Identifying Glioblastoma Associated Functional Modules Based On Multi-genomic Data

Glioblastoma multiforme(GBM)is a common primary brain tumor in adults,with a median survival rate of 12–15 months.The genome characterization of a large number of GBM patients has revealed the extensive complexity of molecular alterations,such as DNA mutation,copy number alterations(CNAs)and dysfunctional miRNA,and extensive heterogeneity.However,identifying the causal molecular alterations conferring GBM initiation and progression remains a challenge.In this paper,we identified the dysfunctional miRNA-mRNA regulatory module(dMiMRMs)in GBM by integrating the matched miRNA and mRNA expression profiles as well as gene oncology(GO)annotaton information and protein interaction netwrok.The common,subtype-specific and individualized dMiMRMs were identified to characterize the synergetic mechanism of miRNA in GBM.Due to multi-layer factors(DNA copy number,methylation,transcription factors(TF)and miRNAs)could affect the gene expression,we proposed a systematic method called CMDD(Core Modules Driving Dysregulation in cancer)by further integrating DNA mutation,copy number,methylation,mRNA and miRNA expression,miRNA and TF target interaction data and protein interaction network.This method identified the core gene modules of genetic alteration by characterizing dysregulated networks associated with genetic alterations.The high genetic heterogeneity among GBM individuals promoted us to characterize the individualized pathogenesis in GBM individuals.Based on the hypothesis that genetic alterations contribute to the carcinogenesis of individuals by dysregulating gene expression in some key pathways in individuals,a method called IndividualizedPath was further proposed to identify genetic alterations and their downstream risk pathways from the perspective of individuals through combining DNA copy number,gene expression data and topological structures of biological pathways.These three proposed methods were applied to GBM associated multi-omicsdatasets,respectively.Using the method identifying dMiMRMs,we identified five common dMi MRMs using all GBM samples,three of which were associated with the prognosis of GBM patients and were better predictors of prognosis than miRNAs or mRNAs alone.Different GBM subtypes shared common dMi MRMs,and some subtype-specific dMiMRMs were observed.Furthermore,personalized dMi MRMs were identified,suggesting significant individual differences in different GBM patients.Compared with the previous methods,our method not only can identify the dynamic regulation of miRNAs but also mine their co-regulated specific functions.Through further integrating multi-omic data,CMDD identified a core gene module of17 genes including seven known drivers of GBM.The genetic alterations of the module showed significant association with shorter survival of GBM.We further classified the module into two gene sets based on genetic alteration pattern across GBM patients,and found that one gene set directly participated in the glioma pathway,while the other indirectly regulated the glioma pathway,mostly,via their dysregulated genes.Both of two gene sets were significant contributors to survival and helpful for classifying GBM subtypes.To characterize the individualized pathogenesis in GBM individuals,the method IndividualizedPath identified 394 gene-pathway pairs in 252 GBM individuals.The same risk pathways were affected by different genes in distinct groups of GBM individuals with a significant pattern of mutual exclusivity.A global landscape of gene-pathway pairs showed that EGFR linked with multiple cancer-related biological pathways confers the highest risk of GBM,and GBM individuals with EGFR-pathway pairs showed a significantly poor prognosis.In addition,we found that some rare copy number alterations with large effect on contribution to numerous cancer-related pathways in GBM individuals.In summary,we proposed three methods based on multi-omics datasets to identify GBM associated functional modules: the method for identifying dMiMRMs based on the mathched miRNA and mRNA expression profiles;the CMDD method foridentifying core gene modules of genetic alteration by further integrating matched DNA mutation,copy number,methylation,mRNA and miRNA expression prifiles;and the method IndividualizedPath for identifying GBM individualized genetic alterations and their affected risk pathways by combining DNA copy number and gene expression in GBM individuals.Our methods not only can identify the driver molecular alterations underlying GBM,but also characterize their dynamic mechanism.These driver molecular alterations and their dynamic mechanism formed GBM associated core modules,which is helpful for understanding the pathogenesis of GBM.Also,our methods offer the possibility to identify personalized cancer mechanisms,which will help to develop personalized treatment for cancer patients.