Interpret Large Networks through Filtering

Introduction

The mammalian heart is a complex structure with highly specialized cells working under a tightly regulated environment. The cardiovascular research community still largely uses conventional approaches (e.g. transgenic mice) and thus focuses on a small group of highly cardiac-specific genes (Lusis and Weiss, 2010). Recently, the emergence of powerful -omics technologies, including microarray and ChIP-Seq, has generated large amount of data that are often non-intuitive to cardiac biologists. One important way to bridge this gap is by creating human-readable visualization (Gehlenborg, et al., 2010).

VISIONET (Visualizing Transcriptomic Profiles Integrated with Overlapping Transcription Factor Networks) was developed as an in-house tool tailored to our biological questions. Being custom-made, VISIONET has features not yet addressed by other popular general-purpose visualization platforms, including CellDesigner (Funahashi, et al., 2003) and Cytocape (Shannon, et al., 2003). In a VISIONET graph, the network topology represents the transcription factor networks constructed from ChIP-Seq datasets, and the node colors represent the transcriptomic profile obtained from microarray experiments. The transcriptomic profile can then be used as a filter to limit the number of nodes visible to the biologist users, allowing the identification of important genes based on human expertise.

The unique node-filtering feature of VISIONET provides biologists with more human-readable visualization than generated by other tools (CellDesigner, Cytoscape). It implements a custom layout algorithm that designed for transcription factor networks. For transcription factor networks, we found the VISIONET custom layout performs faster than established layout techniques including Fruchterman-Reingold (Fruchterman and Reingold, 1991), Harel-Koren (Harel and Koren, 2001), and Sugiyama (Sugiyama, et al., 1981).

Gata4-Tbx20 network

Gata4-Gata6 network

To get started

VISIONET requires two main inputs from the users:

  • (1) the (gene 1, gene 2) tuples that describe the transcription factor network
  • (2) the (gene, value) tuples for the microarray intensity

Input files (1,2) are provided in tab- or comma-separated format, and (1) also accepts GraphML format (Brandes, et al., 2002). Alternatively, the user can supply the raw ChIP-Seq peak list for (1) in the standard COD format, and the raw microarray file for (2) in the in tab- or comma-separated format.

We have provided users with sample test cases, which contain the Gata4-Tbx20 transcription factor network. The ChIP-Seq datasets for Gata4 (GSM558904) and Tbx20 (GSM734426) were obtained from the NCBI GEO database. The microarray intensity for mouse heart and tail fibroblasts were obtained from our previous study (Furtado et al., 2014a-b, Nim et al., 2014).

For network file (1): choose one of the following examples:

For microarray intensity file (2): choose one of the following examples:

To make your own input files for (1) and (2), please refer to the format guideline in VISIONET User Guide.

A desktop companion of VISIONET (running in Microsoft Excel environment) is provided as an additional conveniene to users here. The source code for VISIONET is available here. Refer to the VISIONET User Guide for additional instructions.

Whenever ready, please proceed to VISIONET

References:

Gehlenborg, N., et al. (2010) Visualization of omics data for systems biology, Nature methods, 7, S56-68.

Lusis, A.J. and Weiss, J.N. (2010) Cardiovascular networks: systems-based approaches to cardiovascular disease, Circulation, 121, 157-170.

Funahashi, A., et al. (2003) CellDesigner: a process diagram editor for gene-regulatory and biochemical networks, Drug Discovery Today: BIOSILICO, 1, 159-162.

Shannon, P., et al. (2003) Cytoscape: a software environment for integrated models of biomolecular interaction networks, Genome research, 13, 2498-2504.

Fruchterman, T.M.J. and Reingold, E.M. (1991) Graph drawing by force-directed placement, Software Practice and Experience, 21, 1129-1164.

Harel, D. and Koren, Y. (2001) A Fast Multi-scale Method for Drawing Large Graphs. Proceedings of the 8th International Symposium on Graph Drawing. Springer-Verlag, pp. 183-196.

Sugiyama, K., Tagawa, S. and Toda, M. (1981) Methods for Visual Understanding of Hierarchical System Structures, Systems, Man and Cybernetics, IEEE Transactions on, 11, 109-125.

Furtado, M.B., et al. (2014a) Cardiogenic genes expressed in cardiac fibroblasts contribute to heart development and repair, Circulation research 114, 1422-1434.

Nim, H.T., Boyd, S.E., and Rosenthal, N.A. (2014). Systems Approaches in Integrative Cardiac Biology: Illustrations from Cardiac Heterocellular Signalling Studies. Progress in Biophysics and Molecular Biology, doi:10.1016/j.pbiomolbio.2014.1011.1006.

Furtado, M.B., Nim, H.T., Gould, J.A., Costa, M.W., Rosenthal, N.A., and Boyd, S.E. (2014b). Microarray profiling to analyse adult cardiac fibroblast identity. Genomics Data 2, 345-350.

 

Monash University         SBI Australia