ABI Meeting Minutes 2008-02-27

Present: Peter Hunter, Tommy Yu, Randall Britten, Andrew Miller, Catherine Lloyd, James Lawson

Apologies: Mike Cooling, Poul Nielsen

Review of last week's action items:

1.) "Justin will begin the process of addressing this issue by writing a proposal to update the Graphing and Metadata Specification."

  • Justin has started work on this, was unsure whether the version of the Draft Graphing and Metadata Specification on the cellml.org website is the latest version.

  • Andrew said that it probably is, but that it might be worth emailing David Nickerson to see if he has a more recent version.

Action Item 1.) Justin will add this issue to the tracker and continue his work on this.

2.) "Catherine will work with John to help fix the CellML model, and then put it up on the repository"

  • John has been working on getting his model working in PCEnv and COR as well as JSim. Catherine said the issue appears to be that JSim is more lenient on extraneous equations than PCEnv or COR.

3.) "Andrew said that while the 1.2 features obviously aren't fully defined yet, he will be happy to do this"

  • Andrew has created "What's new?" to describe the differences between both CellML 1.0 & 1.1 and the proposed differences between CellML 1.1 & CellML 1.2

4.) "Peter will talk to Mike regarding IP"

  • Issue resolved.

Agenda for this week:

Agenda Item 1 - Catherine: "Taishin Nomura and Yoshihisa Kurachi (from Osaka) are visiting the ABI, 28th & 29th February"

  • We have the level 10 seminar room booked from 10am-1pm for presentations and discussions from our guests.

  • Their plane gets in at 9:20am so this booking may have to be reassessed.

Agenda Item 2 - Catherine: "Posting rights to the team@cellml.org mailing list for Zoar Engelmann (ABI,) John Davidson (ABI) and Penny Noble (Oxford).

  • The group consents to posting rights for these individuals.

Action Item 2.) Andrew will update the posting permissions for team@cellml.org

Agenda Item 3 - Randall, James: "Curation - prioritising models to work on"

  • James and Randall have been discussing the repository and curation priorities - this issue has been raised previously but new perspectives were included.

  • The issue of which models should be curated, in what priority was raised.

  • James noted that approximately 50% of the models in the repository are now functional to some extent.

  • Randall suggested that at some point, we will need to decide whether it is worth the effort to fix difficult models or whether the repository would be better served by diluting these non-functional models by adding new models that do work.

  • James asked whether the ultimate aim of the repository is currently to provide a proof-of-principle for CellML, or to provide a comprehensive data resource?

  • Peter believes that it should really be the former, with the goal of bringing more people on-board into the community.

  • Eventually, once we have built a strong, model contributing community, we will be able to bill ourselves more effectively as a data resource

  • The outcome of this discussion was essentially that we should continue to hone the models we currently have, and add new models mainly where we are requested to do so by the authors of those models, and those models written by groups we work with.

    • While we are working with an author on a specific model, we should try and curate all of the models by that author.

  • James noted that this is what he is primarily working to achieve - particularly with the cardiac electrophysiology models.

  • Peter emphasised his view of the importance of session files with the SVG-Javascript technology, particularly for commonly used models.

    • Justin is working on a tool to allow automation of creating such sessions, but has so far spent little time on it.

    • James noted that it is actually easier and less time-consuming than he previously thought to create these files manually, but there is still a high barrier for casual users to do this.

Agenda Item 4 - Catherine: "'What's New?' webpage"

  • This item refers to the recently created webpage describing the proposed new features for CellML 1.2

  • This article, which was recently created by Andrew, is currently within the 'specifications' section of the website.

  • David Nickerson has raised some objections via the tracker, and has made some changes to this page. He suggests that this page be a wiki page rather than a page under the specifications section.

Agenda Item 5 - Randall: "Proposal handling, tracker item 347"

  • David Nickerson supports linking items within the page describing what will be new in CellML 1.2 to their corresponding tracker items.

  • David suggested in the discussion of this tracker item that it would also be good to link these items to the appropriate sections of the draft CellML 1.2 specification.

  • He also suggested we need to change the title of this page from "What's new?" to something like "Features proposed for CellML 1.2"

    • Action Item 3.) Catherine will change the name to something more appropriate

  • Workflows for creating proposals were discussed

    • We should use the tracker for submitting, discussing and voting on proposals.

    • We could use a form similar, for example, to the form used for extracting information from tracker issue submitters about curation issues.

"Making decisions on proposals"

  • This related discussion dealt specifically with the CellML specification and the development thereof, and ideas for processes of government of the community (if this is indeed required.)

  • Having discussed the use of the tracker for developing and discussing proposals, the issue of voting on contentious issues within such proposals was discussed.

