GEne Network GEnerator (GeNGe)

Category Cross-Omics>Pathway Analysis/Tools

Abstract The GEne Network GEnerator (GeNGe) is a web-based application for the generation and analysis of gene regulatory networks (GRNs) providing different simulation steps with interactive user interfaces.

An arbitrary number of networks and simulation data can be generated to provide for instance benchmark data for ‘reverse engineering’ applications.

Furthermore, GeNGe offers features for the topological characterization of GRNs. The simulation results can be used to define critical network nodes and suitable candidates for perturbation experiments and thus guide future experimental work.

The GRNs are used to set up a deterministic ordinary differential equation (ODE) system. The gene regulatory model system is composed of instances of mRNAs and proteins acting as transcription factors (TFs) and their corresponding target genes.

Non-linear kinetics based on the logic described by Schilstra and Nehaniv (2008) are used to describe the influence of sets of independently or jointly binding TFs on the expression of a gene.

Various dynamics can be modeled, such as oscillation and bistability.

Global perturbations (network noise) or local perturbations of a single or multiple network nodes can be simulated and the resulting time series are visualized in order to display the perturbation effects.

All results can be downloaded and used for validation of reverse engineering methods or studies of the network dynamics.

Moreover, GeNGe offers features for the topological characterization of GRNs. Network parameters are computed, such as in- and out-degree distributions, average path lengths and clustering coefficients.

Furthermore, by varying parameters of the kinetic laws or by choosing different kinetics, in silico analyses can be performed, e.g. on the effects of knock-downs (partial knock-downs) of a single gene or groups of genes.

The results can be used to define critical network nodes and suitable candidates for perturbation experiments and thus guide future experimental work.

Functionality --

The workflow of GeNGe is divided into three (3) levels. In the first level, the network level, networks are added to a network repository that will be used for further analyses and simulations.

GeNGe provides several pre-defined GRNs, such as a part of the developmental network in sea urchin described by Davidson et al. (2002), artificial networks and network motifs.

Furthermore, the upload of user-defined networks, in the form of tables or adjacency matrices, is supported. Various ‘artificial networks’ can be generated such as random networks, scale free networks and networks composed of small regulatory network motifs (Barabási and Oltvai, 2004; Bollobás et al., 2003; Lee et al., 2002).

Network parameters can be adjusted by the user to generate networks with specific topological characteristics.

Furthermore, it is possible to change any network by adding or deleting nodes and edges as well as associated regulation strengths.

TFs are assumed to bind independently on the DNA. Nevertheless, sets of jointly binding TFs can be specified.

The networks can be visualized and diverse topological measures are calculated, e.g. in- and out-degree distributions, average path lengths and clustering coefficients.

In the next level, the kinetic level, kinetics of the model is specified.

Degradation of mRNA and protein can be modeled by a linear or a Michaelis-Menten kinetic.

The translation is described by a linear kinetic law. For the transcription dynamic, different nonlinear kinetic laws can be selected.

In the third level, the simulation level, parameters of individual kinetic laws can be specified or set randomly.

Based on the network topology, the kinetics and the parameters, an ODE system of the network is set up and exported to PyBioS simulation engine via a web-service based Application Programming Interface (API) (Wierling et al., 2007).

Note: PyBioS is a system for the modeling and simulation of cellular processes. PyBioS acts as a model repository and supports the generation of large models based on publicly available information like data of the Reactome database (see G6G Abstract Number 20267).

Besides un-perturbed time series analyses, global perturbations (such as Gaussian noise) as well as single or multiple local network perturbations (e.g. knock-downs) can be introduced and the resulting 'steady states' of the system are computed.

The procedures can be repeated with different settings and used in an iterative way. All resulting time series can be visualized.

For knock-down experiments, the ratio of each network node of the knock-down and control simulations is calculated and visualized in the network graph.

All results, including the networks in the format of Systems Biology Markup Language (SBML), time series and simulation parameters can be downloaded for further analyses.

System Requirements

Web-based.

Manufacturer

Manufacturer Web Site GeNGe

Price Contact manufacturer.

G6G Abstract Number 20381

G6G Manufacturer Number 101736