Cerebral infarction (CI) remains a major cause of high mortality and long-term disability worldwide. The exploration of biomarkers and pathogenesis is crucial for the early diagnosis of CI. Although the understanding of metabolic perturbations underlying CI has increased in recent years, the relationship between altered metabolites and disease pathogenesis has only been partially elucidated and requires further investigation. In this study, we performed an integrated metabolomics and lipidomics analysis on 59 healthy subjects and 47 CI patients. Ultimately, 49 metabolite and 68 lipid biomarkers were identified and enriched in 24 disturbed pathways. The metabolic network revealed a significant interaction between altered lipids and other metabolites. Using receiver operating characteristic curve (ROC) analysis, a panel of three polar metabolites and seven lipids was optimized in the training set, which included taurine, oleoylcarnitine, creatinine, PE(22:6/P-18:0), Cer 34:2, GlcCer(d18:0/18:0), DG 44:0, LysoPC(16:0), 22:6-OH/LysoPC, and TAG58:7-FA22:4. Subsequently, a support vector machine (SVM) model was constructed and validated, which showed excellent predictive ability in the validation set. Thereby, the integrated metabolomics and lipidomics approach could contribute to a comprehensive understanding of the metabolic dyshomeostasis associated with the pathogenesis of underlying CI. The present research may promote a deeper understanding and early diagnosis of CI in the clinic. All raw data were deposited in PRIDE (PXD036199).
Read full abstract