Version 7 (modified by detwiler, 11 years ago) (diff)


Meeting minutes from 3/28/2007

Date: 3/28/2007 10:30AM
Attending: Dan S, Gary, Linda, Jim, Josh, Dan C, Todd, Onard, Max, Alan

During todays meeting a few of the participants had a chance to give a general overview of work that they have been doing that pertains to the new Ontology View grant. Below is a brief summary of those overviews:

Todd Detwiler

One of the early requirements of this grant is to determine the specification requirements for defining views, and then augmenting (or creating new) a query language with the necessary expressivity. Some of my initial thoughts included:

  • Mentioned RQL and RVL, which were considered in grant proposal
  • Looked into SparQL
    • Has W3C recommendation for RDF(S) query language
    • With CONSTRUCT operator SparQL fulfills closure property (results in RDF)
    • Missing support for regular paths (like regular expressions for paths)
  • SparQLeR
    • SparQL extension to include regular paths
    • SparQLeR processor not yet available
  • AMIA poster proposal looks into composing a SparQleR (or SparQL) query with a SparQLeR (or SparQL) view query.

Dan Cook

Structure ontologies have poor representation (if any) of what these structures do. BioEngineering mathematical models have a lot of information about processes. Dan discussed work that he and Max are undertaking to join static structure to dynamic physiology models.

  • Anatomy by anatomist does not always look like a physiologists view of anatomy
  • need a [symbolic] model in the middle, ApplModel
    • Extract just what we need from ref ontology
    • FMA subset plus other info as needed for simulation modeling
    • Current model built as Protege ontology and as an extension of a general ApplModel Ontology.
  • Why do we need this?
    • Many non-standard ways of representing models and for annotating elements of these models
      • interoperability
      • standardization of tools (i.e. visualization)
    • Aligning/integrating multi-scale models
    • Modularization

Onard Mejino

Onard is continuing work on the FMA as well as researching potential use cases (like RadLex and [BIRNLex] as suggested by Daniel Rubin).

  • Making sure that the FMA as a reference ontology conforms to sound ontological principles
  • Content enhancement, extending FMA types and relations as needed
  • The FMA is meant to be flexible enough to incorporate or reconcile multiple views
    • Reconciling views for interoperability
    • Note: automated alignment not specified as part of this grant, so any alignment required will be largely a manual effort.
  • Cleaning up views as they are incorporated
    • Imposing more rigid adherence to ontological principles
    • Modifying reference ontology where it cannot currently reconcile new view elements.

Back to OntologyViews