Research On Gene Regulation Network For Cancer Genomics Data

Cancer is a disease that seriously threatens human health.Its occurrence and development are related to the co-regulation of genes and the interaction between genes and the environment.It is a complex cellular physiological process.The development of genomic technology improved the genetic data of various diseases and made it possible to study individual differences and disease occurrence at the genome-wide level.Understanding the gene function and regulation mechanism of organisms is of great significance for the scientific research and clinical treatment of cancer.This paper built a transcription factor-microRNA co-regulatory gene network model for cancer genomics data.First,the bioinformatics analysis for constructing cancer-associated transcription factors and microRNA co-regulatory gene network models were studied in this paper.Then,the inference method of the feedforward loop kinetic model of transcription factor-microRNA coordinated regulation was studied in this paper.Finally,the model proposed in this paper was verified by actual data.In this paper,cervical cancer-related genes and microRNA data were obtained from the TCGA database,and the differentially expressed genes and microRNAs related to the disease were screened by using the logarithm of the multiple analysis and the false discovery rate of the characteristic genes as criteria.A co-expression network of differential genes and differential microRNAs were built by weighted gene co-expression network analysis.The positional weight matrix was used to predict the binding site of the transcription factors,thereby determine the transcription factor related to the regulation of genes.A transcription factor-microRNA regulatory relationship pair was obtained with the TransmiR database.Finally,23 gene-transcription factor-microRNA feedforward loop motifs were obtained.Based on the differential equations model,the kinetic model of the feedforward loop motif of the gene regulatory network is built.Considering the complexity of noise which exists in the gene data,the variance stabilization process is used to derive the warping function for different noise forms.Model inference and parameter identification are implemented under the Gaussian process regression framework.Aiming at the transcription factors-microRNA cooperative regulation network,this paper studied a multi-transcription factors and microRNA cooperative regulation network model,and uses the Markov chain Monte Carlo algorithm to solve the model and infer the activities of the latent transcription factors and microRNA.According to experiments on the gene network model of multi-transcription factors co-regulation with the mixed noise form,the results on simulated data proved that the model has higher identification accuracy and reliability,the results on real biometric data proved that the model has better approximation ability and explanatory power.