  • Bugzilla provides a voting system which we have not used to date - issues have generally been discussed on the mailing lists and in meetings and decisions arrived at by consensus.

  • The increasing size of the CellML community raises the prospect that we should work to create a fair and efficient democratic process for making decisions on difficult issues.

  • It was proposed that where possible, the rule of majority should be respected.

  • However in some cases, it will certainly be valuable to lend increased weight to certain individuals' views, where those individuals have a high stake or expertise in the issue. This idea warrants further discussion - it appears that we will need some test cases to judge the merit of vote-weighting. This is also a politically sensitive issue - ala "All animals are equal but some animals are more equal than others"....

  • It was proposed that in the case that decisions can not be arrived at by voting, the CellML community should appoint a panel comprising a small, odd number of people representing the various interests within the CellML community who will weigh the mood and attitudes of the community and make decisions.

    • Peter suggested this panel might consist of David Nickerson, Alan Garny, Poul Nielsen, Randall Britten and another member.

  • In the case that a decisions still can not be reached by the panel, the group suggested that Poul could make the decision unilaterally.

  • James suggested that as the community grows, and specifically, as the amount of funding directed at CellML and related projects grows, we may need some sort of constitution. Currently, however, we should try and keep the process simple.

Agenda Item 6 - Peter: "3 tier databases - discussions with Shankar Subramanium"

  • Shankar has been behind the IT infrastructure development of many large bioinformatics projects, including the Alliance for Cellular Signalling project. He has recently visited Auckland and has had some interesting discussions with Peter.

  • One of Shankar's suggestions was that the CellML PMR should be a 3-tier database.

  • This architecture has been used extensively in Shankar's projects and he is convinced that it will help us enormously when scaling up the repository with respect to the data it will contain and increasing numbers of users.

  • We will need to tailor our metadata systems to this end.

  • Tommy believes that we are already doing this to some extent.

  • Action Item 4.) Peter asked Tommy to write him an email discussing how we are currently approaching this issue, which Peter will forward to Shankar to see whether he thinks we are in fact taking an appropriate approach, and what further suggestions he has.

  • Additionally, Shankar is interested in being able to automatically generate CellML models from proteomic, genomic and other 'omic' databases.

    • Andrew noted that SBML already is being used for these purposes, and noted that this provides further argument for the requirement of an efficient, lossless CellML <-> SBML translation system.

    • He also noted that to achieve this we will also need to look into how to encode certain information contained within SBML files into metadata in the case where CellML does not use equivalent structures.

Agenda Item 7 - "PCEnv progress update"

  • Andrew has now started spending some of his 10 hours per week working on PCEnv.

  • He has added a code generation function similar to that on the website to PCEnv.

    • This feature currently generates C code - he is investigating how to generate code in other languages.

  • Justin has release PCEnv 0.3.1, which contains a fix for the problem with loading models with piecewise equations in the Linux build

  • He has also backported a fix for a unicode bug to PCEnv 0.3

  • Action Item 5.) Justin will announce the release of PCEnv 0.3.1 on the cellml-discussion@cellml.org mailing list

Agenda Item 8 - "PMR2 progress update"

  • Tommy has been making significant progress on the intrastructure for 'exposing' models within the PMR2 prototype.

Agenda Item 9 - "CellML 1.2 Specification progress update"

  • Andrew has worked through Alan Garny's list of requests for modifications to the current CellML 1.2 Specification draft and has implemented these changes.

  • Most of Alan's request were relatively minor and elicted no disagreement from Andrew

Agenda Item 10 - Catherine: "Update from Andrew on Washington synthetic biology workshop"

  • Andrew sent an email to one of Herbert Sauro's postdocs after last week's meeting asking for information which he can not access from the conference's website, which is inaccessible, but has received no reply.

  • It was suggested that Poul follow up on this.

  • Action Item 6.) Poul to email Herbert Sauro regarding this matter

    • Poul was absent from the meeting and therefore must be informed of this request.

Further items:

"Intellectual property and copyright issues regarding the CellML repository"

  • We need to decide what license we offer models under, and display this information prominently on the website.

  • We need to make sure that the license that models are published in journals allows us to replicate and distribute these models.

  • There is to be a seminar on IP and copyright issues organised by Mike Cooling directly after this meeting. Members of the team attending this seminar will inquire as to our situation and request advice.

  • Peter noted that a number of journals now advocate the provision of CellML or SBML models along with the submission of papers, such as PLoS, Annals of Bioengineering, Biophysical Journal, Journal of Experimental Physiology, etc.

    • We should display a list of those journals that endorse provision of CellML models with paper submissions in a prominenet position on the repository website.

    • We should actively endeavour to increase the number of journals that hold this policy - Peter is willing to write to editors to encourage this.

  • Action Item 7.) James will investigate which journals hold this policy and create a comprehensive list of such.