Scientific programme

Welcome and workshop overview

Poul Nielsen, The Auckland Bioengineering Institute


Following on from the success of the inaugural CellML workshop in 2007, the aim of this second meeting is to bring together the growing community of CellML developers and users, to discuss recent developments and future plans, and to provide an opportunity to update the community on ideas which were discussed at the last workshop. In particular, I will provide and overview of the some of the work which has occurred at the Auckland Bioengineering Institute over the past year; including changes in the CellML specification and tool development, and I will also introduce some topics which have involved the broader CellML community.


CellML-compatible Integrated Development Environment for Multilevel Modeling of Biophysical Function and Database

Yoshiyuki Asai, Osaka University

It is important to develop dynamical model of biophysical functions with multilevel structure. There proposed several languages, such as CellML, to describe such models including not only mathematical equation but also related meta-information. If the model has a large scale with multilevel structure, it is difficult to write down the model by hand on, for example, a text editor. We are developing an integrated development environment (IDE) of the models with graphical user interface providing intuitive handling of each component in the models. The IDE can import and export models described in its own format as well as in CellML format. We also try to develop database of models built up on the IDE which can be accessed via the Internet.


openCMISS and FieldML/CellML

Chris Bradley, University of Oxford

openCMISS is the open source distributed memory redevelopment of the computational engine CMISS. openCMISS solves bioengineering problems using finite element analysis, boundary element and collocation techniques. This talk will focus on how openCMISS interacts with FieldML in order to solve partial differential equations and how openCMISS interacts with CellML in order to define source terms for these equations.


Modular CellML models with examples

Mike Cooling, The Auckland Bioengineering Institute

Making CellML models available is an important part of model communication, but does not necessarily lead to easy reuse. Recently a number of concepts designed to help convert biological pathway information into modular CellML models have been advocated. These are illustrated in the context of signal transduction. Modularity is discussed from the point of view of pathway construction and the replication of experimental results via the application of virtual protocols in a key pathway in cardiac disease.


FieldML: data structures for modelling

Richard Christie, The Auckland Bioengineering Institute

Physiology simulations are amongst the most complex of numerical analyses. They involve smooth “organic” geometries, anisotropy, spatial variation of numerous material properties and cell types, large deformations and all other types of non-linearity. Model sizes can become very large, requiring techniques such as embedding and adaptive meshing to reduce solution degrees of freedom.

Modelling software developed at the Bioengineering Institute deals with this complexity by expressing all problem variables as fields, with support for specialised and high order basis functions and embedding of one field in another. This talk will outline existing field representations, and will explain the main concepts, data structures and feature requirements we intend to support in future. We call this developing data specification FieldML, which will ultimately be the XML language used to serialise these field representations.


Visualization of CellML models

Sarala Dissanayake, The Auckland Bioengineering Institute

The aim of this project is to build a framework to visualize CellML models. This involves:

  • developing an ontology which could be used to represent the biological and biophysical concepts that are modeled in CellML models;
  • developing a set of rules for binding these biological and biophysical concepts to CellML models;
  • implementing a visual language that can be used to represent all biological  processes;
  • developing a specification for building visual templates that support this language and the rules for binding them to biological and biophysical concepts within the ontologies;
  • developing a visual editing tool, that combines the visual language and the ontologies visualize CellML models.


Excitable cell and tissue modelling using CellML/FML

Socrates Dokos, The University of New South Wales

Representation of excitable cell models in CellML allows for the ready fitting of new and existing models to user-supplied experimental data. In addition, a variety of cell models can be easily incorporated and exchanged into whole-tissue models to study the influence of specific cell models on spatial propagation phenomena. In this presentation, we describe results of excitable cell fitting using a novel CellML parameter optimisation tool, as well as our recent work in two markup field representation languages, ModelML and FML. ModelML is focused on the relational information between field information such as geometrical data to the relevant mathematical model stored as CellML. These relations can be categorized into groups such as point, boundary and subdomains. In contrast, FML is responsible for storing geometrical or field information. Currently, our focus is on boundary representation using Bezier/BSpline curves and surfaces and using mesh representation to characterise complex biological objects. Examples of CellML model optimisation will be given for various cardiac pacemaker models, as well as their incorporation into FML tissue descriptions. Ensuing model equations are solved for using the COMSOL Multiphysics finite-element package.


Moving from COR to PCEnv/COR

Alan Garny, University of Oxford

Cellular Open Resource (COR) is an environment for the editing and execution of models described in the CellML format, and so is Physiome CellML Environment (PCEnv). Both environments have very similar goals and it was thought best for the two of them to be merged into one single environment (PCEnv/COR). As part of this process, we decided to revisit the approach taken by both COR and PCEnv, and came up with a possible new graphical user interface for PCEnv/COR.


