Characterization of Lipid Epi-Metabolites/Reactions as Diagnostic Biomarkers for Idiopathic Pulmonary Fibrosis by Lipidepifind

Lipidomics enables comprehensive profiling of lipid species, providing a powerful approach for studying disease pathogenesis and identifying biomarkers. Conventional lipidomics workflows rely on the annotation of lipid species using curated biochemical databases. However, many unidentified lipids are absent from these databases, which limits the discovery of functional or pathological biomarkers. To address this limitation, we developed a method to systematically identify structurally modified lipids, referred to as epi-metabolites, in biological samples. First, 1479 parent lipids were identified across different biological matrices using liquid chromatography-high-resolution mass spectrometry. Next, the MS characteristics of all potential lipid epi-metabolites were predicted using a metabolic network expansion strategy comprising 62 reaction types and were subsequently screened in raw test sample data. To ensure reproducibility and improve analytical efficiency, we implemented this workflow in a user-friendly Shiny app (https://xinguang-liu.shinyapps.io/metabolite_mz_predictor/) and an open-source R package, Lipidepifind (https://github.com/Xinguang-Liu/Lipidepifind). Lipid epi-metabolites were then identified using four validation criteria: MS/MS analysis, retention order filtering, retention time calibration, and collision cross-section validation. Finally, differential structural modifications and putative enzyme-mediated reactions were characterized through multivariate statistics and epi-metabolic reaction enrichment analysis. We demonstrated the utility of this approach in idiopathic pulmonary fibrosis (IPF), identifying 725 lipid epi-metabolites in patient serum. Two enriched lipid oxidative cleavage reactions were observed, and two epi-metabolite biomarkers, phosphatidylcholine (16:0/9:0 (CHO)) and (16:0/5:0 (COOH)), were identified in IPF patients. This method outperformed conventional database-based strategies in matching and identifying lipid epi-metabolites and revealed differential lipid modifications and enzymatic processes in IPF.