Abstract

Objective The occurrence of cardiovascular disease is linked to endoplasmic reticulum stress (ERS). Nevertheless, the molecular basis of ERS-related genes (ERSRGs) in coronary artery disease (CAD) remains poorly investigated. This study aimed to assess human CAD transcriptome data, pinpoint and verify ERS-associated hub genes in CAD, and identify biomarkers linked to ERS in CAD. Methods Candidate genes for CAD were identified by analyzing three datasets. The overlap between differentially expressed genes(DEGs), key module genes from WGCNA, and a specific gene set was investigated. PPI network was then used to refine the biomarkers, and machine learning were used for additional analysis. Then, diagnostic genes were examined using single-gene gene set enrichment analysis (GSEA). Furthermore, the interactions between biomarkers and various immune cells were investigated.Networks regulating miRNA-mRNA-TF relationships and identifying potential interacting chemicals were successfully established. Finally, to further explore the possible involvement of the discovered biomarkers in CAD, RT-qPCR and endoplasmic reticulum stress in vitro model experiments was performed. Results 63 candidate genes were identified through the intersection of DEGs, a specific gene set, and critical module genes. Functional analysis linked them to T cell receptor signaling and Golgi transport.LR, SVM, and XGBoost models based on three biomarkers showed moderate diagnostic capability with AUC >0.6.Experimental validation confirmed that TMED2 and SEC22B were significantly downregulated in a cellular model of ERS in CAD, which was reversed by 4-PBA treatment.Immune infiltration analysis revealed significant correlations with various immune cells. Based on the CTD, some interacting chemicals were predicted such as cyclosporine, which might serve as potential relevance for CAD by targeting the corresponding biomarkers. Conclusion This study identified three biomarkers, confirming TMED2 and SEC22B as