Related Experiment Video
Updated: Jun 28, 2025
Author Spotlight: Integrated Multi-Omics Analysis for Unveiling Multicellular Immune Signatures in Clinical Heart Attack Cohorts
Published on: September 20, 2024
Identification of diagnostic biomarkers and immune cell infiltration in coronary artery disease by machine learning,
Xinyi Jiang1,2,3, Yuanxi Luo1,2,3, Zeshi Li1,2,3
1Department of Cardio-Thoracic surgery, Nanjing Drum Tower Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Peking Union Medical College Graduate School, Nanjing, China.
Background:
Coronary artery disease (CAD) is still a lethal disease worldwide. This study aims to identify clinically relevant diagnostic biomarker in CAD and explore the potential medications on CAD.
Methods:
GSE42148, GSE180081, and GSE12288 were downloaded as the training and validation cohorts to identify the candidate genes by constructing the weighted gene co-expression network analysis. Functional enrichment analysis was utilized to determine the functional roles of these genes. Machine learning algorithms determined the candidate biomarkers. Hub genes were then selected and validated by nomogram and the receiver operating curve. Using CIBERSORTx, the hub genes were further discovered in relation to immune cell infiltrability, and molecules associated with immune active families were analyzed by correlation analysis. Drug screening and molecular docking were used to determine medications that target the four genes.
Results:
There were 191 and 230 key genes respectively identified by the weighted gene co-expression network analysis in two modules. A total of 421 key genes found enriched pathways by functional enrichment analysis. Candidate immune-related genes were then screened and identified by the random forest model and the eXtreme Gradient Boosting algorithm. Finally, four hub genes, namely, CSF3R, EED, HSPA1B, and IL17RA, were obtained and used to establish the nomogram model. The receiver operating curve, the area under curve, and the calibration curve were all used to validate the accuracy and usefulness of the diagnostic model. Immune cell infiltrating was examined, and CAD patients were then divided into high- and low-expression groups for further gene set enrichment analysis. Through targeting the hub genes, we also found potential drugs for anti-CAD treatment by using the molecular docking method.
Conclusions:
CSF3R, EED, HSPA1B, and IL17RA are potential diagnostic biomarkers for CAD. CAD pathogenesis is greatly influenced by patterns of immune cell infiltration. Promising drugs offers new prospects for the development of CAD therapy.
