Identification of atherosclerosis biomarkers through metabolic signatures and immune microenvironment analysis

Atherosclerosis is a chronic immune-metabolic inflammatory disease characterized by dysregulated lipid metabolism. Despite its recognized metabolic basis, biomarkers that reflect coordinated metabolic and immune features remain limited. This study aimed to identify diagnostic biomarkers that capture immune-metabolic features of atherosclerosis. Transcriptomic data were obtained from the GEO database. GO and KEGG analyses were performed. WGCNA identified hub genes. ssGSEA was used to evaluate immune cell infiltration. Unsupervised clustering identified atherosclerosis subtypes. Feature genes were selected using machine learning algorithms. Their diagnostic performance was assessed using receiver operating characteristic curve analysis. Integrated transcriptomic and metabolic analyses identified 17 metabolism-related hub genes associated with atherosclerosis. Based on their expression profiles, patients in the training set GSE100927 were stratified into two subtypes. Cluster 2 showed higher immune cell infiltration than Cluster 1. DGKZ and UAP1 were prioritized as feature genes and demonstrated excellent diagnostic performance in GSE100927 (DGKZ: AUC = 0.983, 95% CI: 0.961–1.000; UAP1: AUC = 0.949, 95% CI: 0.906–0.993), with validation in GSE57691 (DGKZ: AUC = 0.900, 95% CI: 0.780–1.000; UAP1: AUC = 0.822, 95% CI: 0.598–1.000) and moderate stage discrimination in GSE28829. Immune correlation analysis showed that DGKZ broadly positively correlated with immune infiltration (max r = 0.83 for MDSCs), whereas UAP1 exhibited predominantly negative correlations (max r = − 0.65 for immature B cells). DGKZ and UAP1 were also strongly negatively correlated in the training set (r = − 0.769), supporting their association with distinct immune-metabolic states. DGKZ and UAP1 are potential diagnostic biomarkers for atherosclerosis and reflect immune-metabolic heterogeneity across molecular subtypes.