GeneNetworkBuilder (GNB) is a web appliation for discovering the transcriptional regulatory network for a given transcription factor (TF) of Caenorhabditis elegans, Homo sapiens and so on, using ChIP-chip (ChIP-seq) combined with gene expression profile from either RNA-seq or expression microarray experiments.
Given a TF, a gene list with altered gene expression and a list of genes that contains binding sites for the given TF in its regulatory region, GNB generates a transcriptional regulatory network to show the direct and indirect targets of the TF and how the indirect targets are connected with miRNAs or other TFs.
The expression gene list can either be a simple gene list or a matrix containing fold change and p-value in this format for filtering, which can be easily generated by Bioconductor package LIMMA, edgeR or REseq. If a matrix is provided, users have the flexibility to filter the expression results by setting threshold for fold change and p-value prior to building network.
By default, the regulatory network is generated using protein DNA interactions (PDI) determined by yeast one-hybrid (Y1H) system Edgedb and the predicted targets of miRNAs from microCosm Targets. Another regulation network (TFBF) is generated using diverse transcription factor binding features revealed by genome-wide ChIP-seq by Wei Niu et.al.. In additon to EDGEdb and TFBF, TRANSFAC, UniProbe and Jaspar are provided as alternatives to build the network.
Please click here to view the workflow of GeneNetworkBuilder.
The network figure is drawn by R package Rgraphviz and Cytoscape Web. The given TF is depicted in rectangle, the direct targets are in octagon shape, and the indirect targets are drawn as circles. The size of nodes is proportional to the number of connections immediate downstream. Up-regulated genes are colored red and the down-regulated genes are colored green. If the microarray results do not contain fold change, all the genes with altered expression will be colored yellow. The miRNAs are colored pink. Genes not present in the expression gene list are shown in cyan.
The network can be viewed interatively or exported as image (png, svg, pdf) or xml (xgmml, graphml, sif) for publication or further analysis.
In the interactive mode, the network can be trimmed or filtered to shown only up-regulated or down-regulated network.
In addition, GO enrichment analysis is optionally included to indentify subnetworks enriched with interesting gene ontologies.