Identification of hub genes in vestibular Schwannoma by bioinformatics analyses and machine learning methods

Abstract

Abstract Background Vestibular Schwannoma (VS) is one of the causes of severe hearing loss with poor therapeutic effect and low quality of life. Lacking effective biomarkers may result in underdiagnosis. Therefore, by exploring the mechanism of sensorineural hearing loss, applying new technologies to optimize the effect of gene therapy to restore auditory function will be a key scientific problem that needs to be solved urgently.This study aims to identify possible hub genes and pathways that may provide useful insights into the underlying pathogenesis and inform integrated prevention and treatment for VS. Methods We performed an integrated analysis using multi-omics data to search biomarkers for VS tumorigenesis. This story starts with two gene expression datasets (GSE108524 and GSE39645) collected from the Gene Expression Omnibus (GEO) database to screen the hub genes. Protein-protein interaction (PPI) network was constructed to select hub genes. Hub genes were validated by network topology analysis, biological expression analysis, and machine learning methods-two algorithms (KNN and SVC). The functional annotation and potential pathways of hub genes and known biomarkers were further discovered by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. We further performed gene set enrichment analysis (GSEA) to identify functions that varied between normal and tumor tissues. CIBERSORT score was conducted to check the heterogeneity of immune cells among tumor tissues. Results We identified 425 DEGs between normal tissue and VS from gene expression data. Three hub genes (EGFR, CAV1, and PPARG) were selected according to the PPI network. The average accuracy of 5 algorithms in machine learning methods is 0.956. GO and KEGG analysis found out signaling pathways were significantly enriched pathways for DEGs. There was obvious Geneset enrichment in the grouping of PPARG and CIBERSORT scores on 107 tumor tissues found that the types of immune cells with high and low expression of PPARG were different. Conclusion The dysregulation of three genes may be involved in the pathogenesis of VS, furthermore, they may be used for prognosis and new therapeutic targets. Among these hub genes, types of immune cells with high and low expression of PPARG were different.