Integrating Systems Biology Markup Language (SBML) with NEURON

TitleIntegrating Systems Biology Markup Language (SBML) with NEURON
Publication TypeConference Paper
Year of Publication2014
AuthorsBulanova, A. S., Mcdougal R. A., Neymotin S. A., Mutai V. K., Lytton WW., & Hines M. L.
Conference NameSociety for Neuroscience 2014 (SFN '14)
KeywordsSFN, Society for Neuroscience
Abstract

The NEURON simulator software is widely used by the computational neuroscience community for electrophysiology modeling. We have recently extended it to provide support for reaction-diffusion dynamics (RxD class). Many reaction simulations, written for a variety of different cell types in various organs, have been developed in the Systems Biology Markup Language (SBML), a standard XML-based format supported by over 200 software packages. We therefore have developed routines to import SBML simulations into NEURON, enabling NEURON users to import a large number of previously developed cell biology models and use them in computational neuroscience research. These can then combine with NEURON's electrophysiological simulation capabilities, and with other intracellular reaction models written in the new RxD syntax. This allows combined use of 2 extensive collections: the neuroscience ModelDB (http://senselab.med.yale.edu/modeldb/) and the SBML BioModels (http://www.ebi.ac.uk/biomodels-main/) databases. Import and export of SBML models is handled using the libSBML library (http://sbml.org/Software/libSBML). A critical feature is a method to match state variables across the different models and different modeling levels. We also give the user facilities to make sure that semantically identical parameters are identified across models and to manage parameters, regardless of their origin. The importation procedure is as follows: 1) an electrophysiology model is loaded from ModelDB or constructed de novo; 2) NEURON loads SBML model descriptions (and creates corresponding SBMLModelComponent objects) 3) the user identifies semantically identical names across the component model name spaces; 4) Diffusion constants are associated with appropriate SBML states; 5) appropriate RxD objects (rxd.Region, rxd.Species, rxd.Reaction) are instantiated; 6) the user adjusts parameters and makes simulation runs of the model.