Is Biobank name an institutional identification?

[zhengj2007]

From MBrochhausen@gmail.com on July 08, 2011 13:53:22

Would biobank name be an institutional identification? Were is the definition of what an institution is according to IAO? If biobank name is not an institutional identification, were would it go? Wouldn't it make sense to have textual identification items as one class with (possible) author identification and institutional identification as sub-types? Is it necessary to give those two a special treatment as opposed to person identification, dog identification and painting identification?

Thanks, Mathias

Original issue: http://code.google.com/p/information-artifact-ontology/issues/detail?id=116

[zhengj2007]

From mcour...@gmail.com on July 08, 2011 11:29:29

Those were added by Larry, who had a specific use case of parsing literature information if I recall correctly - thus the specific author and institution identification classes. He may be able to add more information.

Owner: HunterOn...@gmail.com

[zhengj2007]

From HunterOn...@gmail.com on July 08, 2011 11:54:20

Melanie is right, the original motivation for this was to pick out the sorts of institutions that are associated with authors in publications. For a variety of in-scope use cases there is a use for naming something that identifies an institution and something that identifies a person, e.g. text spans in biomedical publications. It is also used elsewhere in the ontology (e.g. institution list). I'm not convinced dog or painting identifications are in scope. There are also "centrally registered identifiers" (CRID, see IAO_0000574 et seq) which might be better for (licensed) dogs and (inventoried) paintings. I agree it would be of potential value to have a parent term that subsumes these classes, but we had a long argument over what was or wasn't an identifier -- it seems that any information content entity can be conceived of as an identifier. There is a synopsis of this discussion under the obsolete term "_identifier"

As to whether A biobank name is an institutional identification, it seems to me that if there is a specific biobank being named, it would indeed be a "particular institution" and then meet the definition.

Larry

[zhengj2007]

From alanruttenberg@gmail.com on July 08, 2011 12:25:50

It depends on whether we define an institutional name as any name of an organization, or whether we define institutional identification more narrowly for Larry's case, as an IA that is created with the more specific objective of identifying organizations in certain kinds of texts, and only appearing as parts of such. We want to have room nearby for names that serve other purposes, such as legal or regulatory requirements.

[zhengj2007]

From MBrochhausen@gmail.com on July 08, 2011 12:35:22

Alan: yes, I figured that this is the intention behind that, but it would be nice if that would be apparent from the definition of either institutional identifier or textual entity.

Larry: OK I'll treat it like an institution and I'll talk to Bill Hogan about adding "institution" to OMRSE. Some definitions of what a biobank is do not fit the more narrow definitions of institutions, but it seems that "biobankers" aren't to sure themselves. In the project I am working with the opionions range from 2 freezers to something similar to the UN ;-)

Thanks, Mathias

[zhengj2007]

From HunterOn...@gmail.com on July 08, 2011 12:36:57

The current textual definition is for the broader case: "A textual entity intended to identify a particular institution" which I think would subsume, for example, a legal name of an institution as well as the ones used by authors in publications. There is no IAO definition of an institution, which strikes me as out of scope (institutions aren't information content entities). Checking the NCBO annotator, there are 11 terms named or related to "institution" returned, for example, this one in transmed: http://bioportal.bioontology.org/ontologies/42758/?p=terms&conceptid=transmed:TMO_0025 which has xrefs to several others. It doesn't appear to me that any foundry ontology defines "institution."

[zhengj2007]

From MBrochhausen@gmail.com on July 08, 2011 12:59:01

Right. I'll go with your recommendation and subsume Biobank Name here. And I already tlaked to Bill Hogan and we are looking into defining institution for OMRSE since, as you said, it is out of scope of IAO.

Thanks again, Matt

[zhengj2007]

From alanruttenberg@gmail.com on July 08, 2011 13:05:45

OBI defines 'organization' out of necessity, and I'm guessing institution is either same or subclass of that. http://purl.obolibrary.org/obo/OBI_0000245 The definition you cite is too narrow for "organization", and maybe too narrow for even "institution" - do institutions in your intended sense necessary "act legally in the same way as a person"?

It's annoying that the crossrefs aren't hyperlinks. ho hum...

dailymed:organization - http://dailymed.nlm.nih.gov/dailymed/search.cfm?startswith=organization&x=0&y=0 linkedct:agency - http://static.linkedct.org/page/linkedct/agency UMLS:C1272753:
Semantic Types Idea or Concept Definitions NCI/PT | An established society, corporation, foundation or other organization founded and united for a specific purpose, e.g. for health-related research; also used to refer to a building or buildings occupied or used by such organization. NCIt:Institution- same as above Birnlex:2085 - http://bioportal.bioontology.org/ontologies/29684?p=terms&conceptid=birnlex_2085 LNC:LP76237-4 - http://bioportal.bioontology.org/ontologies/44774?p=terms&conceptid=LP76237-4 ??? ACGT:institution - no definition SNOMED:385437003 - http://snomed.vetmed.vt.edu/SCT/ISA.cfm?SCT_ConceptID=385437003&CFID=2386243&CFTOKEN=16757988 I think I'll stick with obi:organization.

[zhengj2007]

From alanruttenberg@gmail.com on July 08, 2011 13:07:47

It is outside the scope of OMRSE as well. Clinical institution, medical institution, etc would be in scope for that.

[zhengj2007]

From MBrochhausen@gmail.com on July 11, 2011 07:19:07

Why should OMRSE refrain from representing institution when OBI represents organization"? We will, of course, represent it out of necessity as well. OBI "organiztion" probably is a good start, but from my few courses in social science I seem to remember that there was some significant difference between organizations and institutions. I'll check that. If we represent institution in OMRSE we will, of course, give it to anyone creating an OBO Foundry General Social Entity Ontology as soon as someone starts it. In the meanwhile I think we need proper definitions of things. Since Medical Insitutions are in scope and since no Ontology of General Social Entities exists, we need to give a tentative definition that tells that what kind of a thing a Medical Institution is apart from its differentia. Otherwise we will loose ourselves and only play games with words.

[zhengj2007]

From alanruttenberg@gmail.com on July 11, 2011 08:49:40

regarding "Why should OMRSE refrain from representing institution when OBI represents organization"? We will, of course, represent it out of necessity as well. OBI "organiztion" probably is a good start, but from my few courses in social science I seem to remember that there was some significant difference between organizations and institutions. I'll check that. "

I didn't say OMRSE should refrain from this, I only noted the scope, in response to what seemed to be an implication that it was. But perhaps you implied that as you think that ontologies should stay in scope when possible ;-)

Interested to hear the difference between organization and institution - particularly those that will matter in the context of OMRSE.

[zhengj2007]

From MBrochhausen@gmail.com on July 11, 2011 09:14:22

Well, I think I menat to imply that given that we are refering to more primitive terms in our defintion we need to be ready to define those and not hide behind out-of-scope-arguments ;-). As I said once the big social entity ontologies within the Foundry arrives I'd happily hand institution over to them, so no harm will be done (but a lot of good might be achieved regarding re-use, becuase external user might be able to understand what we talk about in our ontology).

Furthermore, I am not sure I promised anything like a difference between organizatio and institution that matter to OMRSE, but let me check with the experts and then I'll let you know....