Identification effective diagnosis biomarkers and immune cell infiltration in liver cancer by bioinformatics analysis and machine learning algorithm

Abstract

Abstract Objective The aim of this study was to identify the key diagnosis biomarkers of abnormal expression and immune infiltration in liver cancer based on bioinformatics analysis and machine learning algorithm. Methods Three microarray datasets were obtained from Gene Expression Omnibus database, of which GSE88389 and GSE121248 were defined as training sets and GSE45436 was defined as a validation set. Differentially expressed genes (DEGs) were identified and functional and pathway enrichment analysis was performed by Gene Ontology terms (GO), Kyoto Encyclopedia of Genes (KEGG), disease enrichment analysis (DO) and gene enrichment analysis (GSEA). Tumor biomarkers for liver cancer were identified through Lasso and support vector machine (SVM) and validated in the GSE45436. CIBERSORT was performed to analyze the relation between the diagnosis biomarkers for liver cancer and immune cell infiltration. Results A total of 39 differentially expressed genes (DEGs), including 6 up-regulated and 33 down-regulated genes, were obtained based on expression fold change and significance. Analysis of GO, KEGG, DO and GSEA pathways indicated that these DEGs were enriched in collagen-containing extracellular matrix, collagen trimer, plasma lipoprotein particle, bile secretion, tryptophan metabolism, retinol metabolism, chemical carcinogenesis - DNA adducts, C-type lectin receptor signaling pathway, acute porphyria, lung squamous cell carcinoma, cell cycle, DNA replication, proteasome and ribosome. Combination analysis of Lasso and Support vector machine (SVM), three diagnosis value genes of CAP2、CXCL14 and TMEM27 for liver cancer were obtained by Venn diagram. Compared with normal tissue, immune infiltration analysis demonstrated that naive B cells, memory B cells, regulatory T cells and macrophages were highly expressed in hepatocellular carcinoma tissues, while plasma cells and T cells were low expression. Conclusion We identified CAP2、CXCL14 and TMEM27 as potential biomarkers for liver cancer and that can mediate immune cell activity in liver cancer.