Background: Atherosclerosis (AS) is a complex cardiovascular disorder driven by endothelial cell dysfunction and immune microenvironment dysregulation. We identified novel endothelial-related diagnostic biomarkers through multi-omics integration and machine learning approaches.
Methods: Single‑cell atlas of AS was constructed from scRNA-seq data using the Seurat. Endothelial cell‑specific co‑expression modules and hub genes were identified via high-dimensional WGCNA (hdWGCNA), and key endothelial‑associated differentially expressed genes (DEGs) were obtained by integrating these modules with differential expression analysis. Diagnostic genes were screened using LASSO regression and SVM-RFE using glmnet and caret packages, respectively. Their correlations with immune cell infiltration were assessed by single-sample GSEA (ssGSEA) and the CIBERSORT algorithm. Finally, the binding capacity of the encoded proteins to potential therapeutic agents was evaluated through drug-target prediction using the Enrichr platform and the DSigDB database, followed by molecular docking simulations.
Results: A total of 66 endothelial cell-associated DEGs were identified, from which four core feature genes (ANXA2, DBN1, ZNF385D, and IL6ST) were screened using machine learning approaches. Immune infiltration analysis revealed a global increase in immune cell infiltration (e.g., activated B cells, T cells, and macrophages) in atherosclerotic lesions, with the four genes showing significant correlations with specific immune subsets, while single-cell data further confirmed T cells, macrophages, and B cells as the predominant cellular components in the plaque microenvironment. Molecular docking results demonstrated strong binding capabilities of ANXA2 with thalidomide and IL6ST with resveratrol, with binding energies of -6.7 kcal/mol and -7.4 kcal/mol, respectively.
Conclusion: Our findings provided new insights for the targeted AS therapy.
扫码关注我们
求助内容:
应助结果提醒方式:
