The Future of the Foundational Model of Anatomy

Date: November 12-13, 2009

Venue: Stanford University

Organizers: Onard Mejino, Natasha Noy, Alan Ruttenberg

Agenda

Thursday, November 12

Preliminaries

8:30 Welcome - Mark Musen, Barry Smith, Cornelius Rosse

9:00 Short introductions by participants

9:15 Objectives of the FMA group for the meeting - Jim Brinkley

9:30 Presentation by FMA group explaining semantics, deficiencies in

current representation, questions and targets for OWL (Onard and Cornelius)

11:00 Requirements brought in from OBO Foundry/Semweb (Alan)

11:15-11:30 The FMA and its ontological commitment(s) (Stefan)

Review and discussion of current approaches

11:30 Current approaches to translation - Christine Golbreich

12:05 Current approaches to translation - Natasha Noy

12:40 Lunch

1:30 Current approaches to translation - Chris Mungall

2:10 Introduction to OWL 2 and its features - Uli Sattler

2:50 Break

Details

3:15 Who's using FMA, and how? - Onard Mejino

3:45 Presentation of specific challenges - Onard Mejino

Discussion sections addressing issues identified above, focused on cardiovascular system as exemplar

4:15 Single/Multiple inheritance and inferred hierarchies

    Problem: Single inheritance hierarchy prohibits multi-supertype   

assignments.

    Examples: 
        - primary incisor tooth can be a subtype of either incisor   

tooth or primary tooth

        - right female breast can be a subtype of female breast or   

right breast

        - proximal phalanx of thumb can be a subtype of either   

phalanx of thumb or proximal phalanx of finger

        - question: can OWL automatically infer one of the supertypes   

in the inferred hierarchy

Friday, November 13

9:00

Representation of different contexts using the same relation.

    Examples: 
        - prostate can be regionally subdivided into different parts   

using different contexts;

            a. classically into anterior lobe, median lobe, right   

lateral lobe, left lateral lobe and posterior lobe

            b. histologically into peripheral zone, central zone,   

transition zone, and peri-urethral zone

            c. surgically into right median lobe, left median lobe,   

right lateral lobe, left lateral lobe, right dorsal lobe and left dorsal lobe

        - heart can subdivided into different contexts: 
            Right side and left side 
            Biatrial part and biventricular part 
            T7, T8, T9, T10 parts 

These are all valid regional parts but the slot has regional part can only accommodate one context, in the case of the FMA it’s the classical anatomical representation and the rest are relegated to the attributed slot attributed part because it would be inappropriate to create has regional part 1, has regional part 2 and has regional part 3 slots

9:45 Review of relations and their usage.

  - Axiomization 
  - Use of Attributed/reified relationships 
     - Are they necessary? 
     - If necessary how to represent them in OWL 

10:30 Break

11:00 Post-Coordination

11:45 What can be inferred? Quality assurance - error and consistency checks

12:30 Lunch

Moving forward

1:30 Review of goals and outlines of possible solutions - Jim Brinkley

2:00 Technical methods to achieve interoperability and orthogonality:

2:10 OWL Modularity – Uli Sattler

2:20 Ontology views – Todd Detwiler (10 min)

2:30 Cross-references and semantic web linking methods – Alan

Ruttenberg

3:00 Break

3:30-5:00 Discussion, action items, and future plans – Jim Brinkley, Alan Ruttenberg, Olivier Bodenreider

5:00 Closing remarks Cornelius Rosse, Mark Musen, Barry Smith