2006-04-12 Conference Call Minutes

Participants: Frank, Gary, Ranjani, Joanne, Emek, Ken, Dan, Andrea, Alan (Jeremy via Skype)

  1. Development status update

    1. DXTrackWorkplan status update (Emek/Gary)

Gary will Ken and Frank will send examples of gene regulation from INOH and TRANSFAC to the discussion list Emek discussed states with Paul Thomas, Carl Shaefer, Chris Sander and people from Cell Signaling Technologies. Emek is working on a new proposed implementation of states that he will hope to post to the wiki by the end of this week. He is trying to isolate the reference data on physical entities as a utility class - the protein class remains as the only protein class in the main class hierarchy (chidren of the entity class). There will be a way to state unknowns and groups them into groups. Emek did his best to integrate ideas from existing proposals.

  1. How to get going on splitting use cases/implementation proposals. Alan_Ruttenberg/Re-expressing a use case (Alan) (action item from last call)

Other people should read this. Emek read it and will try it out and give comments. Action item for Emek for next meeting.

  1. Late 2005 Proposal for BioPAX Development/Working groups proposal status update (Alan)

Naming of the tracks is still an open issue. It seems that doing things in the OWL way will require a greater reworking, so that may factor into the decision.

  1. BioPAX Validator / PAXerve interoperability (Ranjani)

The BioPAX validator + paxerve projects will be combined. The paxerve API and the rules will be shared. Alan suggests that OWL validation should be a step for the validator to catch new classes of mistakes. Ranjani - Currently this is not done because it impacts performance and that users should run the OWL validator separately. Alan - the BioPAX validator should do both because it is clearer for the user. Emek and Ranjani will report regularly to the list about this project and ask for feedback from the group.

  1. [WWW] debugging the bug E.coli data merge project (Jeremy/Alan) ([WWW] code repository) status update

-majority of EcoCyc errors found (ID inconsistencies among compounds) have been fixed in EcoCyc by their curators, with the exception of one, where there was a disagreement. There were 30-40 mistakes in ~800 compounds. Only from comparing two databases. Was able to represent compounds and reaction as classes and found identical compounds and reactions with the reasoner. One discovery in reaction matching is that there is an asymemtry with left and right vs. right and left - the reasoner can't deal with this unless you do something special like represent both directions as 2 reactions, but that incurs a space cost. Project is still ongoing. - Gary-Are you evaluating rules, like SWRL? Alan will discuss this in the Manchester report. Alan has only implemented OWL and queries, but not rules. Andrea has been using Jena rules and is now an expert. Adrew- Clean, infer, custom pathway inference + validation are uses for rules. Eyeball is an application in the Jena stack that may help here. Alan is hoping that he can give a tutorial by June on these various pieces with examples.

  1. [WWW] INOH BioPAX Level 2 export status update (Ken)

INOH has an XSLT converter from INOH files to BioPAX Level 2 in pre-beta status. They already converted all INOH files to BioPAX and Ken can send some examples to the discussion list for feedback before releasing. Because it's an XSLT conversion, most extra molecular role ontology is in the xref property, so that is a recognized limitation. They mapped everything that they can to Level 2 and the code is working. Alan - where do the sameAs on individuals statements come from? Ken - used sameAs to state that some individuals are identical as extra annotation. Alan is interested in seeing the reasoning for using these - asks Ken to send that to the list - also XML parsers should realize they are there as the information is important.

  1. [WWW] http://cancer.cellmap.org MSKCC Cancer Cell Map BioPAX Level 2 support (Gary)

Cancer cell map is a new pathway database focused on pathways that play a role in Cancer. The pathways are now available in BioPAX Level 2.

  1. [WWW] https://cabig.nci.nih.gov/ caBIG meeting update

Emek, Ranjani, Guanming, Ethan, and other caBIG participants recently met about BioPAX at the annual caBIG meeting in DC. Emek set up a conference call with Anand Kumar from OBO to find out how OBO and BioPAX can interoperate. Also, some new data consumers there were more interested. There was a lot of discussion on semantic web technologies and OWL, but the audience was lost on those issues. QPACA is a new caBIG tool that can read BioPAX format now at some level.

  1. Governance. Gary action item to provide alternative wording for BioPAX Governance (action item from last call) - Gary was not able to get feedback on this yet, so he will re-read and send to interested people (so far Alan + Joanne) for feedback.

  2. BioPAX workshop organization - there is some desire to focus this meeting on discussing the proposals in on the DXTrackWorkplan page - basically to be focused on BioPAX features. Alan wants a clear statement of goals for the meeting. Action item: Emek will post the results of the survey monkey page. Alan would have liked to have a process to discuss what the meeting was about - because the way the survey questions and conclusions from them need to be discussed. Gary mentions that in his mind, the meeting will be about dxtrack proposal. Andrea doesn't think you should differentiate. Alan's main point is to announce and create things in order not to alienate portions of the community.

--- these items were not able to be discussed and are tabled until next time ---

  1. Meeting report on 2006-04-03_to_07 Informal Workshop on formal semantics for BioPAX at Manchester

  2. Future meetings

    1. Potential workshop with Barry Smith (Alan) (action item from last call)

  3. Feedback on biopax.org (Alan) (action item from last call)

last edited 2006-04-12 17:08:52 by bader