Font Size: a A A

Research An Gene Regulatory Network Models In Gliomatosis Cerebri

Posted on:2019-11-08Degree:MasterType:Thesis
Country:ChinaCandidate:L ZhaoFull Text:PDF
GTID:2404330590965994Subject:Systems Science
Abstract/Summary:PDF Full Text Request
High-throughput technology generates a large amount of bioinformatics data,and analysis of these data is the focus and difficulty of system biology research.Usually,cluster algorithms is used to find genomes with similar expression patterns,construct gene regulatory networks,and analyze Interaction relationship with genes.Traditional clustering algorithms cluster on all experimental conditions and can only find non-overlapping clustering genes,while a large number of genes may participate in multiple biological pathways.The biclustering algorithm can solve these problems,but the biclustering algorithm also has many defects.This study uses a single column vector clustering method to optimize the dual clustering algorithm based on greedy random adaptive search process.The gene expression data of yeast cells was used to verify the effectiveness of the optimized algorithm.Experiments show that the optimized dual clustering algorithm can produce more and better dual clustering results.The double clustering found in this paper covers the gene expression matrix.The rate reached 34.77809%.The algorithm was applied to glioma gene expression data and 34 co-expressed genes were obtained.Through detailed analysis and evaluation of several existing classical gene regulatory network models,it was found that the Bayesian network model is suitable for incomplete data sets.It uses a directed acyclic graph model to reveal the regulation of gene expression relationship based on statistical hypotheses.By calling the bnlearn package in R.using 34 co-expressed genes a gene regulatory network of gliomas was constructed.The experiments showed that HAS2,MICAL2,GML,MARK4,SART3,RPP14,CYP11B2,SEMA3 F,CHMP2A,NPPA,ZKSCAN5,SAE1,CTSO,ZNF324 B,GLYAT,DLAT,DNMT1,C20orf43,this 18 genes have a regulatory relationship,which may be related to the pathogenesis of glioma.In this study,a biclustering algorithm based on a greedy random self-adaptive search process was optimized to discover more co-expressed genes in gliomas and a gene regulatory network for gliomas was constructed.For the next step to study the pathogenesis of glioma lay the foundation.
Keywords/Search Tags:Gliomatosis Cerebri, biclustering algorithm, gene regulation networks
PDF Full Text Request
Related items