0.35) with the gold standard than did method B (concordance 0.67, RMSE 1.49 and mean bias 1.26). Methods D and E were simpler to perform than method C, and method D was slightly more accurate than method E. CONCLUSION Based on precision, accuracy and simplicity of implementation, it is recommended that, when method A cannot be used, methods D and E are preferred to estimate MP from vented in vitro systems. (c) 2018 Society of Chemical Industry