Exploring the Common Gene Signatures Between Myocardial Infarction-Reperfusion Injury and the Gut Microbiome Using Bioinformatics

Abstract

Background: This bioinformatics report attempts to explore the cross-talk genes, transcription factors (TFs), and pathways related to myocardial ischemia-reperfusion injury (MIRI) as well as the gut microbiome. Method: The datasets GSE61592 (three MIRI and three sham samples) and GSE160516 (twelve MIRI and four sham samples) were selected in the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) identification (p < 0.05 and |log FC (fold change)| ≥1) together with functional annotation (p < 0.05) was implemented. The Cytoscape platform established the protein-protein interaction (PPI) network. Genes associated with gut microbiome disorder were extracted based on the DisGeNET database, and those associated with MIRI were overlapped. The Recursive Feature Elimination (RFE) algorithm was adopted for selecting features, and cross-talk genes were predicted by the Support Vector Machine (SVM) models. A network encompassing cross-talk genes along with the TFs was thereby established. Result: The MIRI datasets comprised 138 shared DEGs, with 101 showing up-regulation whereas 37 showing down-regulation. Notably, the PPI interwork for MIRI contained 2517 edges along with 1818 nodes. By using RFE and SVM methods, six feature genes with the highest prediction were identified: B2m, VCAM-1, PDIA4, Ptgds, Mlxipl, and ACADS. Among these genes, B2m and PDIA4 were most highly expressed in MIRI and the gut microbiome disorder. Conclusion: B2m and PDIA4 were identified to be significantly correlated with candidate cross-talk genes of MIRI with gut microbiome disorder, implying a similarity between MIRI and Gut microbiome disorder (GMD). These genes can serve as an experimental research basis for future studies.