Construction And Functional Analysis Of Transcriptional Regulatory Network In Mouse Brain

Transcriptional factors are the key components for the control of gene expressions, which plays a key role in the development and diversity process and other functions of mammalian brain, and can determine the functions of cells and their response to environments. Many hard and excellent works focused on this topic, aiming to obtain a better understanding about the mechanism of which transcriptional factor invloves in a specific region and how they interact. By performing biological experiments to comfirm or reject hypotheses is important. Alternatively, using binformatics tools to analysis public released data to construct a transcriptional regulatory netowork which can predicts tens of thousands of transcriptional regulatory interactions simultaneously is also a natural choice in the era of omics.Microarray is the key data which can be used to construct a transcriptional regulatory network, which can measures the expression of genes simultaneously under specific conditions. When used for samples from brain, it is a powerful technique for deriving information about specific brain regions.We constructed a transcriptional regulatory network in mouse brain with the use of microarray data. Based on the constructed transcriptional regulatory network, topological analyses and functional analyses were performed.Detailedly, microarray data was downloaded from the website of GEO (Gene expression Omnibus) database, followed by a series of quality control operations. Subsequently, we used ARACNE (Algorithm for the Reconstruction of Accurate Cellular Networks) to build a transcriptional regulatory network in mouse brain, based on5,964differential expressed genes that were selected out by ANOVA (Analysis of Variance). The transcriptional regulatory network we built consists of5,964nodes and89,138links, with each node represents a gene and each undirected link represents a predicted transcription regulation.Topological analyses suggested that the degree distribution of the transcriptional regulatory network we built is approximately scale-free for nodes which has a degree larger than10, consistenting with that in previous reports. Besides, the network has a diameter of6, which is quite small. In the subsequent analysis, FANMOD (Fast Network Motif Detection) was used to search network motif in the constructed transcriptional regulatory network. Totally5motifs made up of3or4nodes were found comparing with sub-graphs in random networks. Then we searched for the hub genes (n=315) based on the degrees of nodes. By only includes hub genes and the edges linked them, we built the hub gene sub-network. This sub-network totally has315nodes and13,142links, with a diameter of3, the degree distribution of which cannot be fitted with power law. Gene-annotation enrichment analysis was performed for the hub genes which revealed that many GO (Gene Ontology) terms are over-represented among hub genes.Sox10, a high-mobility-group transcriptional regulator in early neural crest, which can regulates and interacts with many genes, playing a key role in promoting oligodendrocytes terminal differentiation. In order to discover reliable hypotheses for the subsequent test experiments in our co-operating labortary, a Sox10sub-network was built by including Sox10itself and all its direct neighbors (n=67) in the transcriptional regulatory network. We reported several pairs of interactions in this sub-network, such as Sox10-Plpl, Sox10-Gfap, Sox10-Nkx2.2, Sox10-Mobp, Sox10-Cdh6, and Sox10-Elavl3, with extra support from literature mining, which may gives directions for test experiments.With the construction of the transcriptional regulatory network, we predicted tens of thousands of transcriptional regulatory interactions simultaneously. With indepedent support from literatures, several highly potential transcriptional regulatory interactions were reported, along with some new target genes which has not been reported before. Through the construction of transcriptional regulatory network in mouse brain and the related analyses, we offered dicections for our co-operating labortary, which is meaningful for the reseaches of transcription factors in mouse brain and other related studies.