Forward Looking Session - The Future of Model Sharing

This panel will discuss the complex technical, sociological and societal challenges to realistic sharing of mathematical and computational models. The discussion will include the review of model sharing efforts of the IMAG Multiscale Modeling (MSM) Consortium over the last three years, a summary of outcomes from the MSM 2007 Model Sharing workshop, and highlight efforts that are developing and efforts that have achieved successful outcomes. This panel will also discuss the future of journal requirements for publications to facilitate model reproducibility.

1. Grace Peng (5 mins) - Introduction to the Panel

Theme 1 - Model Sharing discussions, tools and standards

2. James Glazier (10 mins), summary of 2007 MSM Model Sharing Strategies - One Day Mini-Workshop

3. Jim Bassingthwaighte (15 mins), WG10 discussions and Herbert Sauro, Standards

Theme 2 - Sharing platforms, success stories, sociological discussions

4. Karin Johansson (10 mins), EC Community Codes‎

5. Taishin Nomura (10 mins), Japanese Community Codes, summary report here Japan

6. Peter Hunter and Frank Sachse (20 mins), CellML Dialogue

7. Joy Ku and Trent Guess (20 mins), SIMTK Dialogue

Back to Overall SIAM Schedule

Abstracts

James A. Glazier , Maciej Swat

Biocomplexity Institute, Indiana University

MSM 2007 Model Sharing Workshop

Model Sharing Mini-Workshop organized by Multi-Scale Modeling consortium in April 2007 covered a range of topics related to the use of markup languages (MLs) and scripting languages in model definition and description, the role of standards and ontologies in model building, maintenance and curation, adherence to modeling standards in research, and the role that standards play in enabling model sharing and model integration. The goal of the workshop was to define best strategies leading to the creation of widely usable model description standards, better model shareability and exchangeability, and increased awareness of benefits coming from standards compliance.

In this talk we will review major issues associated with model sharing and model definitions and revisit main recommendations made by workshop participants concerning implementations of standard-compliant computational modeling tools

James B. Bassingthwaighte, and Gary Raymond

University of Washington, Seattle

Toward Model Sharing via Modular Modeling and Standards for Reproducible Modeling

The establishment of a database of models which can be run over the web or downloaded to platforms anywhere represents the current effort of personnel at the National Simulation Resource for Circulatory Transport, Metabolism and Exchange to make it easy to share and to reuse developed models. REinvention, so often the hallmark of REsearch, is wasteful. A model at www.physiome.org can be modified over the web to create a new model and then downloaded to a users laptop to develop further. While we are pushing to raise our own standards for writing clear, definitive, meaningful models, and to give full credit for the origins of these models, we are also pushing to have other investigators present complete models on which we might build. At the same time, we offer our resource to researchers worldwide to deposit their models at www.physiome.org for public dissemination; this will help fulfill the funding agencies desires for making data and models available. To further this purpose we have succeeded in devising a method for bring models together in an automated way so that modules can be substituted for one another in a composite model. The next step is to make JSim's MML automatically convertible into CellML and SBML repositories for broader distribution or to develop a repository that includes both of those and also handles PDEs and ODEs to broaden the range of applications to biology.

Joy P. Ku¹, Randy Radmer¹, Simbios Staff¹, Jeanette Schmidt¹, Russ Altman¹, Scott Delp¹, Trent Guess²

¹Simbios, Department of Bioengineering, Stanford University, ²Department of Mechanical Engineering, University of Missouri - Kansas City

Simtk.org: Using Technology to Encourage the Sharing of Research Tools and Data

Simbios (http://simbios.stanford.edu/), one of the National Centers of Biomedical Computing, has developed Simtk.org (http://simtk.org/), a web-based tool designed for the sharing of simulation software tools and models. We will present some of the issues that researchers have about sharing their results and discuss how Simtk.org addresses these issues, providing perspectives from both users and developers. Media:SIAM2008ModelSharing_KuGuess.ppt

Taishin Nomura, Yoshiyuki Asai, Kenichi Hagihara, and Yoshihisa Kurachi

The Center for Advanced Medical Engineering and Informatics, Osaka University

An open platform for physiome and systems biology: insilicoML and IDE as a CellML compatible model sharing environment

The number of mathematical models describing biological functions published in peer reviewed journals and complexity of each of those models are rapidly increasing, leading to difficulties in reproducing simulated dynamics of the published models for reuse. As a solution to this problem and to promote sciences and the knowledge integration, under MEXT funding, we have been developing a model description format (insilicoML or ISML) representing hierarchical/modular structure of biological entities and functional operations of one module onto others. ISML forms a triplet (ISML, ISDB, ISIDE) at http://www.physiome.jp with ISDB for model database and ISIDE for constructing and simulating models of physiological systems. ISML is compatible and complementary with CellML of IUPS Physiome project, the pioneering effort to overcome the problem. The triplet can thus be enhancing the stream of model sharing and contributing to diversity of building blocks useful to develop in silico human.

Peter Hunter¹, Catherine Lloyd¹, Poul Nielsen¹ and Frank Sachse²

¹Auckland Bioengineering Institute, University of Auckland, New Zealand, ²Nora Eccles Harrison Cardiovascular Research and Training Institute, University of Utah

The XML markup language CellML is designed to help with model reuse by (i) providing a precise and unambiguous definition of a model (equations, units and biological meaning via ontological terms in the metadata), and (ii) allowing model components to be imported. This discussion between a developer and a user of CellML will address the challenges of model reuse in practice.