In pursuit of a working model: the curation process

Catherine Lloyd, The Auckland Bioengineering Institute

CellML evolved as a solution to the problem of a lack of consistency between the code modellers use to develop and run a computational model, and the text and equations in which the same model is published. To emphasise this problem, there are currently 300 models in the CellML model repository, and we have yet to directly translate a published paper into a working CellML model. Type errors, lack of unit consistency, and missing equations and/or parameter values are just some of the problems we face. Model curation is an ongoing process, and our ultimate aim is to have a CellML model recreating the published results. For this we employ a variety of tools, and where possible, we seek the help of the original model author. As the CellML community grows, so too does the general awareness of the CellML project. Models are just starting to be published in press concurrent with the CellML source code, but we still face the considerable task of curating the existing repository.


The Physiome CellML Environment: status update

Justin Marsh, The Auckland Bioengineering Institute

The Physiome CellML Environment (PCEnv) is an integrated tool for editing and simulating CellML models. Over the past year, PCEnv has gained full model editing capabilities, as well as a number of key usability and performance improvements. In addition, key services in the CellML API, which are used by PCEnv, have been improved, allowing a wider range of CellML models to be correctly processed.


The future of the CellML specification

Andrew Miller, The Auckland Bioengineering Institute

There have been efforts to start development of the CellML 1.2 Specification. I will describe the process by which we have started developing the CellML 1.2 Specification, and how community members can get involved in this process. In addition, I will briefly highlight some of the key proposals that are under consideration for inclusion in the CellML 1.2 Specification.


How to describe mathematical models of cellular physiology?

David Nickerson, National University of Singapore

This question is what I will be exploring during this presentation. How much and what  detail is required to provide a complete description of a model such that it can be both  human and machine interpretable? What technology is best suited to such descriptions? How can new implementations of existing models or even new instantiations of existing implementations be quantitatively validated? These are a few of the issues we have been considering recently. In this presentation I will be discussing these issues and presenting some initial progress we have been making in trying to find a viable solution.



GIMIAS – Graphical Interface for Medical Image Analysis and Simulation

Maarten Nieber, Universitat Pompeu Fabra, Barcelona

In the past few years a considerable amount of work has been put to develop new disciplines and extend existing ones with the interest of reproducing in silico the pathophysiological behaviour of the human body (e.g., medical imaging, medical image analysis, computational models for biology and physiology, computational biomechanics, etc.). Still,  there is a lack of software tools that allow specialists (e.g., from research or medicine) to integrate all these methodologies in a single platform. The current work presents an overview of GIMIAS, a Graphical Interface for Medical Image Analysis and Simulation. This framework is being developed as an open source effort that intends to fill this gap. GIMIAS has been build over widely used open source C++ based frameworks like The Visualization Toolkit  (VTK, suported by Kitware Inc.), The Insight Toolkit (ITK, also supported by Kitware Inc.), The DICOM Toolkit (DCMTK, supported by Offis in Germany), Medical Imaging Interaction Toolkit (MITK, developed at Division of Medical Informatics, Deutsches Krebsforschungszentrum (DKFZ), Germany) among others. By developing these platform as an open source software project, we intend to encourage the collaboration between governmental, industrial and academic institutions in a normalized way for developing their business and technologies.


A database approach to model analysis: methods and challenges

Steve Niederer, University of Oxford

I will discuss the recent progress at the University of Oxford in developing an tools for performing multiple virtual experimental protocols on multiple models encoded in CellML. This project has identified gaps in model and experimental protocol definitions that we hope will motivate the development of community standards to allow extensive comparison of models within the ever growing CellML repository.


First things first: Some practical thoughts about mathematical modeling in biology

Jose Puglisi, University of Chicago

In this talk I will cover some practical issues that newcomers encounter as they incursion in this novel discipline. Also I will outlined the academic profile of the researcher interested in applying mathematical techniques to a biological problem. It will be aimed to an audience of novel students or researchers that are starting in this field.


Methods for semantic cell modelling and convenient simulation

Shimayoshi Takao, Kyoto University

In this talk, three topics are first outlined; current Kyoto model, a semantic model editor, and a description language for experimental protocols. Next, a method for setting up cell simulations by analyzing model equations based on graph theory is presented.


Combining models using CellML: rat cardiac excitation-contraction coupling as an example

Jonna Terkildsen, The Auckland Bioengineering Institute

Using CellML 1.1, we will demonstrate the process of formalised model reuse by combining three separate models of rat cardiomyocyte function (an electrophysiology model, a model of cellular calcium dynamics, and a mechanics model) into one integrative model of rat electromechanics.  During this process, we will highlight and classify the challenges associated with combining models, and provide some suggested solutions.


CellML model development with version control system

Tommy Yu, The Auckland Bioengineering Institute

With CellML models becoming increasingly complicated, with importable components from various sources being made available from within the same model or elsewhere. Sometimes a model developer may want to work with a specific component, only to later find out they made a change that broke some models that import this component. The modeler may wish there was a time machine that would let her go back in time to see what might have gone wrong – enter a version control system.  In this talk we will discuss developing CellML models using a version control system, or source code management system. We will look at what Mercurial is, and how that fits into the CellML repository design.