Screening and functional verification of key genes in acute myeloid leukemia / 白血病·淋巴瘤

ABSTRACT

Objective:To explore the key genes related to the development, progression and prognosis of acute myeloid leukemia (AML) based on bioinformatics, and to analyze their functions.Methods:The chip expression profile GSE84881 data set of AML patients including 19 AML samples and 4 normal tissue samples was downloaded from the gene expression omnibus (GEO) database. GEO online tool GEO2R was used to screen the differentially expressed genes (DEG). The DAVID online database was used to make gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis of DEG. The STRING online database was used to analyze the protein interaction (PPI) network of DEG, and the key genes were screened by using the Cytoscape software. The weighted gene co-expression network analysis (WGCNA) was used to build co-expressed network and obtain the central genes.LC-Bio online platform was used to construct Venn diagram and the key genes and central genes in PPI were crossed to finally obtain the true key genes. RNA-seq datasets GSE2191 and GSE90062 of human tissues were downloaded from GEO database to verify the screened key genes. Kaplan-Meier method was used to analyze the effects of key genes on the overall survival (OS) of AML based on the data of GEPIA database.Results:A total of 247 DEG were identified in GSE84881 data set, including 112 up-regulated genes and 135 down-regulated genes. According to the results of GO enrichment analysis, 247 DEG were mainly enriched in the regulation of signal transduction and cell proliferation in the biological process (BP); the cell composition (CC) revealed that these genes were mainly involved in the cytoplasm and exosomes; the molecular function (MF) analysis showed that these genes were mainly enriched in protein binding and calcium binding. Further KEGG pathway enrichment analysis showed that these 247 DEG were mainly involved in NOD-like receptor signal pathway and interleukin 17 (IL-17) signal pathway. And then the 12 key genes were obtained from PPI. WGCNA software was used to screen 13 central genes from GSE84881 dataset and finally 1 real key gene EGF was obtained after taking intersection. Kaplan-Meier method showed that OS time of AML patients in EGF high expression group was decreased than that in EGF low expression group, and the difference was statistically significant( P = 0.044). Conclusions:EGF may be an important diagnosis and treatment target of AML and may become a potential biomarker for clinical treatment and prognosis prediction of AML.