김가빈 (울산대학교)

장귀주 (울산대산학)

유지영 (서울아산병원)

김경곤 (울산대학교)

김가빈 (울산대학교)

장귀주 (울산대산학)

우종규 (서울대학교)

성제경 (서울대학교)

유지영 (서울아산병원)

김경곤 (울산대학교)

Distinguishing microbial contributions to the fecal metabolome from host- and diet-derived signals remains challenging because conventional approaches generally infer microbial involvement from relative abundance differences. Germ-free (GF) mice served as a negative control, and features detected in specific pathogen-free (SPF) feces but absent from all GF samples were classified as microbiota-associated by presence/absence.

 Fecal samples from GF and SPF mice (n=10/group) underwent single-aliquot methanol/methyl tert-butyl ether (MTBE)/water biphasic extraction, yielding polar and lipid fractions. Both fractions were analyzed by liquid chromatography–ion mobility–tandem mass spectrometry (LC–IM–MS/MS). MetaboScape annotation integrated accurate mass, isotopic pattern, collision cross section (CCS), and MS/MS spectra. After filtering, 577 metabolite features and 500 lipid species were retained.

 Of 1,077 features, 48 were detected in SPF samples but in no GF replicate and were designated SPF-exclusive. Partial least squares discriminant analysis separated GF and SPF groups along component 1. The SPF-exclusive set included secondary bile acids, tryptophan catabolites, sulfonolipids, N-acyl glycines, and fatty acid esters of hydroxy fatty acids. These findings show that microbiota-associated remodeling extends beyond polar metabolites to diverse lipids. Combining a GF negative control with CCS-supported annotation provides a stringent framework for prioritizing microbiota-associated metabolites and lipids.