Bioinformatics Advance Access originally published online on June 16, 2006
Bioinformatics 2006 22(16):2051-2052; doi:10.1093/bioinformatics/btl322
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
JCell—a Java-based framework for inferring regulatory networks from time series data
Centre for Bioinformatics Tübingen (ZBIT), University of Tübingen Sand 1, 72076 Tübingen, Germany
*To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT 1 INTRODUCTION2 CONCEPTSREFERENCES |
|---|
Motivation: JCell is a Java-based application for reconstructing gene regulatory networks from experimental data. The framework provides several algorithms to identify genetic and metabolic dependencies based on experimental data conjoint with mathematical models to describe and simulate regulatory systems. Owing to the modular structure, researchers can easily implement new methods. JCell is a pure Java application with additional scripting capabilities and thus widely usable, e.g. on parallel or cluster computers.
Availability: The software is freely available for download at http://www-ra.informatik.uni-tuebingen.de/software/JCell
Contact: christian.spieth{at}uni-tuebingen.de
|
1 INTRODUCTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CONCEPTSREFERENCES |
|---|
Within the last few years, researchers obtained large amounts of datasets from experiments. For example, fermentation experiments yield data series of the metabolic system of yeast and DNA microarray technology allows the measurement of gene expression levels for a whole genome. These techniques provide new insights into activities of system components under different biochemical and physiological conditions and can therefore be used to extract time-dependent relationships of interacting molecules like RNA, or proteins.
Although an increasing number of such microarray and metabolic datasets exist, still only few software packages are available to determine regulatory networks from time series datasets. Existing software solutions that are freely available were developed for evaluation procedures addressing specific biological questions. Graphical tools like GenMAPP (Dahlquist et al., 2002) allow only for visualizing microarray data on pre-compiled pathways. Simulation frameworks such as CellDesigner (Kitano et al., 2005) or BioNetGen (Blinov et al., 2004) focus on the modeling of network topologies and their simulation, without the ability to learn network dependencies from experimental data. Tools for inferring such dependencies, such as GeneNetwork by Wu et al. (2004) or the SBToolbox (Schmidt and Jirstrand, 2006), on the other hand provide a limited number of mathematical models or focus only on a few optimization methods and rely on commercial software.
JCell is an application to address the problem of combining algorithms and models to infer regulatory systems from time series data. First, It aims to provide a framework structure that enables researchers to use these methods without deeper knowledge of the underlying mathematics. Second, one of the design goals is the independence of commercial software to ensure the spirit of free research. Further on, owing to the modular design it can be easily extended with new algorithms and strategies. For this purpose, we implemented non-mandatory interfaces to other packages such as MATLAB, Maple, JavaEvA, evolvica and libEA (Fig. 1).
|
2 CONCEPTS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CONCEPTSREFERENCES |
|---|
The proposed application is aimed to support biological and medical researchers in generating hypotheses about topologies and kinetic parameters of regulatory systems. This is done by searching for dependencies in time series data. JCell was initially designed to handle gene expression data from microarray experiments and was later extended on metabolic data because they have similar properties with respect to the mathematical modeling of the underlying systems. To reconstruct regulatory systems, JCell tries to model time series data by finding a topology of the network, together with a mathematical description and the corresponding kinetic parameters of the given dynamic systems.
The actual computational search is performed mostly with Evolutionary Algorithms (Bäck et al., 1997) such as Genetic Algorithms, Evolution Strategies, Differential Evolution, Particle Swarm, Genetic Programming, Memetic Algorithms and several multi-objective optimizer together with direct heuristics, if applicable. Overall, the framework comprises more than 10 different optimization strategies including their parallel and distributed versions that allow for running the tool on high-performance clusters. Further on, the program provides several mathematical models for simulating regulatory systems, like Boolean Networks, linear and pseudolinear Systems, S-Systems, H-Systems and arbitrary differential equations. We also implemented models to simulate intra-cellular systems stochastically (e.g. Bayesian Networks), where a deterministic approach is unfavorable as it is in the case for immune relevant signaling cascades. For the inference process, multiple novel strategies have been developed at our institute and implemented in JCell, including iterative knock-outs, topology identification and graph minimization problems. The corresponding publications and a user guide/tutorial can be found on the website. In addition, JCell enables the user to edit the network topology, thus adding biological information of the network into the inference process, enforcing the algorithm to find networks with specified properties. This can be done manually, using a built-in graph editor, or by importing topologies from external sources. For example, pathways can be incorporated by querying public databases like KEGG (Kanehisa et al., 2000) and TransFAC (Wingender et al., 2001) to import topology graphs into JCell (Spieth et al., 2005a). The imported pathways are used as templates for the inference process. They can either be fixed throughout the inference, so that the tool only fits the kinetic parameters, or they can be used as an initial guidance structure that can be altered by the framework.
