OpenCOR

OpenCOR is a free, open-source, cross-platform modelling environment for organising, editing, simulating and analysing biological models expressed in CellML (CellML), an XML-based markup language for encoding systems of mathematical equations (ODEs, DAEs and algebraic relations) that describe cellular and physiological processes. It was created by Alan Garny and Peter J. Hunter at the Auckland Bioengineering Institute (University of Auckland) and grew out of two predecessor tools, COR and OpenCell. It runs on Windows, Linux and macOS, both as a GUI application and from the command line, and is written predominantly in C++.

The application is built as a plugin-based "empty shell" to which capabilities are added by enabling selectable plugins (editors, solvers, views) backed by non-selectable infrastructure plugins. To simulate a model, OpenCOR translates the CellML mathematics to C, compiles it at runtime via LLVM/Clang, and integrates it using solvers from the SUNDIALS library: CVODE for ODEs, IDA for DAEs and KINSOL for non-linear algebraic systems, alongside simpler built-in Forward Euler, Heun and Runge-Kutta methods. Unit consistency across equations is checked during authoring.

Beyond CellML itself, OpenCOR reads and writes SED-ML (Simulation Experiment Description Markup Language) documents to make simulation experiments reproducible, and packages models together with their SED-ML descriptions as COMBINE Archives (.omex). It integrates directly with the Physiome Model Repository (PMR), letting users browse and load hundreds of curated CellML/FieldML models without leaving the tool. This places it in the standards-driven computational-systems-biology ecosystem alongside CellML tooling such as libCellML and, on the parallel SBML side, SBML.

Typical use cases include authoring and curating single-cell electrophysiology and biochemical-pathway models, running and visualising simulations, and exchanging reproducible experiments with collaborators. In graph-data and format-conversion terms its relevance is narrow: it consumes and emits domain-specific systems-biology formats rather than general graph interchange formats, and it does not natively round-trip to SBML (only partial, community-driven CellML/SBML translation exists). Other honest limitations noted by its authors are less complete CellML authoring than the older COR, slower execution than bespoke-compiled predecessors owing to the generic LLVM path, and a focus on single-cell rather than multiscale or spatial modelling.