Identification\nof Blood Biomarkers in Ischemic Stroke\nby Integrated Analysis of Metabolomics and Proteomics

We aimed to uncover the pathological\nmechanism of ischemic\nstroke\n(IS) using a combined analysis of untargeted metabolomics and proteomics.\nThe serum samples from a discovery set of 44 IS patients and 44 matched\ncontrols were analyzed using a specific detection method. The same\nmethod was then used to validate metabolites and proteins in the two\nvalidation sets: one with 30 IS patients and 30 matched controls,\nand the other with 50 IS patients and 50 matched controls. A total\nof 105 and 221 differentially expressed metabolites or proteins were\nidentified, and the association between the two omics was determined\nin the discovery set. Enrichment analysis of the top 25 metabolites\nand 25 proteins in the two-way orthogonal partial least-squares with\ndiscriminant analysis, which was employed to identify highly correlated\nbiomarkers, highlighted 15 pathways relevant to the pathological process.\nOne metabolite and seven proteins exhibited differences between groups\nin the validation set. The binary logistic regression model, which\nincluded metabolite 2-hydroxyhippuric acid and proteins APOM_O95445,\nMASP2_O00187, and PRTN3_D6CHE9, achieved an area under the curve of\n0.985 (95% CI: 0.966–1) in the discovery set. This study elucidated\nalterations and potential coregulatory influences of metabolites and\nproteins in the blood of IS patients.