Identification of five characteristic ferroptosis-related genes as novel biomarkers for ischemic stroke using bioinformatic and machine learning strategies

Abstract

AbstractFerroptosis is crucial in neuronal cell death, associated with various neurological disorders. However, the role of ferroptosis-related genes (FRGs) in ischaemic stroke (IS) has yet to be well elucidated. We downloaded IS-related gene information and FRGs from the Gene Expression Omnibus (GEO) and Ferritin databases (FerrDb). 22 IS-related DE-FRGs were obtained. Functional enrichment analysis showed that these genes involve multiple regulatory pathways related to IS pathogenesis, including redox, amino acid metabolism, and cell cycle. Subsequently, DUOX2, MDM2, EGFR, MAP3K14, and TRIM46 were identified as core marker genes among the 22 DE-FRGs by lasso and SVM-RFE algorithms. The construction of the nomogram using the five marker genes had excellent diagnostic value. In addition, CIBERSORT analysis showed that changes in the immune microenvironment of IS patients might be associated with TRIM46, MDM2, and DUOX2. In addition, a total of 66 drugs targeting two characteristic FRGs were obtained. The ceRNA networks revealed complex regulatory relationships based on characteristic FRGs. These findings provide new insights into the diagnosis and treatment of IS.