Exploring biomarkers for ischemic stroke through integrated microarray data analysis

Stroke is the third leading cause of disability-adjusted life years worldwide, and drugs available for its treatment are limited. This study aimed to explore high-confidence candidate genes associated with ischemic stroke (IS) through bioinformatics analysis and identify potential diagnostic biomarkers and gene-drug interactions. Weighted gene coexpression network analysis (WGCNA) and differentially expressed genes (DEGs) were integrated to identify overlapping genes. Then, high-confidence candidate genes were screened by least absolute shrinkage and selection operator (LASSO) regression. Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic value of high-confidence candidate genes as biomarkers for IS. The NetworkAnalyst database was used to construct the TF-gene network and miRNA-TF regulatory network of the high-confidence candidate genes. The DGIdb database was used to identified gene-drug interactions. Through the comprehensive analysis of GSE58294 and GSE16561, 10 high-confidence candidate genes were identified by LASSO regression: ARG1, LY96, ABCA1, SLC22A4, CD163, TPM2, SLC25A42, ID3, FAM102A and CD79B. FAM102A had the highest diagnostic value, and the area under curve (AUC), sensitivity and specificity values were 0.974, 0.919 and 0.936, respectively. The HPA database demonstrated that 10 high-confidence candidate genes were expressed in the brain and blood in normal humans. Finally, DGIdb database analysis identified 8 gene-drug interactions. We identified IS-related diagnostic biomarkers and gene-drug interactions that potentially provide new insights into the diagnosis and treatment of IS.