Predicting CO-regulatory Factors Of Serum Response Factor(SRF) And Constructing Regulatory Network

The transcription factors and the cis-regulatory elements correct identification is the keyto the regulation of gene expression, in-depth study of the inner relation between them isextremely important to build the regulatory network with target genes. serum responsefactor(SRF) is Multi-functional transcription factor that is ubiquitous in mammalian. TheDNA binding sites of SRF CC(AT)6GG named CArG element is extremely conservative inmammalian evolutionary history, so the cis-element CArG mutations may produce SRFimproper regulation and lead to many diseases. The conventional sequence analysis method tofind functional CArG elements tend to a high false positive rate, and that restrict theconstruction of the regulatory network between SRF and its co-regulator and their targetgenes. In this study, we searched the CArG elements from the promoter sequence comprising4,000bp up and4,000bp downstream from the transcription start site (TSS) in humangenome, and took advantage of46-way vertebrate genomes comparison data，then identifiedthe conserved CArG elements4744in human genome. Combining with expressiondata(normal VS silence SRF), we further confirmed the SRF target genes652under specificexperimental conditions. In order to construct the regulatory network, we clusterd the SRFtarget genes by function, genes with the same or similar functions were clustered in the sameclass. We took the sequences comprising500bp up and downstream from CArG elements ofSRF target genes in each functional class to predict the co-regulatory with SRF, and selectedthe transcription factor that play an important regulatory role under the specific experimentconditions by penalized multivariate regression, and find that SRF may interact with differentco-regulatory factors to regulate the transcription of target genes in different functional classes,in other words, the regulation model of SRF in different functional classes is different. Afterthe above process, we have found SRF target genes and co-regulatory factors under thespecific experiment condition, at last we constructed the regulatory networks between SRFand co-regulatory factors and their target genes in each functional class.