Identification of Hub Genes and Immune Infiltration in Coronary Artery Disease: A Risk Prediction Model

Purpose: Our study aimed to establish a prediction model for coronary artery disease (CAD) that integrates immune infiltration and a gene expression signature.

Methods: 613 differentially expressed genes (DEGs) and 12 hub genes were screened via the GSE113079 dataset. The pathway enrichment analysis indicated that these genes (613 DEGs and 12 hub genes) were closely associated with the inflammatory and immune responses. Based on the differentially expressed miRNA (DEmiRNA)-DEG regulatory network and immune cell infiltration, the Lasso algorithm constructed a CAD risk prediction model containing the risk score and immune score. Then, ROC-AUC and polymerase chain reaction (PCR) were performed for validation.

Results: Six hub genes (PTGER1, PIK3R1, ADRA2A, CORT, CXCL12, and S1PR5) had a high distinguishing capability (AUC > 0.90). In addition, the miRNAs targeting 12 hub genes were predicted and intersected with the DEmiRNAs, and the DEmiRNA-DEG regulatory network was then constructed. Two LASSO models and a novel CAD risk prediction model were constructed through LASSO regression analysis, and they both accurately obtained the risk of CAD. The CAD risk prediction model shows good performance (AUC = 0.988). We also constructed a valid nomogram, and PCR results verified three downregulation hub genes and one upregulation gene in the CAD risk model.

Conclusion: We demonstrated the molecular mechanism of the hub genes in CAD and provided a valuable tool for predicting the risk of CAD.