Disulfidptosis-related gene in acute myocardial infarction and Its diagnostic value and functions based on bioinformatics analysis and machine learning

Liwen ChenJinru Wei^*Guoxiong Deng^*Guien Xu

• Fifth Affiliated Hospital, Guangxi Medical University, Nanning, China

Acute myocardial infarction (AMI) is a leading cause of morbidity and mortality. Disulfidptosis, a newly identified form of programmed cell death, remains poorly studied in AMI. This study aimed to identify disulfidptosis-related genes in AMI and evaluate their diagnostic potential using bioinformatics and machine learning.Microarray datasets GSE60993 and GSE61144 were obtained from the GEO database. Differentially expressed disulfidptosis-related genes were identified using FerrDb V2 and differential expression analysis. A Protein-Protein Interaction (PPI) network was constructed, and key hub genes were screened using Cytoscape. Functional enrichment and correlation analyses were conducted, along with drug target prediction. Machine learning methods—Support Vector Machine (SVM), Lasso regression, and random forest—were applied to narrow down hub genes. Their expression levels were validated using dataset GSE61144.We identified 16 differentially expressed disulfidptosis-related genes, mainly enriched in pathways including “regulation of actin cytoskeleton organization,” “cell cortex,” “cadherin binding,” and “D-glucose transmembrane transporter activity.” Top 10 hub genes included ACTB, RAC1, IQGAP1, FLNB, MYL6, ABI2, DBN1, PRDX1, SLC2A1, and SLC2A3. After machine learning screening, DBN1, RAC1, and SLC2A3 were selected as final candidates. DBN1 and SLC2A3 showed significant differential expression in GSE61144 with AUC ≥ 0.7.In conclusion, this study identified key disulfidptosis-related genes in AMI patients, providing insights into their molecular mechanisms and offering potential biomarkers for diagnosis and therapeutic targeting.