Gene Expression Analysis Reveals Novel Gene Prognostic Signatures Involved in Medulloblastoma Subgroups

PurposeMedulloblastoma is a malignant childhood tumor with four molecular subtypes: WNT, SHH, G3, and G4. The prognosis of these four molecular subtypes is different. WNT has the best prognosis, followed by SHH, and G3 and G4 subtypes have the worst prognosis. This study aimed to identify various molecular subtypes of medulloblastoma can independently predict the prognosis of patients and provide specific treatment way for them.MethodsBased on the data in the GSE37418 dataset, the WGCNA method was used to find the genes most related to these molecular subtypes, and then the top ten hub genes in each subtype were found through the cytohubba plug-in of cytoscape. GO pathway enrichment of four interested modules was used, and then GSE85217 was used for clinical trials in single-factor Cox regression analysis .ResultsThe information was subjected to single-factor Cox regression analysis, and twelve hub genes with the most significant prognostic effects on medulloblastoma were found, and then these genes were subjected to multi-factor Cox regression analysis on each molecular subtype, and finally GNG3 was determined .The combination of CALCB and GCGR can predict the development of SHH well (p=0.0011, AUC=0.734), SOCS3 and HOXC10 can better predict the development of G4 (p=0.044, AUC=0.618), and the combination of ADCY8 and LHX3 can predict G3 Development (p=0.034, AUC=0.675).ConclusionThis report showed a possible evidence that OS-related features of various molecular subtypes of medulloblastoma can independently predict the prognosis of patients with each subtype of medulloblastoma, and provide new therapeutic targets for them.