Mbirn: Ontology Updates

From NAMIC Wiki
Revision as of 13:19, 18 December 2006 by Andy (talk | contribs) (Update from Wiki)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
Home < Mbirn: Ontology Updates

Note: BIRNs ontology resources web page http://nbirn.net/Resources/Users/Ontologies/index.htm


2006.02.02 OTF Update (C. Fennema Notestine)

  • BIRN needs common vocabularies for multi-database communication
    • to find relevant/similar information
    • to link table columns and values (“database mark-up”)
    • to find related information through a higher level semantic network
    • E.g., a scientific query such as “Study the relationship between hippocampal volume and recognition memory performance in AD.” would require data gathering requiring knowledge about clinical diagnosis, cognitive assessment, neuroanatomy, and MR scan acquisition.
  • Our work proceeds at two levels, first for basic terminologies to mark-up the database tables for database registration and integration, and second for higher level conceptual ontologies for scientific queries.
  • BIRN OTF Workshop (1/26-27/2006) at the BIRN-CC focused on basic requirements.
    • Table mark-up process using BIRN’s BONFIRE tool and terms from UMLS and NeuroNames sources (http://imhotep.ucsd.edu:7873/knowme/bonfire.htm)
    • Selected existing or created new concepts, identification numbers, and related definitions to label table information, such as “handedness,” “right handedness,” Edinburgh Handedness Inventory,” “hippocampus,” “T1,” “memory,” and “California Verbal Learning Test”
    • Through this process we have created a BIRN Lexicon of terms, IDs, definitions, and semantic types that we are continuing to develop.
    • For the higher level structure related to cognitive, psychiatric, and behavioral concepts, mBIRN and fBIRN are working together and in collaboration with BrainMap (http://www.brainmap.org/) to develop initial hierarchies of terms.
  • BIRN OTF officially began collaboration with the Center for Biomedical Ontology (CBiO; http://www.bioontology.org) with a workshop (2/28-3/1/2006) at Stanford with primary support from Daniel Rubin, M.D. (Stanford University; Exec. Dir. CBiO), Mark Musen, MD, PhD (Stanford University, PI), and Barry Smith (University at Buffalo).
    • Discussed existing BIRN Lexicon with respect to ontological principles, semantic types, and tools for development (e.g., Protégé, OWL).
    • Grethe, Martone, and Bean as well as Dave Kennedy, Ph.D. (MGH) are also attending the more general CBiO Image Ontology workshop (3/23-25/06) aimed to further promote broad-based interoperability of biomedical image and imaging ontologies through the coordination of current ontology development efforts in the imaging domain.
  • Current work focuses on the expansion of the BIRN Lexicon and finalizing the curation process. With this base structure, we are developing the “Multiscale Investigation of Neurological Disease” ontology to knit together experiment and results.
  • Next telecon April 10th, 2006
  • Next face-to-face meeting of OTF to be held at UCLA following the Mouse BIRN AHM

2006.02.02 (C. Fennema Notestine)

  • The two-day Ontology "Boot Camp" held at the BIRN-CC 1/26-27 was a success.
    • BIRN-CC and Morph, Mouse, and Function BIRN testbeds were all represented. C. Fennema-Notestine and B. Ozyurt represented the Morph BIRN.
    • mBIRN and fBIRN, as the human testbeds, will coordinate all efforts in ontological development given many of the same needs.
    • All testbeds will work together to link anatomical and other demographic and/or experimental terms as appropriate
    • Specific advancements include:
      • identifification of key concerns and limitations of existing ontologies
      • development of best practices for submitting new concepts and relationships to existing ontological structure (based on UMLS) via BIRN's Bonfire tool
      • markup of existing HID database table names and values to allow mediated queries, from terms such as gender to assessments such as the California Verbal Learning Test.
      • planned followup meetings to begin cognitive domain defintions between m&fBIRN
      • provided feedback to BIRN-CC and to the NCBiO for future development efforts
  • Relationship with NCBiO has been solidified; NCBiO will hold a specific workshop for the BIRN OTF on Feb 28-Mar 1, 2006 Working Group Meetings.

