Identification of hub biomarkers of myocardial infarction by single-cell sequencing, bioinformatics, and machine learning

Abstract

BackgroundMyocardial infarction (MI) is one of the first cardiovascular diseases endangering human health. Inflammatory response plays a significant role in the pathophysiological process of MI. Messenger RNA (mRNA) has been proven to play a key role in cardiovascular diseases. Single-cell sequencing (SCS) technology is a new technology for high-throughput sequencing analysis of genome, transcriptome, and epigenome at the single-cell level, and it also plays an important role in the diagnosis and treatment of cardiovascular diseases. Machine learning algorithms have a wide scope of utilization in biomedicine and have demonstrated superior efficiency in clinical trials. However, few studies integrate these three methods to investigate the role of mRNA in MI. The aim of this study was to screen the expression of mRNA, investigate the function of mRNA, and provide an underlying scientific basis for the diagnosis of MI.MethodsIn total, four RNA microarray datasets of MI, namely, GSE66360, GSE97320, GSE60993, and GSE48060, were downloaded from the Gene Expression Omnibus database. The function analysis was carried out by Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Disease Ontology (DO) enrichment analysis. At the same time, inflammation-related genes (IRGs) were acquired from the GeneCards database. Then, 52 co-DEGs were acquired from differentially expressed genes (DEGs) in differential analysis, IRGs, and genes from SCS, and they were used to construct a protein-protein interaction (PPI) network. Two machine learning algorithms, namely, (1) least absolute shrinkage and selection operator and (2) support vector machine recursive feature elimination, were used to filter the co-DEGs. Gene set enrichment analysis (GSEA) was performed to screen the hub-modulating signaling pathways associated with the hub genes. The results were validated in GSE97320, GSE60993, and GSE48060 datasets. The CIBERSORT algorithm was used to analyze 22 infiltrating immune cells in the MI and healthy control (CON) groups and to analyze the correlation between these immune cells. The Pymol software was used for molecular docking of hub DEGs and for potential treatment of MI drugs acquired from the COREMINE.ResultsA total of 126 DEGs were in the MI and CON groups. After screening two machine learning algorithms and key co-DEGs from a PPI network, two hub DEGs (i.e., IL1B and TLR2) were obtained. The diagnostic efficiency of IL1B, TLR2, and IL1B + TLR2 showed good discrimination in the four cohorts. GSEA showed that KEGG enriched by DEGs were mainly related to inflammation-mediated signaling pathways, and GO biological processes enriched by DEGs were linked to biological effects of various inflammatory cells. Immune analysis indicated that IL1B and TLR2 were correlated with various immune cells. Dan shen, san qi, feng mi, yuan can e, can sha, san qi ye, san qi hua, and cha shu gen were identified as the potential traditional Chinese medicine (TCM) for the treatment of MI. 7-hydroxyflavone (HF) had stable combinations with IL1B and TLR2, respectively.ConclusionThis study identified two hub DEGs (IL1B and TLR2) and illustrated its potential role in the diagnosis of MI to enhance our knowledge of the underlying molecular mechanism. Infiltrating immune cells played an important role in MI. TCM, especially HF, was a potential drug for the treatment of MI.