2005 AHM Planning: Ontologies Working Group
- Morphometry BIRN: Chris Fennema-Notestine
- Function BIRN: Jessica Turner
- Mouse BIRN: Maryann Martone, Bill Bug
- BIRN CC: Jeff Grethe, Amarnath Gupta
- NCRR: Carol Bean (Ontology Task Force co-Chair)
Tuesday, 2-6: Ontology Working Group
- 2-2:30: Brief introduction to ontologies and their importance to BIRN (Maryann) High level and directed toward domain scientist
- o BIRN ontologies
- o Lay out what ontologies are needed
- o Presentation of BIRN ontology policies and procedures, in the context of those developed by other large projects, with time for open discussion. It was suggested that this topic be revisited later in the session.
- o Describe the role of the task force
- o What should be accomplished in this meeting
- 2:30-3:30: Introduction to UMLS and Gene Ontology (Carol and Bill, 20 min each + 20 min discussion)
- o What are they; what do they do; what don’t they do
- o General issues in the use of any ontologies and what are limitations of current ontologies in general for ontological engineering task
- 3:30-4:15: Current testbed experience with ontologies for data integration
- Mouse BIRN (Maryann)
- FBIRN (Jessica)
- MorphBIRN (Christine)
- 4:15-4:45 Break
- 4:45-5:45: Panel discussion (Ontology Task Force): what areas of interest to BIRN are not adequately covered by existing ontologies, including discussion of mapping animal models onto human disease (Part 1)
- 5:45-6:00: Wrap up (Maryann and Carol)
Wednesday: Data integration and ontologies
- 8-8:30: Data integration and ontologies: brief overview (Jeff)
- 8:30-9:00: Mapping between ontologies (Bill)
- o Practical examples: UMLS and GO
- o What’s being done in BIRN: mouse BIRN and FBIRN
- o How to deal with the problem that concepts in an ontology domain have a context and the matter of mapping between them may not be straightforward.
- 9:00-9:30: Extending Ontologies (Jeff)
- o Working with Bonfire
- o Semantic concordance
- o Curation of Bonfire terms
- o Evolution of Bonfire
- 9:30-10:00 Discussion: Animal models to human disease: inter-test bed facilitation (Bill)
- 10:00-10:30: Break
- 10:30-12:00: Working with ontologies (Amarnath, Maryann)
- o Roles of programmers vs roles of domain scientists
- o Demonstrations of ontology creation and mediator associated tools
- o What should be mapped in the database
- o How should the mapping be represented in the database
- o What about semantic types?
- o Ontology concept-based queries
- 12:00-12:30 Plans and milestones for coming year; goals for test beds by the Spring AHM’s (Jeff)
BIRN AHM: Ontology Workshop Session Materials
Goal of the AHM Ontology Working Group: The goal of the AHM ontology working group session is to discuss the use and development of ontologies within BIRN, providing practical examples and advice. We will cover a general introduction to ontologies and their use within BIRN. We will provide overviews of existing ontologies that will be of use to BIRN as well as tools for constructing, maintaining, extending and expressing ontologies. We will also highlight existing ontology development efforts by BIRN participants and collaborators and try to identify areas where new ontologies are needed. In the joint ontology-data integration session, we will go over practical aspects of using ontologies to enhance interoperability of BIRN data sources. By the end of the meeting, we hope that we will have formulated a clear set of policies and procedures for developing and working with ontologies in BIRN and a set of goals for moving forward.
The meeting agenda was prepared by the BIRN Ontology Task Force:
• Carol Bean (co-chair) NIH • Maryann Martone (co-chair) Mouse BIRN, BIRN CC • Amarnath Gupta BIRN CC • Bill Bug, Mouse BIRN • Christine Fennema-Notestine Morph BIRN • Jessica Turner FBIRN • Jeff Grethe BIRN CC
Goals of the Ontology Task Force (OTF):
Create an ontology prototype that relates brain structure and function. While a “complete” ontology is beyond the scope of this project, the aim is to develop, using existing structures where appropriate, a set of knowledge structures sufficient to allow integration across multiple databases and relevant data types.
Identify and assess existing ontologies and terminologies for summarizing, comparing, merging, and mining datasets. Relevant subject domains include clinical assessments, demographics, cognitive task descriptions, imaging parameters, and derived (fMRI) data.
Identify the resources needed to achieve the ontological objectives of individual test-beds and of the BIRN overall. May include finding other funding sources, making connections with industry and other consortia facing similar issues, and planning a strategy to acquire the necessary resources.
BIRN Knowledge Sources:
Why are they important?
The Biomedical Informatics Research Network (BIRN) project is creating a high bandwidth and large capacity network for linking databases created at multiple centers concerned with human disease and associated animal models. A key component in the BIRN is the establishment of an architecture for linking and navigating across distributed data sources. Each participating site is creating an individual database around the type of experimental and derived data it produces. These databases will be linked together to allow cross querying and data integration across a series of federated databases, containing multi-dimensional and multi-scale information relevant to the understanding of human disease. To facilitate data exchange and database interoperability, each of the BIRN databases is mapping their schema and data to shared knowledge sources.
What is a shared knowledge source?