2005.12.15 (C. Fennema Notestine)

  • BIRNs ontology resources web page is now live at http://nbirn.net/Resources/Users/Ontologies/index.htm
  • Primary outcomes of recent OTF meetings include:
    • Forging relationships with National Center for Biomedical Ontology (NCBO) http://bioontology.org/ and National Center for Ontology Reseearch (NCOR).
    • Pursuing specific interactions with NCBO
      • Outreach component for guidance in ontology development similar to the path taken with Gene Ontology (http://www.geneontology.org/)
      • To examine the use of Protege ( http://protege.stanford.edu/ ), an ontology management tool
      • Organizing specific BIRN testbed projects to interact with Daniel Rubin, NCBO
      • Arranging BIRN Ontology meeting at NCBO in Feb '06 for face-to-face discussions
      • BIRN will be represented at special NCBO upcoming workshop on imaging ontologies on March 24-25th by Carol Bean and Jeff Grethe
  • Within BIRN we are planning a two-day Ontology "Boot Camp" for Jan 26-27, 2006. Please see Working Group Meetings for more information.

2005.11.21 (C. Fennema Notestine)

  • The BIRN AHM meeting brought together the Ontology Task Force and Working Group participants for a cross-testbed strategy meeting. The OTF includes NIH (Carol Bean), MBIRN (Christine Fennema-Notestine), fBIRN (Jessica Turner), MouseBIRN (Maryann Martone, Bill Bug), and BIRN-CC (Amarnath Gupta, Jeff Grethe).
  • Working group objectives include:
    • Educate BIRN participants on the use of ontologies and associated tools for data integration
    • Develop a set of ontology resources for BIRN participants, based on existing ontologies where possible
    • Identify areas that are not well covered by existing ontologies for possible development.
    • Develop a clear set of policies and procedures for working with ontologies
  • The OTF and Working Group also met with the Data Integration Working Group to build the bridges between Ontologies/Terminologies and databasing/mediation.
  • Current work aims to develop prototypes within testbeds with existing ontology structures to determine development needs. For example, the current BIRN-CC Bonfire tool will extend existing terminology structures and may be further developed.
  • OTF members are working with individuals from the National Center for Ontology Research and National Center for Biomedical Ontology to attend a “boot camp” for creating shared ontological frameworks for handling biomedical images and image processing activities. This meeting is scheduled for March 2006.
  • An extension of this work is to learn more about Protégé software created specifically for ontology development purposes.

2005.10.17 (C. Fennema Notestine)


2005.10.04 (C. Fennema Notestine)

  • Agenda for the BIRN AHM Ontologies Working Group can be found at 2005 AHM Planning: Ontologies Working Group
    • Tuesday meetings will center on introducing ontologies and the current experiences of the BIRN testbeds to define BIRN-specific needs
    • Wednesday meetings will overlap with the Data Integration Working Group to focus on mapping and extending ontologies, inter-testbed facilitation, and how BIRN will be working with ontologies

2005.09.20 (C. Fennema Notestine)

  • Recent Updates from BIRN Ontology Task Force bimonthly meetings (7/29, 8/12, 9/2, 9/16):
    • Recent efforts have focused on gathering information on existing ontologies to discuss needs and limitations relevant to BIRN testbed requirements:
      • 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.
        • UMLS, Gene Ontology, Mammalian Phenotype Ontology, NeuroNames, Systematized Nomenclature of Medicine (SNOMED) Ontology, BrainMap are examples of reviewed ontologies
      • Mapping between ontologies, extending existing ontologies (e.g., Bonfire), and curation policies are topics for the AHM during the Data integration and Ontologies overlapping session
    • BIRN Web Resource Page under development. This BIRN community page for ontology resources will be visible from the main BIRN site.
    • The OTF proposes that BIRN should take a leading role in mapping across human disease and animal models.
  • Draft goal of AHM Ontology Working Group Meeting:

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.


2005.07.26 (C. Fennema Notestine)

  • BIRN Ontology Task Force meeting held July 15th, 2005 to focus on goals and planning for AHM.
    • Ontology Progress (for overall goals, see May 11th notes below)
      • Using the 20 subject public mBIRN database demo, work has begun on a small model of data integration across sites using the mediator (B. Ozyurt, J Grethe, A. Gupta, C. Fennema-Notestine)
      • Within testbed samples exist for guiding development and planning of overarching ontology goals.
      • Necessary areas of focus include curation of entries and relationships and semantic concordance issues, i.e., linking all “related” terms.
    • AHM Overarching Goals:
      • To develop a set of coherent and well-articulated set of policies for the development and usage of ontologies in BIRN that is agreed to by the BIRN community at large.
      • To provide a set of resources and tools that is available to BIRN scientists for use and development of ontologies
      • To formalize interactions with other communities who are working with ontologies and database federation
      • Resulting outline for AHM can be found at 2005 AHM Planning: Ontologies Working Group
      • Testbed specific needs to be outlined at the next task force meeting
  • Next Ontology Task Force meeting scheduled for Friday, July 29th at 10am. Meetings will occur once a month at a minimum; mailing list to be developed via BIRN-CC.