Another feature of JCell is the ability to simulate regulatory networks to examine effects of particular components in silico (Streichert et al., 2004). This enables researchers to study regulatory networks and the impact of drug treatment on a computer to reduce expensive ‘wet lab’ experiments. Publications on applications in Artificial Embryology are listed on the website.
One of the main goals was to develop a user-friendly framework that is able to run on most computer platforms without requirements of special hardware. To ensure this independence, JCell is written completely in Java. Furthermore, it can be run using Java WebStart in a standard web browser. JCell comprises a simple and self-explanatory user interface that presents only the most important parameter values. For most of the remaining hidden settings default values that showed good performance in preliminary experiments are used. Nevertheless, all algorithm settings are accessible via configuration files. These files can be either loaded from the interface or committed in the command line for batch mode. This enables experts to further tune the algorithms.
Beside the already implemented ability to import pathways in ASCII, XML and KGML (KEGG Markup Language), we plan comprehensive support of SBML (Systems Biology Markup Language) for the next release of JCell.
|
|
Acknowledgments |
|---|
This work was partially supported by the National Genome Research Network (NGFN) from the Federal Ministry of Education and Research in Germany under contract number 0313323. JCell was supported by the State of Baden-Württemberg with the doIT Software Research Award 2005.
Conflict of Interest: none declared.
|
FOOTNOTES |
|---|
Associate Editor: Chris Stoeckert
Received on April 11, 2006; revised on June 5, 2006; accepted on June 7, 2006
|
REFERENCES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CONCEPTSREFERENCES |
|---|
Bäck, T., Fogel, D.B., Michalewicz, Z. Handbook on Evolutionary Computation, (1997) Oxford University Press.
Blinov, M.L., et al. (2004) BioNetGen: software for rule-based modeling of signal transduction based on the interactions of molecular domains. Bioinformatics, 20, 3289–3291
Dahlquist, K., et al. (2002) GenMAPP, a new tool for viewing and analyzing microarray data on biological pathways. Nat. Genet, . 19, 19–20[Medline].
Kanehisa, M. and Goto, S. (2000) KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res, . 28, 27–30
Kitano, H., et al. (2005) Using process diagrams for the graphical representation of biological networks. Nat. Biotechnol, . 23, 961–966[CrossRef][Web of Science][Medline].
Schmidt, H. and Jirstrand, M. (2006) Systems Biology Toolbox for MATLAB: a computational platform for research in systems biology. Bioinformatics, 22, 514–515
Spieth, C., et al. (2005a) Inferring regulatory systems with noisy pathway information. German Conf. Bioinform, . P-71, 193–203 LNI.
Spieth, C., et al. (2005b) Predicting single genes related to immune-relevant processes. Comput. Intell. Bioinform. Comput. Biol, . 2005, 461–468.
Spieth, C. (2006) Modeling gene regulatory systems. PhD Thesis, University of Tübingen,.
Streichert, F., Spieth, C., Ulmer, H., Zell, A. (2004) How to evolve the head-tail pattern from Reaction–diffusion systems. Proceedings of Nasa/DoD Conference on Evolvable Hardware , pp. 261–268.
Wingender, E., et al. (2001) The TRANSFAC system on gene expression regulation. Nucleic Acids. Res, . 29, 281–283
Wu, C., et al. (2004) GeneNetwork: an interactive tool for reconstruction of genetic networks using microarray data. Bioinformatics, 20, 3691–3693
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  F. Ortega, K. Sameith, N. Turan, R. Compton, V. Trevino, M. Vannucci, and F. Falciani Models and computational strategies linking physiological response to molecular networks from large-scale data Phil Trans R Soc A, September 13, 2008; 366(1878): 3067 - 3089. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||