Shared knowledge sources include controlled vocabularies, metathesauri, taxonomies, spatial atlases and ontologies. Spatial atlases will be discussed in a different session at the AHM. For the purposes of this discussion, we will define a controlled vocabulary as a set of terms with definitions; a taxonomy as a classification hierarchy, e.g., Purkinje cell is a neuron, and an ontology as a set of concepts with relations between them that may include both “is a” and “part of” relationships (meronyms) as well as others. A metathesaurus provides mapping between concepts contained in different knowledge sources. UMLS has a large metathesaurus. In most of these knowledge sources, each concept is assigned a unique identifier that distinguishes it from all other concepts. In most of the discussions, we often don’t distinguish among these different types of sources and refer to them all as “ontologies”. Participants should keep in mind, however, that as we become more familiar with knowledge sources and what they can and cannot do, the above distinctions will become important. For a formal definition of ontologies, see http://www-ksl.stanford.edu/kst/what-is-an-ontology.html.
Developing the BIRN Knowledge Sources
The official position of the OTF is that BIRN should use existing ontologies wherever possible, that we should extend them as necessary if the coverage of concepts is not sufficient and that we should develop ontologies or ontology mappings only when they are needed. The process of extending and maintaining ontologies can be challenging, and we are gathering information as to how these tasks are handled in other large ontology initiatives, e.g., Gene Ontology. New ontologies may be needed if no existing ontology covers a required area or if existing ontologies provide a substantially different conceptual view of a field that cannot be reconciled with BIRN’s view. Any ontology work performed by BIRN should be aligned with other efforts and provided back to the source ontology and/or to the UMLS. This same position is extended to ontology tools and formalisms.
Although BIRN is an open, diverse and fluid environment, the use of ontologies for enhanced interoperability will be pointless if we allow random use of ontologies. The OTF recommends that there be a set of ontologies that are approved for use and a set of policies and procedures for adding or creating additional knowledge sources. Current knowledge sources that are currently in use include UMLS, GO, LOINC, SNOMED, NEURONAMES. Mapping between these sources is provided by UMLS. Extensions to these are handled by the BIRN Bonfire tool. The OTF recommends that we develop a set of criteria for including ontologies as BIRN knowledge sources. At least some of these criteria will align with those of the OBO (Open Bio-Ontologies consortium), e.g., ontologies must be orthogonal to existing ontologies, they must have unique identifiers and definitions for concepts. The OTF will assume an oversight role for the addition of ontologies to the BIRN knowledge sources, that is, they will ensure that proposed ontologies meet the criteria.
Topics for discussion at AHM
Policies and procedures:
The OTF will develop a set of policies and procedures with input from the Ontology Working Group designed to assist BIRN test beds and general users with using and developing ontologies with the goal of increasing interoperability of BIRN data resources and tools. There are many different levels on which ontologies can be considered: • Content (terminology and concepts) • Inter-ontology mapping • Formalism (knowledge representation) • Tools These aspects will be considered separately. The Policies and Procedures will include: 1) recommendations for mapping sources to ontologies; 2) Providing a set of ontologies that should be used for this purpose; 3) Providing a set of procedures for adding, extending or developing ontologies; 4) Providing guidelines and recommended tools for mapping data sources to ontologies
Developing resources for BIRN participants on the use, development and importance of ontologies.
A BIRN community page for ontology resources is being developed that will be visible from the main BIRN site. This page will provide the policies and procedures, access to the BIRN knowledge sources and general information on ontologies that will be helpful to the community. One point of discussion at the AHM should be what type of resources would be helpful to those using ontologies and those creating ontologies.
For each ontology in the BIRN Knowledge Sources, an information sheet will be provided givng a concise summary of: o What it covers o How it is used o How BIRN can benefit from it o Within a test bed o Across a test bed
Thus far, FACT sheets have been developed for Neuronames, the Mammalian Phenotype Ontology and SNOMED. They are available from the web page that is being developed for BIRN ontology resources , currently viewable at: http://nbirn.net/Resources/Users/Ontologies/
Because ontologies are so central to the mission of the BIRN project, and because BIRN doesn’t have any formal BIRN-wide ontology activities, the OTF is planning on conducting a series of workshops for BIRN participants to help them become familiar with ontologies and working with them.
Human Disease-Animal model
While ontology development has not been a main focus of the BIRN project, we believe that there is one area where the combined expertise of BIRN will be invaluable: mapping between human diseases and animal models of human diseases. The OTF proposes that BIRN should take a leading role in mapping across human disease and animal models. Levels to consider: o Human Anatomy – Mouse anatomy 1. Other groups have made headway in this and we can utilize their efforts o Human disease – animal model 1. Less has been done here. Good opportunity for interaction with MGI 2. BIRN CC has been developing disease maps o Human functional assessment – animal functional assessment 1. Mapping cognitive tasks across species
This activity is viewed as a long term goal and will be performed in collaboration with other groups that have also expressed interest. At the AHM, we will discuss which areas are critical for promoting cross-test bed query capability.
Useful Links for Background Material
http://www.loni.ucla.edu/twiki/bin/view/MouseBIRN/MouseBIRNOntologyRelatedResources Resource set up by Mouse BIRN containing useful articles and links
Protégé web site. Protégé is a tool for creating, maintaining and viewing ontologies. Provides some educational materials on Ontologies 101. See also http://www-ksl.stanford.edu/kst/what-is-an-ontology.html
Possible Participants at AHM
- Morphometry BIRN: Christine Fennema-Notestine, Burak Ozyurt
- Function BIRN: Jessica Turner, Dingying Wei, Judy Ford, Angus MacDonald, Ayse Belger, Cindy Wible (FBIRN Ontology group)
- Mouse BIRN: Maryann Martone, Bill Bug,
- BIRN CC: Xufei Qian, Amarnath Gupta, Daniel Wei, Jeff Grethe(part-time), Mark James (part-time)
- NCRR: Carol Bean