2005.06.9 (C. Fennema Notestine)

  • MouseBIRN AHM has resulted in:
  • Current mBIRN work:
    • Review relevant posted manuscripts
    • Review and search new posted Bio-Ontology links
  • BIRN Ontology meeting to follow

2005.05.12 (C. Fennema Notestine)

  • BIRN-CC is working with the public demo HID site (with 10 subjects in each of two databases) to mimic data sharing across two different sites. The work is aimed to improve the mediator and its ability to mediate at a higher, semantic level. This level of query ties into the Ontology project and the requirements of projects such as MAD. This new mediator will be released in June.
  • Ontology group will meet to more forward specific use-case and technical needs, with representatives from all testbeds and BIRN-CC.
  • As of June 1st, fBIRN HID sites should be up and running for test cases.
  • Proposed test case for communication between databases:
    • HID to HID with basic level communication and move up to the semantic level integration which will rely on the ontology development.
    • HID and XNAT database mediation. Initial case could be within the MAD project between UCSD and WashU. Database registration between XNAT and the mediator will be worked on first, once the new mediator release occurs.

2005.05.11 (C. Fennema Notestine)

  • On-going discussion of use of UMLS semantic types in the KnowMeUMLS/Bonfire browser. Currently we have the capability to see semantic types related to terms, although we cannot search on semantic types. The search process is related only to the metathesaurus in UMLS. We are considering extending the current prototype to accomodate this.
  • Ontology Goals for:
    • Morphometry: Continue to investigate and develop ontological extensions for clinical assessments/ratings. Concept mapping of database table, fields and values to established ontologies and terminologies (esp. via UMLS) with goal of supporting increased database interoperability.
    • Function: Continue to investigate and develop ontological extensions for experimental designs (in collaboration with BrainMap project; http://www.brainmap.org). Concept mapping of database table, fields and values to established ontologies and terminologies (esp. via UMLS) with goal of supporting increased database interoperability.
    • BIRN-CC: Continued development of web interface (a component of the BIRN 2.0 software release; http://imhotep.ucsd.edu:7873/knowme/bonfire.html) to add new terms and relationships to BONFIRE. This continued development will add additional support for advanced browsing of UMLS extensions. Work closely with test beds in the mapping of their databases to the ontologies being defined in support of BIRN data integration efforts.
    • Mouse BIRN: Concept mapping of database table and field names and values to established ontologies and terminologies (esp. via UMLS) with goal of supporting increased database interoperability. Observations: Overall, NeuroNames provides good basis for terminology, but lacks the necessary relationships. That is, UMLS lacks sufficient contextual specificity, but many terms could be related via class relationships or by combinations of terms. UMLS has conceptual bias toward clinical and human terminology. Semantic types for imaging are poor.
  • Ontology Task Force Report to BIRN Executive Committee 12 April 2005
    • Meetings
      • 07 March 2005 Attendees: Carol Bean (NCRR), Christine Fennema-Notestine (Morph BIRN), Jeffrey Grethe (BIRN CC), Amarnath Gupta (BIRN CC), Maryann Martone (Mouse BIRN), Jessica Turner (Func BIRN); Discussed current status, perceived needs, and plans regarding ontology efforts at individual test-beds. Discussed mission and objectives of Task Force.
      • 16 March 2005 Attendees: Burak Ozyurt (m&fBIRN); Christine Fennema-Notestine (mBIRN); and BIRN-CC's Jeff Grethe, Amarnath Gupta, Edward, Vadim Astakhov, and Xufei Qian; Discussed interaction of Mediator and ontologies.
    • Task Force Objectives
      • 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.
    • Proposed tasks
      • Form a "use-case task force" of one person from each test-bed group to create a few interaction scenarios suitable to define one or two representative use cases for interoperability. Identify which databases to mediate for the use-cases. Following this, a technical-specification task force involving the technical people from each group would see what it means to implement the use-cases using a simple ontology, and in the process discover impediments.
      • Extend research examples from BIRN projects to a relationship-rich ontology based on the UMLS. Assess UMLS coverage of needed terms and relationships. Identify and create extensions and modifications to the existing UMLS via the BONFIRE tool necessary for representing and linking data.
      • Based in part on the created “ontology” structure, the BIRN-CC’s Mediator tool should be extended to provide an integrated view of the BIRN shared database infrastructure. The search engine needs look-up tables with task descriptions, score interpretations, etc., for clinical assessments.



