Platform 5: Software & Modelling

Principal Investigator

Professor Peter Hunter, University of Auckland

Associate Investigators

Dr Koray Atalag, Dr Chris Bradley, Dr Jagir Hussan, Dr David (Andre) Nickerson,
Professor Poul Nielsen, Dr Soroush Safaei

ABI Software Team

Dr Richard Christie, Dr Alan Garny, Hugh Sorby, Alan Wu, Tommy Yu, Noel Zeng


This Platform is developing the software programmes and modelling frameworks that underpin the biophysically based modelling of the MedTech CoRE. These tools and services are used by all Flagships. Much of the work in this Platform is also linked to the international Physiome Project. There are five parts to Platform 5, briefly described here along with their current status.

1. Physiome standards

A logo for the Physiome Project. A human figure standing mid-stride, with a DNA strand trailing behind the figure.Biophysically based models of physiological processes capture physiological phenotype in a quantitative and predictive form. They underpin research in physiology in the same way that nucleotide sequence underpins research in genetics, since there is no other way to deal with the complexity of physiological systems in a quantitative fashion. Most importantly, biophysical models bring with them the constraints on physiological behaviour imposed by the conservation laws of physics. Physiological function depends on these laws of nature as much as it depends on molecular biology. To ensure that biophysically based models and data are encoded in a reproducible fashion, we are developing three XML standards: CellML (for biophysical models based on ordinary differential and algebraic equations), FieldML (for partial differential equations and spatially varying models) and BiosignalML (for time-varying signals). Another standard (SED-ML) is being developed in collaboration with the molecular systems biology community to encode the simulation protocols – again to ensure reproducibility of stated results. A website has been developed to support the modelling community using these standards.(

2. OpenCOR software for CellML models

A software package called OpenCOR (available from is being developed to create, visualise and run CellML models. The interface is shown below.

A screen shot of OpenCOR user interface.

The OpenCOR editing environment allows the creation and editing of complex mathematical expressions, as illustrated here.

A screenshot of a complex mathematical formula in OpenCOR viewer.

A repository for models and data, based on the CellML standard, called the Physiome Model Repository or PMR, has been established ( A tutorial on the use of OpenCOR for models encoded in CellML and SED-ML and stored on PMR is available at

3. OpenCMISS software for FieldML models

At the heart of physical laws is the concept of continuous fields - and much of physiology is the interaction between these fields and the molecules that support life. Examples are the relationship between a continuous temperature field and the energy of its constituent particles, or a pressure field representing the energy density (J/m^3) of energetic particles, or the oxygen concentration field in a tissue and its relationship with oxygen sensitive proteins. The ‘particles’ in physiology are molecules (proteins, carbohydrates, lipids) and of course the complexity of their formation, structure and interaction is what biomedical science and drug discovery is all about. Physiological function and the symptoms of disease, however, appear at the tissue/organ scale and the field laws of physics are an essential component of the physiological phenotype. There are only four equations, or equation systems, needed to capture the laws of physics (conservation of mass, linear and angular momentum, and energy, respectively) at the scale of continuous fields. These are the Navier-Stokes equations (dealing with fluid mechanics), the equations of finite elasticity (dealing with the mechanical behaviour of solid materials), Maxwell’s equations (dealing with the behaviour of electro-magnetic) and reaction-diffusion equations (dealing with conservation of mass for chemical species represented by their concentration fields). We are developing a modelling standard (FieldML) and software libraries that support both simulation and visualisation based on this standard (OpenCMISS:Iron and OpenCMISS:Zinc). Some examples of FieldML models being used in the MedTech CoRE are shown below. See for details.

A composite image of various models of human organs - the heart, lung, hand, thigh and baby.

4. User interfaces for MedTech projects

To facilitate the use of the Physiome Project standards for the modelling activities in the MedTech CoRE, we are developing a number of reusable modules and Graphical User Interfaces (GUIs) that link in the standards-based simulation and visualisation libraries described above. These components are designed such that application developers are provided with high level abstractions to process medical images, medical data and simulation algorithms. These components are organised as software modules and GUI plugins. A schematic of this framework is shown below.

An architecture diagram of MedTech User Interface projects.

The MedTech CORE Software library is a collection of software modules that abstract high level processes such as segmentation, mesh generation, simulation etc. These modules are mainly built using OpenCMISS libraries and other open source libraries like ITK (which provides methods for image registration, segmentation and filtering), VTK, CGAL etc. (which provide meshing capabilities). Modules do not directly interact with the user; rather they enable plugins with processing capabilities. The application framework consists of plugins that implement workflows by linking user interactions with modules and other plugins. Plugins use the QT widget framework for GUI and the Common Toolkit (CTK) for their lifecyle management and inter-plugin communication. The framework enables developers to build applications by creating workflows using existing plugins or adding new plugins and modules as required. Each development contributes to the MedTech CORE software ecosystem and maintains a consistent user experience across applications. Developers can also build applications that do not use plugins and the CTK framework rather just the capabilities offered by the modules.

5. Medical informatics and Electronic Health Records

The modelling standards described above include metadata that provides the biological and biophysical meaning to the mathematical terms in the models. To use these models as descriptions of physiological phenotypes in order to interpret clinical measurements, we need to map the CellML and FieldML model metadata to Electronic Health Records (EHRs) and in particular the openEHR standard ( An exploratory project is now underway using openEHR Archetypes ( to achieve this mapping. Implementation of modelling workflows with clinical data in a hospital setting is being implemented via the VPH-Share project (

The objective is to represent measurement data and associated clinical information using Archetypes and then map to CellML and FieldML parameters for model validation and to create next generation of personalised and predictive decision support tools at the bedside.

Archetypes allow for capturing of clinical data and context as an indivisible whole so that the exact meaning can be preserved and safely acted upon across different systems. The following example illustrates the blood pressure measurement archetype comprising:

  1. Data: holding the actual measurement data (e.g. systolic and diastolic blood pressure in mmHg with max and min allowable values)
  2. Protocol: holding protocol of the measurement such as cuff size (e.g. adult, child)
  3. State: holding state information such as patient position (e.g. lying, sitting)
  4. Events: depicting whether it is a one off measurement or 24 hour average
A map of ideas related to Blood Pressure.

By using shared biomedical ontologies (e.g. FMA, Gene Ontology) and clinical terminology (e.g. SNOMED CT, LOINC) to annotate both computational and clinical models, automated reasoning and new knowledge discovery will be possible which will ultimately help reach the goals of the Physiome Project. The project will extend PMR to extract and store archetype meta-data which will allow searching of both models and associated clinical data.