Primary glioblastoma transcriptome data analysis for screening survival-related genes

Purpose The aim of this study was to screen survival-related genes for glioblastoma (GBM). Methods GSE53733 was downloaded from Gene Expression Omnibus (GEO) database, including 16 short-term (ST), 31 intermediate (IM), and 23 long-term (LT) survivors. Analysis of variance was used to analyze the expression in three groups. The genes with P < .01 were screened as differentially expressed genes (DEGs). Soft clustering was performed using Mfuzz to mine the expression patterns of differential genes in three groups of overall survival (OS) classification. The cytoscape plugin clueGO was used for functional enrichment analysis. The protein interaction between differential genes was extracted from the STRING V10 database, and the protein-protein interaction (PPI) network was constructed and displayed with cytoscape. The hub genes were verified by quantitative reverse-transcription polymerase chain reaction. Results Total 662 DEGs were obtained among three groups and enriched in 12 clusters. The overlap analysis between clusters could classify these 12 clusters Cluster A and B. Total 264 OS.DEGs were contained in Cluter A and Cluster B, and enriched in 28 Gene Ontology terms, such as trophoblast giant cell differentiation (P value = 6.18E-04), muscle fiber development (P value = 9.09E-04), and negative regulation of stem cell differentiation (P value = 1.76E-03). The top five nodes with highest degree in OS.PPI were HDAC1, DECR1, RASL11A, PDIA3, and POLR2F. The expression of DECR1 and POLR2F was significantly lower, while the levels of HDAC1 and PDIA3 were highly expressed in GBM tissues. Conclusion DECR1, POLR2F, HDAC1, and PDIA3 might be potential key genes affected the overall survival time of patients with GBM.