Differential Regulation Analysis Reveals Dysfunctional Regulatory Mechanism Involving Transcription Factors And MicroRNAs In Gastric Carcinogenesis

Gastric cancer(GC)is one of the most incident malignancies and the leading causes of cancer death around the world,particularly in Eastern Asia.Although lots of genes and microRNAs(miRNAs)have been associated with gastric carcinogenesis,the systematic underlying regulatory mechanisms remain unclear.To explore the dysfunctional mechanisms of GC,this work developed a novel method that identifies carcinogenesis relevant regulatory relationships,which is characterized by quantifying the difference of regulatory relationship between stages and identifying differentially regulated links(DRLs)by integrating differential expression,differential regulation,and the regulation contribution of the regulator to the target.Firstly,based on the transcriptomic datasets of GC including paired mRNA and miRNA derived from TCGA Database,this research applied the strategy of differential coexpression analysis(DCEA)to identify a set of genes/miRNAs whose correlation of expression pattern changed significantly from normal to cancer.The set of genes/miRNAs were used to construct conditional combinatorial gene regulatory networks(cGRNs)involving both transcription factors(TFs)and miRNAs respectively for normal and cancer.Enrichment of known cancer genes/miRNAs and predicted prognostic genes/miRNAs was observed in each cGRN.Then this work built a quantitative method to measure the level of differential regulation for every regulatory relationship between normal and cancer,which could rank the known cancer genes/miRNAs significantly higher.By integrating differential expression,differential regulation,and the regulation contribution of the regulator to the target,DRLs were identified.Many cancer related pathways were presented among genes in DRLs.Meanwhile,three master regulators TCF7L1,TCF4,and MEIS1 were identified by combining DRLs and survival data,and testable hypotheses of dysfunctional mechanisms underlying gastric carcinogenesis related to them were generated.Furthermore,the fine-tuning effects of miRNAs were also observed.Besides,the differential regulation analysis framework was applied to a Chinese GC expression dataset GSE54129 to rank the differentially regulated genes(DCGs)and relationships(DCLs)that most related to GC.Followingly,we selected one of the highest rank TFs CREB1 and its two highly ranked targets,TCEAL2 and MBNL1,and recurred the results of the two predicted differential regulations in normal and cancer through molecular experiment.The mechanism of gastric cancer drawed in this work will provide guidance for future research.We propose that this differential regulation network analysis framework is feasible to gain insights into dysregulated mechanisms underlying other complex diseases and phenotypic changes.