ABI Meeting Minutes 2008-04-30

Present: Randall Britten, Michael Cooling, Peter Hunter, Justin Marsh, Abhishek Tiwari, Tommy Yu

Apologies: Poul Nielsen

Action items from previous meeting:

  • Item 1: Justin has not sent the email yet, he will do so this week while extending the discussion deadline by another week (now Action Item 1).

  • Item 2: Justin had difficulties building the Mozilla snapshot on Windows, hence the building of PCEnv snapshot was delayed. It will be done this week. (now Action Item 2)

Agenda for this week:

Mike had a few topics to bring up. While at UK, Mike had a discussion with different groups that are keen on using CellML.

Support for Stochastic Modeling

  • Certain groups want to know whether CellML will support stochastic modeling, and they may want to see this feature supported in CellML 1.2.

  • Randall noted that Andrew is working on the core specification that does not exclude stochastic modeling

  • A secondary specification could be defined specifically for stochastic modeling, thus no need to force this requirement onto all conforming modeling software.

  • An option that is worth to exploring is to use a paper that was submitted by the Newcastle group that addresses this issue by extending SBML, perhaps we could use it and adapt it for CellML.

Support for Gene Network Modeling

  • Andrew was dealing with this in CellML while working on his PhD. He noted that the CellML that was generated to represent an instance of this network is too verbose to be of practical use.

  • He ended up creating a custom format that is not as verbose to address his problem.

Java Bridge for the CellML API

  • Mike noted that many people see more tools available for SBML without regard to their quality.

  • He also noted that many groups in UK use Java for the software they develop.

  • If we have a Java bridge for the CellML API, we can enable them to use CellML to facilitate their research.

  • Randall said that Andrew previously did work on Python and Java bindings, specifically some way to automatically generate them whenever API changes; however this remains incomplete.

  • Mike pointed out that another group will finish a graph visualization software by June, and they would also like to have support for CellML. As it is written in Java, they would want to know if Java bindings for CellML will be supported by the API any time soon.

    • Action Item 3: Randall to speak with Andrew about status of this.

Programmatic / Web Service Access to the CellML Repository

  • Some group wants to be able to use some sort of web service type access to the CellML Repository for their work. They would like to query the repository for models, for instance.

  • We do not have support for this exactly.

  • Randall would like to speak to them directly to seek their requirements. Mike will send him their contact information.

Ask Poul about CellML <-> SBML Converter / Mid-language Initiative

  • Sarala has been working on the ontology (specifically using OWL) and CellML metadata, which will come into this.

  • Peter noted that there will be a meeting between himself, Sarala and Matt, and they will discuss some of these issues.

CellML Support in BioNessie

  • Randall pointed out we could just show them around the CellML API to help them get started.

  • Mike noted that they specialize / have support for parameter searching and parallel processing of models, and their models are mostly signal transduction models.

That was the topics of interest from Mike.

PCEnv/CellML API progress update

  • Justin reports nothing major, aside from building snapshots.

  • Randall noted that the new snapshot had much improved MathML rendering, but that there still are at least two issues.

  • Firstly, separate pieces of piecewise expressions are not on separate lines, and have no separator at all

  • Secondly, the square root sign does not stretch, which makes it ambiguous.

  • Solutions include patching Mozilla up, but requiring us to maintain another separate patch, or we wait for them to fix it but fix up our XSLT to include brackets to make the equation not appear ambiguous.

    • Action item 4: Justin to add a tracker item relating to this.

  • There was discussion to whether to support replacing the string 'alpha' with the actual Greek letter. In COR there was this feature and it was up to the user to turn it on.

PMR 2 Update

  • Tommy has been working with rdflib and figuring out SPARQL

  • He hopes to have the searching exposed model interface done by next week to finish off the prototyping stage.