2005.04.7 (C. Fennema Notestine)

  • March 7th Roundtable discussion of the "ontology" projects across testbeds to help define end goals, common themes, etc. Attendees: Carol Bean (NCRR), Jeff Grethe (BIRN-CC), Maryann Martone (MouseBIRN), Jessica Turner (fBIRN), Amarnath Gupta (BIRN-CC), and Christine Fennema-Notestine (mBIRN).
    • BIRN Ontology Task Force requested to help coordinate the defined status, mission statement, plan and priorites for action. Information from each testbed to be routed through Carol Bean, for rpesentation at the BIRN Executive Committee meeting April 12.
    • mBIRN needs appear to be common across testbeds, and in line with the original work done in the Mouse BIRN. We are extending this work to include additional cognitive domains, and expanding the Bonfire tool.
    • fBIRN needs are more extensive, particularly in the sense of experimental design and fMRI parameter tracking. These may be dealt with as a separate path through a possible subgroup.
  • March 16th Interaction of Mediator and Ontologies. Attendees Burak Ozyurt (m&fBIRN), Christine Fennema-Notestine (mBIRN), and BIRN-CC's Jeff Grethe, Amarnath Gupta, Edward, Vadim Astakhov, and Xufei Qian.
    • Defined difference between variable NAMES and SCORES with respect to the mediator schema. Content (semantic) is identical, but the storage of information is different.
    • Assessment name table expanded for synonyms, e.g., SEX and GENDER.
    • Assessment score table can be expanded to include synonyms, such as scores representing males to be "1" or "M" or "MALE"....
    • Concept IDs in UMLS/Bonfire will define the synonyms to be used by the Mediator. This may need to be done by site as new information/new databases are registered to the Mediator. If new variable names or scores exist in the new database, the Concept ID links must be created for synonyms.

2005.02.28 (C. Fennema Notestine)

    • BIRN-CC’s AG and XQ have created a developmental prototype for the new web interface that will allow the addition of new terms and relationships via BONFIRE, extending the existing KNOW-ME UMLS browser.
      • Stages for curation exist to
        • 1) allow the user to develop personal "MyBonfire" that requires no external curation process,
        • 2) submit personal additions to the curation team for processing, and
        • 3) addition to shared Bonfire for information that successfully passes the curation process.
      • Development continues to allow options to delete concepts and relationships (these will have to follow the same curation process).
      • Larger scale development aims to improve the graphical interface beyond the basic UMLS scheme.
      • CFN is currently testing prototype for presentation at March MBIRN AHM meeting.
    • DK's made progress on a new tool for navigating rdf data ( see http://simile.mit.edu/piggy-bank/ ).
    • DK's group is currently working on a new view ontology that will serve as the right infrastructure for developing views of biological data.
    • Review of the Platypus semantic Wiki option for development suggests that it may not be aimed at the appropriate user for our current development task. CFN still to review site.

2005.02.02 (C. Fennema Notestine)

    • BIRN-CC’s AG and XQ will develop web interface for adding new terms and creating new relationships via Bonfire, extending the existing KNOW-ME UMLS browser (see BIRN Portal http://imhotep.ucsd.edu:7873/knowme/umls.html ).
    • Current goal is mid-February for testing, with the aim to display this work at the March meeting.
    • In addition to the ontological structure, we need simple “look up” tables with task descriptions, including information on score interpretation, etc. CFN has sent AG the updated descriptions for the basic clinical assessments for reference in the search engine.
    • AG's team is extending the current KNOW-ME ontology browser to include the extensions to UMLS. It would also extend the capabilities of the current browser to issue HID and mediator queries with one or more selected terms from the KNOW-ME interface.
    • DK and CFN have spoken regarding the scope of the ontology project development once the information entry interface is in place.
      • Initial proposal is to start small with a select 10 experienced neuropsychologist types to assist in growing the database with valid information. Ultimately, input resources could grow to a larger set of individuals in which case curation becomes a much more significant issue.
      • DK mentioned Platypus (http://platypuswiki.sourceforge.net/ ), similar to a “semantic WIKI,” that allows relational links between objects and the creation of new terms, as well as some potential graphical presentation tools.
      • DK and CFN will coordinate a demo of some of these existing tools so that we can gauge their usefulness.
    • DK's ongoing work in Haystack aims to provide options for presentation and organization of query results.

For Background information see: