Skip Navigation
NIH | National Cancer Institute | NCI Wiki   New Account Help Tips
Page tree
Skip to end of metadata
Go to start of metadata

Date

Thursday, August 12, 2010

Attendees:

Sherri DeCoronado, Larry Wright, Stuart Bell, Trish Whetzel, Mike Riben, Harold Solbrig, Rick Kiefer, Stuart Turner, Riki Ohira

Notes:

  • Reviewed what we discussed last week, the CORE OMV attributes the group decided should be required. They also decided to use BioPortal as implementation of OMV since it is being maintained.
  • Spent a lot of time on URI, which Sherri did draft an explanation/definition of URI (added definition is in red text)
  • name:
    • Should be required and definition is clear
  • acronym:
    • There were some caveats or issues that would need to be included for this field, although it was not listed as required
    • Definition could be misleading since some groups do not always provide same short name for a given terminology/ontology
    • A note is needed stating that acronym needs to be unique in the context of the service, particularly for terminology/ontologies that are not very well known
    • Q: Is acronym meant to be the name space? A: Depending on what you mean by name space; UMLS-world LOINC is known as NMC. This is why we have IRIs, and there should be a unique IRI for every ontology/terminology.
  • We should put what the caveats are and we should not make it a requirement to get community consensus for those that may have multiple names/URIs/etc
  • description:
    • Should be version independent and recommend having curator's add the source of the description
    • If this may be the one source that people will look at for information about ontologies, it would be helpful to have notations if the source is canonical versus individual creation
    • Additional information/descriptions about terminology may not fall under the required "description" attribute, but could be included in another one that might be Optional
  • creationDate:
    • Suggested to make it Optional, but we may need to include some date
    • Looked at version which is another currently required attribute, but not sure if we would need versionDate as well
    • Suggestion to make creationDate versionDate instead
    • Final Decision: change creationDate to versionDate and make it Optional
  • naturalLanguage:
    • The group felt it should be required, but should also enable the inclusion of multiple languages
    • The definition would specify the primary language
    • Q: For MeSH, how would one populate this field? A: If someone puts the English version of MeSH, the primary language would be English; if the French version is loaded, then the primary language would be French, etc
    • Would need to use double character acronym (IETF language tag; RFC 5646 and RFC 4647)), like en for English
  • hasCreator:
    • The definition should specify the "owner" rather than "contributor" or "author"
    • The person who created the original terminology/ontology and not contributing or authoring versions
    • OMV does not have a definition (based on search in BioPortal)
    • Dublin Core definition for creator: An entity primarily responsible for making the resource
    • Q: For French MeSH, would the creator be NLM or the French group that does the translations? A: The group agreed that it would be NLM since they are maintaining it.
  • hasOntologyLanguage:
    • The group agreed that this should be required and the definition is clear
    • Example: For MeSH in UMLS, the Ontology Language would be RRF
    • Q: Would you include original source language or the language used for that particular ontology? A: We would be looking for the source language for the given ontology.
    • Q: Do we need to note the native ontology language and the "version" language (particularly if they are different)? A:
    • There would be a set of metadata about the canonical version and set of metadata about the version being loaded?
    • Dublin Core contains hasFormat attribute that could be used for multiple canonical sources, for example, SNOMED
    • hasOntologyLanguage should define canonical source, and add hasFormat with definition to mean other accepted formats (not canonical source), which would be Optional
  • resourceLocator:
    • Q: What resourceLocator would we give for a terminology that is in RRF and does not have a URL that can be downloaded?
    • Suggestion since URI is Optional it would make sense the resourceLocator is required
    • The definition seems clear enough
  • version:
    • It should be required and the definition is sufficient
    • Are we being consistent with the use extensions that NCBO has? Stuart will check to make sure.
    • Virtual Ontology has 6 attributes, which include hasVersions, and currentVersions.
      • hasVersion in Protege is the instances of the current version of the ontology, but not configured properly to comment on currentVersion
      • This may be an implementation issue. The version described by OMV is an instance, and we can't know if this is the current version and what the other versions are (not sure if this is an issue), but it is an artifact of implementation
  • We have a few more attributes to go through, but we still need to look at the Provenance.
  • We should have the contracts in place soon and may have to have a brief Kick-Off again once contracts are in place
  • Q: Will we have enough information to start implementing some of the core metadata on VKC or something once we have the core attributes identified and defined? A: Yes.
  • Once we implement it, we can better evaluate the use of this, and see this as an alpha/prototype
  • We want to keep CTS2 submission as closely aligned with the work this group is doing. In the future, it will be helpful to go over where we sit on submission and how to go forward on the alignment
  • We can plan to go over CTS2 alignment at one of these meetings

Action Items:

  1. Stuart Turner - Check NCBO use extensions to make sure we are consistent
  2. Stuart Turner - Will look at provenance information in OMV and other source
  3. Stuart/Sherri - Schedule CTS2 alignment for agenda of future meeting
  • No labels