Bioinformatics Advance Access originally published online on June 22, 2007
Bioinformatics 2007 23(14):1859-1861; doi:10.1093/bioinformatics/btm231
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Stochastic simulation GUI for biochemical networks
1Keck Graduate Institute, 535 Watson Drive, Claremont, CA 91711, USA and 2Department of Bioengineering, University of Washington, Seattle, WA 98195, USA
*To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT 1 INTRODUCTION2 CAPABILITIES3 FUTURE DIRECTIONSAcknowledgementREFERENCES |
|---|
Motivation: This article describes the development of a useful graphical user interface for stochastic simulation of biochemical networks which allows model builders to run stochastic simulations of their models and perform statistical analysis on the results. These include the construction of correlations, power-spectral densities and transfer functions between selected inputs and outputs.
Availability: The software is licensed under the BSD open source license and is available at http://sourceforge.net/projects/jdesigner. In addition, a more detailed account of the algorithms employed in the tool can be found at the Wiki at http://www.sys-bio.org/sbwWiki.
Contact: rrao{at}kgi.edu
Supplementary information: Supplementary data are available at Bioinformatics online.
|
1 INTRODUCTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CAPABILITIES3 FUTURE DIRECTIONSAcknowledgementREFERENCES |
|---|
The complexity of biochemical networks arises from the nature of the interaction of various biochemical species. When large numbers of participants are present, their time evolution can be described by a system of differential equations. However, in cases where the species numbers are low, this has to be replaced by a stochastic approach. A number of algorithms have been developed in the recent years for such problems, improving over the original Stochastic Simulation Algorithm (Gillespie, 1977) and have been incorporated into tools such as COPASI (Hoops et al., 2006), Dizzy (Ramsey et al., 2005) and BioNetS (Adalsteinsson et al., 2004). While most of these tools can be used only for simulation, a few like BioNetS allow construction of probability density histograms and power spectral densities. The application described in this article enables additional analysis capabilities such as construction of correlations between species, power spectral densities as well as transfer functions.
|
2 CAPABILITIES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CAPABILITIES3 FUTURE DIRECTIONSAcknowledgementREFERENCES |
|---|
The user interface and the Gillespie simulator it interacts with were built in C# using the Systems Biology Workbench (Sauro et al., 2003). Users can perform the following tasks, as shown in Figure 1.
|
Data generation: Users can load biochemical models in SBML to perform stochastic simulation. Data can be generated on an evenly spaced time-grid or by updating the species numbers by a specified value. The first option enables users to carry out statistical analysis, compute Power Spectral Densities as well as Transfer Functions.
Probability density: Each data run is converted to a histogram representative of the spread. Averaging these over all the generated runs yields a histogram of the ensemble averaged probability.
Ensemble averages: Population means and standard deviations for all time points are obtained by averaging the entire set of runs. These can be compared with deterministic results for the same model.
Correlations: For data generated using an evenly spaced time grid, auto and cross-correlations (Chatfield, 2004) are computed between all the species in the network for each run as well as the ensemble.
Power spectral densities: PSDs can provide information on noise filtering characteristics of biochemical networks (Simpson et al., 2003). These are computed using the publicly available Exocortex.DSP signal processing toolbox (Exocortex, 2003).
Transfer Functions: These are response functions that relate system outputs to inputs (Williams et al., 1972).
Creating noise sources: This tool allows converion of boundary nodes into floating nodes by means of a SBMLModifer module that interfaces with libSBML (http://www.sbml.org/software/libsbml/).
Testing application: This is a companion tool that has been built to allow users to test the correctness of the simulator implementation and is based on the discrete stochastic model testing suite (http://www.calibayes.ncl.ac.uk/Resources/dsmts/overview).
|
3 FUTURE DIRECTIONS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CAPABILITIES3 FUTURE DIRECTIONSAcknowledgementREFERENCES |
|---|
While the present version allows data generated by the simulator to access the analysis capabilities, future versions could allow experimental data to be loaded into the application to carry out the same analysis for comparison with model results.
|
Acknowledgement |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CAPABILITIES3 FUTURE DIRECTIONSAcknowledgementREFERENCES |
|---|
RRV would like to acknowledge Frank Bergmann for assisting with software and Vijay Chickarmane and Alpan Raval for discussions on stochastic modeling of biochemical networks. RRV and HMS are grateful to generous support from the US DOE GTL Program.
|
FOOTNOTES |
|---|
Associate Editor: Olga Troyanskaya
Received on December 16, 2006; revised on April 17, 2007; accepted on April 26, 2007
|
REFERENCES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 CAPABILITIES3 FUTURE DIRECTIONSAcknowledgementREFERENCES |
|---|
Adalsteinsson D, et al. Biochemical Network Stochastic Simulator (BioNetS): software for stochastic modeling of biochemical networks. BMC Bioinformatics, ( (2004) ) 5, : 24.[CrossRef][Medline].
Chatfield C. The Analysis of Time Series: An Introduction. Chapman & Hall, London, ( (2004) )..
Exocortex DSP. An open source C# Complex Number and FFT library for Microsoft .NET. ( (2003) ) http://www.exocortex.org/dsp/..
Gillespie DT. Exact stochastic simulation of coupled chemical reactions. J Phys. Chem, ( (1977) ) 81, : 2340–2361.[CrossRef][ISI].
Hoops S, et al. COPASI a COmplex Pathway SImulator. Bioinformatics, ( (2006) ) 22, : 3067–3074.
Ramsey S, et al. Dizzy: stochastic simulation of large-scale genetic regulatory networks. J. Bioinform. Comput. Biol., ( (2005) ) 3, : 415–436.[CrossRef][Medline].
Sauro H, et al. Next generation simulation tools: the Systems Biology Workbench and BioSPICE integration. OMICS, ( (2003) ) 7, : 355–372.[CrossRef][Medline].
Simpson ML, et al. Frequency domain analysis of noise in autoregulated gene circuits. PNAS, ( (2003) ) 100, : 4551–4556.
Williams WJ, et al. Biological system transfer-function extraction using swept-frequency and correlation techniques. Med. Biol. Eng. Comput., ( (1972) ) 10, : 609–620..
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. W. Evans, C. S. Gillespie, and D. J. Wilkinson The SBML discrete stochastic models test suite Bioinformatics, January 15, 2008; 24(2): 285 - 286. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||