Multi-omics of host–gut metabolic crosstalk in Parkinson’s disease

An altered gut microbiome has been implicated in Parkinson’s disease (PD), yet its reciprocal effects on systemic metabolism remain unclear. This cross-sectional study included 67 PD patients and 64 non-PD controls and integrated untargeted serum and fecal metabolomics using liquid chromatography-mass spectrometry with 16S rRNA gene sequencing of the gut microbiome. PD patients exhibited lipid dysregulation, including alterations in serum isoprenoids and their bacteria-derived precursor (2-C-methyl- D -erythritol 4-phosphate), sphingolipids, and glycerophospholipids, with perturbations in fatty acid biosynthesis pathways. Multi-omics integration revealed that PD was more closely associated with gut microbiome and serum metabolomics. Serum N -hydroxy- L -phenylalanine was linked to hydrogen sulfide-producing Desulfovibrio , implicating microbial-related amino acid metabolism. Machine learning models predicted cognition in PD more robustly than motor impairment and were driven exclusively by gut-derived features, including short-chain fatty acid-producing taxa, opportunistic pathogens, and fecal lipid metabolites, whereas no serum metabolites contributed. These findings highlight novel microbial-driven lipid and amino acid pathways with potential biomarker and therapeutic relevance in PD.