Abstract
AbstractMetabolomic studies are important to understand microbial metabolism and interaction between the host and the gut microbiome. Although there have been efforts to standardize sample processing in metabolomic studies, infant samples are mostly disregarded. In birth cohort studies, the use of diaper liners is prevalent and its impact on fecal metabolic profile remains untested. In this study, we compared metabolite profiles of fecal samples collected as solid stool and those collected from stool saturated liner. One infant’s stool sample was collected in triplicate for solid stool and stool saturated liner. Comprehensive metabolomics analysis of the fecal samples was performed using NMR, UPLC and DI-MS. The total number, identities and concentrations of the metabolites were determined and compared between stool sample collection methods (stool vs. liner). The number and identity of metabolites did not differ between collection methods for NMR and DI-MS when excluding metabolites with a coefficient of variation (CV) > 40%. NMR analysis demonstrated lowest bias between collection methods, and lowest technical precision between triplicates of the same method followed by DI-MS then UPLC. Concentrations of many metabolites from stool and stool saturated liner differed significantly as revealed by Bland-Altman plots and t-tests. Overall, a mean bias of 10.2% in the Bland-Altman analysis was acceptable for some metabolites confirming mutual agreement but not for others with a wide range of bias (−97-117%). Consequently, stool and stool-saturated liner could be used interchangeably only for some select metabolite classes e.g. amino acids. Differences between the metabolomic profiles of solid stool samples and stool saturated liner samples for some important molecules e.g., ethanol, fumarate, short chain fatty acids and bile acids, indicate the need for standardization in stool collection method for metabolomic studies performed in infants.
Publisher
Cold Spring Harbor Laboratory