Prediction of the potential Diagnostic Markers of Peripheral Arterial Disease by machine learning and the Correlation with the Ferroptosis and Network of Cancer Gene

Background: Peripheral arterial disease (PAD) is a widely and complex disease which also known as arteriosclerosis obliterans(ASO). This disorder causes an increasingly prevalence all around the world. Despite its high prevalence and huge disruptive impact on the social economy globally, a large proportion of patients that suffers from PAD cannot receive early and proper therapy. Even some patients are not diagnosed so clearly that they missed the PAD optimal treatment. However, seldom research on PAD’s early diagnosis was based on bioinformatics assisted by machine learning. And the correlation between PAD, ferroptosis and cancer is still unavailable. The aim at this study is to seek potential diagnostic markers for PAD and the related genes of PAD-ferroptosis and analyze the relationship to malignant tumors. Methods: We used PAD datasets from Gene Expression Omnibus (GEO) database. R software was used to identify differentially expressed genes (DEGs) in PAD. Then we performed functional correlation Bioinformatics analysis such as Gene Ontology(GO) analysis, Kyoto Encyclopedia of Genes and Genomes(KEGG) pathway enrichment analysis and Network of Cancer Gene (NCG) enrichment analysis. The protein–protein interaction analysis (PPI) network was also built and enriched for DEGs. The hub genes were acquired by Cytoscape software. Hub genes were taken to intersection with ferroptosis related genes for acquiring PAD-ferroptosis related genes. And we selected the key gene from hub genes by using support vector machine-recursive feature elimination (SVM-RFE) and least absolute shrinkage and selection operator (LASSO) logistic regression methods. By the use of R software, we drew the ROC curve to evaluate the diagnostic efficiency of PAD. Results: A total of 176 DEGs, containing 53 up-regulated and 123 down-regulated DEGs, were identified. FBXW7 and YWHAE were turned out to be the PAD-ferroptosis related genes. PAD has a potential correlation with many types of cancer. SMARCA4 and YWHAE can be treated as the diagnostic markers of PAD(AUC＞0.8). Conclusion: To summarize, SMARCA4 and YWHAE were identified as diagnostic markers of PAD. FBXW7 and YWHAE were selected to be the PAD-ferroptosis related genes. YWHAE may be the crossing gene among PAD, a part of malignant tumors and ferroptosis. Trial registration: Our reaserch was based on bioinformatic analysis and we obtained the date from Gene Expression Omnibus database.