Cellular open resource (COR): a new environment for cellular and multicellular modelling of cardiac electrophysiology

University of Leeds (2002) J Physiol 544P, S075

Demonstrations: Cellular open resource (COR): a new environment for cellular and multicellular modelling of cardiac electrophysiology

A. Garny, P.J. Noble, P. Kohl and D. Noble

Department of Physiology, University of Oxford, Oxford OX1 3PT, UK

View other abstracts by:


Computer models have become an invaluable tool for both experimental and theoretical research into cardiac electro-physiology. These models have, over the years, matured from simple representations of action potential shapes to mechanistic models with developing predictive power. Theoretical work on the sodium-calcium exchanger has, for instance, predicted its stoichiometry some five years before its experimental confirmation (for details, see Noble, 1995). The iterative interaction between simulation and experimental work is a vital part of model development (Noble & Rudy, 2001). OxSoft Heart (Noble et al. 1998) offers an environment for such an iterative process and has for more than 15 years been widely used as a research tool, as well as a teaching tool. It has, however, been recently discontinued, due mainly to technical issues that prevent DOS-based programs written in Borland Pascal from working on the latest generation of processors. COR (Cellular Open Resource) aims at pursuing the approach taken by programs like OxSoft Heart, while providing an up-to-date environment for cellular and multicellular modelling of cardiac electrophysiology.

Currently, a number of academic groups offer interfaces for cardiac modelling (both at the cellular and multicellular levels). LabHEART (Puglisi & Bers, 2001) and iCell (Demir, ssd1.bme.memphis.edu/icell) provide advanced ionic cell modelling, but they are restricted to a limited number of single-cell models. Both have static interfaces, which make for useful teaching tools (specific interfaces help the setting of tasks), but do not offer enough flexibility for research. Software packages like Cell EditorTM (Physiome Sciences, Inc., www.physiome.com) are more appropriate for the latter, since they allow for the refinement and development of new single-cell models. Cell EditorTM, however, does not currently support multicellular modelling. Also – it is proprietary software. Cellular and multi-cellular models can very appropriately be implemented using CMISS (Hunter, www.cmiss.org). However, CMISS is a package for the advanced user, requires significant training, and is best applied to complex problems involving structural electro-mechanics models.

Based on the above, COR has set out to provide a package that is simple to use, both as a research and teaching tool. It aims at offering the same level of flexibility as programs like OxSoft Heart (voltage and current clamps, I-V curve, ability to pause and resume simulations for interactive parameter changes, etc.). Unlike most of the current interfaces, models are not hard-coded, but stored in files using the CellMLTM format (Hedley et al. 2001). This provides COR with an ‘out of the box’ access to a large database of cardiac single-cell models, such as sino-atrial node pacemaker cells (Zhang et al. 2002), atrial myocytes (Nygren, 1998), Purkinje fibres (Varghese & Winslow, 1994), and ventricular myocytes (Luo & Rudy, 1994; Noble et al. 1998), as well as the ability to share new models with other CellMLTM-compliant programs like Cell EditorTM and CMISS. Models are parsed for correctness before being converted into machine code for the Intel x86 family of processors. The computation of the model is therefore optimised, which is essential considering that COR supports the simulation of multicellular problems (which are intrinsically computationally demanding). Unlike CMISS, COR uses a finite difference technique, instead of a finite/boundary element method, which demands less user effort in setting up the problem. 1D and 2D problems correspond to a strand and a tissue section of single cells, respectively. 2D tissue architecture can be freely defined or inherited from histological samples using bitmaps to define a particular mesh. Individual cells properties and intercellular coupling can be assigned to pre-defines models or values, and they can be interactively modified. This provides an easy to use, yet powerful, environment for cellular and multicellular simulations.

Last but not least – COR is freely available (source code included) from cor.physiol.ox.ac.uk




Where applicable, experiments conform with Society ethical requirements.

Site search

Filter

Content Type