An integrative approach to inferring gene regulatory module networks

Abstract

Background: Gene regulatory networks (GRNs) provide insight into the mechanisms of differential gene expression at a system level. However, the methods for inference, functional analysis and visualization of gene regulatory modules and GRNs require the user to collect heterogeneous data from many sources using numerous bioinformatics tools. This makes the analysis expensive and time-consuming.

Results: In this work, the BiologicalNetworks application-the data integration and network based research environment-was extended with tools for inference and analysis of gene regulatory modules and networks. The backend database of the application integrates public data on gene expression, pathways, transcription factor binding sites, gene and protein sequences, and functional annotations. Thus, all data essential for the gene regulation analysis can be mined publicly. In addition, the user's data can either be integrated in the database and become public, or kept private within the application. The capabilities to analyze multiple gene expression experiments are also provided.

Conclusion: The generated modular networks, regulatory modules and binding sites can be visualized and further analyzed within this same application. The developed tools were applied to the mouse model of asthma and the OCT4 regulatory network in embryonic stem cells. Developed methods and data are available through the Java application from BiologicalNetworks program at http://www.biologicalnetworks.org.

Figure 1. Screen-shot of the Multi-Experiment viewer (Use Case #1, Study 2).

(A) The matrix represent the genes (in columns) co-expressed with the query gene(s) in microarray experiments (in rows). The brightness of blue of the matrix element corresponds to the co-expression value of the gene in an experiment (Eq. 4). The genes and experiments are sorted by average Z-values of genes (Eqs. 1–3). The vertical and horizontal levers allow selecting the highest ranked genes and experiments for building regulatory modules (the selection is shown in a black square). Hovering over the genes and experiments brings up their short description. (B–C) Clicking on the experiment ID brings up the experiment properties and visualization of the expression data. (D) A word cloud that characterizes the found set of experiments described by keywords (ontology terms representing cell types, tissues, diseases, biological processes, etc.). Clicking on the term in the cloud highlights respective experiments. The ‘Recalculate’ button allows the user to recalculate the matrix choosing only the experiments containing selected terms.

Figure 2. Integrative view of the OCT4 regulatory network (Use Case #1, Study 2).

(A) Gene regulatory modular network of OCT4 transcription factor. Grey boxes represent the gene regulatory and co-expressed modules; rectangles represent the genes; red rectangles, the genes with known binding sites; a yellow triangle, the transcription factor; blue edges, TF-target gene relationships; red lines, co-expressed TF-gene pairs. The top module (shown in C), called ‘Module 1′, is highlighted. (B) GenomeBrowser window showing the sequences of the genes and TF binding sites. The OCT4 binding site for the selected in the network (A) Pou2f1 gene is shown. (C) Module Table showing the gene modules, TFs, and functional annotation for each module with Fisher enrichment score (p-value) of GO terms. The top ‘Module 1′ is highlighted. (D) Table of TFs and target genes found in public databases. Gene Pou2f1 (selected in A) is highlighted in magenta. (E) Multi-Experiment Viewer represents the matrix of genes (in columns) co-expressed with the query gene(s) in microarray experiments (in rows). (F) Microarray Gene Expression window showing the hit map and hierarchical tree of clustering data from selected experiments. Pointing out the mouse on the tree vertex shows the significant GO terms for the cluster; ‘Module 1′ is highlighted.

Figure 3. Screen-shot of BiologicalNetworks showing top OCT4 regulatory modules (Use Case #1, Study 2).

The top module is marked in red as it contains OCT4 gene and the genes (marked in red) that are co-expressed with OCT4 in the selected in Study 2 experiments. It is also marked in grey as it contains genes (marked in grey) in which protein products are known to be involved in protein-protein interactions with OCT4 either in human or mouse. And it is marked in blue when it contains genes that have been selected in Study 2 as the mouse or human genes containing known or predicted OCT4 binding sites in the promoters. The ‘G’ column specifies the number of genes in each module. The ‘%’ column represents functional coherence of each module, measured as percentage of genes in the module covered by significant gene annotations (at a specified threshold on p-value). Each module is formed by a part of hierarchical clustering tree and thus represents a hierarchical tree with different terms assigned to different clusters. For each selected and shown GO term, we provide p-value, number of genes assigned to this GO term (the ‘List Hits’ column), number of genes in the tree clusters associated with this term (the ‘List Total’ column), and number of genes with this term among all mouse genes (the ‘Population Hits’ column) in the ontology (the ‘Population Total’ column). Genes with GO terms listed are shown in bold. Column ‘Regulators’ contains transcription factors and regulators (in this case OCT4 only) predicted to regulate a respective module. The search window on the right bottom allows the user to search genes and GO terms in the table.

Figure 4. The modular network inferred for the genes from Module NC_622 (Use Case #2, Study 2).

Grey boxes represent gene regulatory modules; rectangles, genes in the modules; red rectangles, genes from module NC_622; yellow triangles, transcription factors with known binding sites; red triangles, transcription factor that are co-expressed with the genes in the modules; red diamonds, regulators that are co-expressed with the genes in the modules; blue edges, TF-gene binding; red edges, co-expression relationships; grey edges, protein-protein interaction.