The Study On The MicroRNA Involved Tissue-specific Gene Regulation Network

MicroRNAs(miRNAs)are small(-22 nt),non-coding RNAs which could negatively regulate gene expression at post transcriptional stage by binding to the target sites of mRNAs.They were found at the end of the 20th century and have been the hotspots of life science for the last decade.Studies so far have revealed that miRNAs are involved in almost all the key biological processes,such as development,differentiation,apoptosis,cell proliferation and carcinogenesis.By targeting on multiple genes and being significantly different expressed in tissues,miRNA is very important to be responsible for the complexity and tissue specificity of gene regulation network.We decided to study on this problem from a systematic perspective and developed a new method to recognize the cell-specific transcriptional start sites(TSS)of miRNAs.So we can understand the transcriptional regulation of miRNAs more accurately.Based on our predicted miRNA TSS,we further predicted more reliable and cell-specific regulation relationships which are from transcription factors(TFs)to miRNAs.Then we predicted the TF-IncRNA regulation relationships.With data of a large number of miRNAs and protein coding genes,it gives a possibility to construct more comprehensive map of gene regulatory networks by integrating gene expression profiles,transcriptional and post-transcriptional regulation relationships.By analyzing the gene regulatory network,it's capable to understand the function,behaviour pattern and tissue specificity of miRNAs in the network on a systematic level.Corresponding to the order in which we studied the microRNA involved tissue-specific gene regulation network,the research mainly contains three parts:1.Identification of cell-specific microRNA transcriptional start sites(TSS).In general,miRNA genes are firstly transcribed into primary miRNAs(pri-miRNAs)by RNA polymerase(Pol)? or ? in nucleus.Then,the RNase Drosha processes pri-miRNAs into precursor miRNAs(pre-miRNAs),which later are exported into the cytosol and turn to mature miRNAs with the help of Dicer.The process of miRNA transcription is basically the same as that of protein-coding genes;however,the 5' cap of pri-miRNA is quickly degraded after transcription.Therefore,the full-length transcript of a miRNA gene is very difficult to be maintained and it is hard to accurately locate the transcriptional start site(TSS)of miRNA.A workflow to identify the transcriptional start sites(TSSs)of miRNAs was built by integrating the data of H3K4me3 and DNase ? hypersensitive sites(DHSs)and combining the species conservation and sequence feature.The proximal H3K4me3 region to each pre-miRNA in each cell line is considered as the potential region containing the TSS of the miRNA.By applying the workflow to the data for 54 cell lines from the ENCODE project,we successfully identified TSSs for 610 intragenic miRNAs and 765 intergenic miRNAs.For these cell lines,we found 2468 alternative TSSs for intragenic miRNAs and 4748 alternative TSSs for intergenic miRNAs.The comparison between our predictions and a set of experimentally verified TSSs collected from literature showed that our method is of better performance than others.We noticed that the CAGE signal at the position of identified miRNA TSSs was much higher in nucleus than that in cytoplasm which supported the current model of miRNA biogenesis.Among the non cell-specific miRNAs,the intergenic ones tend to use more alternative TSSs than the intragenic ones.Then,the intergenic ones could be regulated in a promoter-specific manner when they activate their TSSs in various cell lines.This study of cell-specific miRNA TSSs could benefit the understanding of the specific expression of miRNAs and be used to construct exact regulation networks in different cells.2.The predictions of regulations related with non-coding RNAs.Besides miRNAs,long non-coding RNAs(lncRNAs)which are longer than 200bp,are increasingly discovered recently.As reported,besides known 20 thousands of protein coding RNA,which occupies about 2%of human genome,three quarters of human genomes could be properly transcribed.So most transcripts are non-coding RNAs and they play an important role in normal development,physiological process and diseases.In this chapter,we predicted the regulations related with non-coding RNAs including miRNAs and lncRNAs.Firstly,we integrated results of predictions from different tools and experimental data,so that we got the relatively reliable miRNA-targets regulations.Secondly,based on the conservation distribution of our predicted miRNA promoters,we systematically predicted TF-miRNA regulations.In particular,we used the data of cell-specific miRNA promoters to research two cell lines from heart and found some heart specific miRNA promoters.For the TFs which could regulate the heart specific promoter of hsa-mir-21,we made a GO function enrichment for them and found that the enriched function is consistent with reported function of hsa-mir-21.Finally,we compared the conservation of promoters of different types of lncRNA with that of miRNA.We found that the promoters of lncRNA were obviously less conservative than those of miRNAs.However,promoters of most types of lncRNAs have a similar conservation distribution at TSS as those of miRNAs.Thus,we predicted the TF-lncRNA regulations with the same method of miRNA,which will be helpful for the research of lncRNAs.3.Analysis of tissue-specific gene regulatory network.The great variety of human cell types in morphology and function is due to the diverse gene expression profiles that are governed by the distinctive regulatory networks in different cell types.Regulatory networks of eight human tissues were constructed in this research,which subsume the regulatory interactions between transcription factors(TFs),miRNAs and non-TF target genes.For these eight constructed regulatory networks of human tissues,we investigated their design principles from the perspective of three levels:the local structure of vertices(i.e.degrees),the small circuits(i.e.network motifs)and the assembly of small circuits.The results show that there are in-/out-hubs of high in-/out-degrees in tissue networks.It was also found that the network motifs were mostly feed-forward loops and miRNA preferred to occur in the tissue-specific motifs.In particular,a common bow-tie framework was found that underlies the motif instances.The bow-tie structures of the eight investigated tissue-specific regulatory networks(TRNs)could classify them into symmetric,input-oriented and output-oriented networks.Such bow-tie framework could serve as the model to further understand the structural adaptation of the regulatory system to the specific requirements of different cell functions.Finally,based on the data of discovered motifs,we attempted to construct a motif database of tissue's gene regulation network(TiMoD),which is hopefully to help the experimental biology,system biology and so on.In conclusion,this study focuses on the miRNA and tries to systematically analyze the role of miRNA in gene regulation network by analyzing network.Firstly,we made a systematic study on the miRNA TSS and its cell specificity.And we found that among the non cell-specific miRNAs,the intergenic ones tend to use more alternative TSSs than the intragenic ones.Then,by integrating different tools and utilizing our predicted miRNA TSSs,we acquired reliable miRNA related regulations.Finally,by integrating all the TF and miRNA related regulations we constructed eight tissues' TRNs.By investigating the features of motifs in different TRNs,we found miRNA preferred to occur in the tissue-specific motifs and mainly occupied the input level of bow-tie structure.This indicates that miRNA is very important in forming the complexity of network.In summary,our research can play an important role in the study of tissue specificity of miRNA,transcriptional regulation of miRNAs,microRNA involved tissue-specific gene regulation network and network motif.