identifier in the Disease Ontology database
List of this constraint violations: Database reports/Constraint violations/P699#Format, hourly updated report, SPARQL, SPARQL (new)
|Distinct values: this property likely contains a value that is different from all other items. (Help)|
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P699#Unique value, SPARQL (every item), SPARQL (by value), SPARQL (new)
|Single value: this property generally contains a single value. (Help)|
Exceptions are possible as rare values may exist. Known exceptions: glucose-6-phosphate dehydrogenase deficiency (Q848343)
List of this constraint violations: Database reports/Constraint violations/P699#Single value, SPARQL, SPARQL (new)
|Qualifiers “mapping relation type (P4390)”: this property should be used only with the listed qualifiers. (Help)|
List of this constraint violations: Database reports/Constraint violations/P699#Allowed qualifiers, hourly updated report, SPARQL, SPARQL (new)
|Scope is as main value (Q54828448), as references (Q54828450): the property must be used by specified way only (Help)|
List of this constraint violations: Database reports/Constraint violations/P699#scope, hourly updated report, SPARQL, SPARQL (new)
- I do not see any arguments pro or contra. Both variants are equals as I see. Wikidata does not have standard policy about database name prefixes. ID without DOID: was used by historical reasons. I think it is better to continue this practice until we does not have some other arguments. — Ivan A. Krestinin (talk) 18:50, 27 July 2015 (UTC)
- I would argue in favor of adding the prefix "DOID:" to all disease ontology identifiers for two reasons. First, on the original source the prefix is an intrinsic part of the identifier (e.g. []). Secondly, external resources such as the miriam registry also include the prefix DOID in its pattern for Disease Ontology. So it seems more common to include the DOID prefix to an identifier from the Disease Ontology. Andrawaag (talk) 20:17, 27 July 2015 (UTC)
- I support these arguments. In addition, IDs with prefix are easier to recognize if one is outside of its Wikidata web page. Also, practically all other biology and chemistry identifiers are imported to Wikidata with their prefixes, retaining the prefix for the disease ontology ID therefore adds to consistency. Sebotic (talk) 21:02, 27 July 2015 (UTC)
- The prefixes are an integral part when working with ontologies, they are part of the URI of each term. In OWL when importing a term to a new/different ontology it is standard practice to do that according to the MIREOT principles. The prefixes are essential for referencing. They should not be removed.Emitraka (talk) 22:14, 27 July 2015 (UTC)
- This rule doesn't seem to be applied consistently; for example the IDs for Uberon are unprefixed: https://www.wikidata.org/wiki/Q23640 - I prefer prefixing but note that it seems impossible to expand this to the OBO PURL using the current system which assumes that the complete ID can be affixed to a base URL. This is because OBO shortform IDs should be understood as CURIEs, and some limited level of CURIE awareness should be built into the URL expansion system on wikidata and wikipedia. This confusion frequently arises, see some of the entries in identifiers.org and see also this discussion which recapitulates the same issues: https://github.com/ga4gh/schemas/issues/165 Cmungall (talk) 15:28, 26 August 2015 (UTC)
We have a number of arguments pro DOID: prefix now. Minor argument about historical reasons is not important in this situation. So I am starting migration process. — Ivan A. Krestinin (talk) 22:17, 1 August 2015 (UTC)
- There is now another service which uses IRI that are different from DOID:\d+ scheme, see, e.g., https://www.ebi.ac.uk/ols/ontologies/doid/terms?iri=http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FDOID_10652# This also presents symptoms, e.g., for paralysis: http://purl.obolibrary.org/obo/SYMP_0000030 — Finn Årup Nielsen (fnielsen) (talk) 20:01, 29 January 2018 (UTC)
I propose not having the prefix be part of the ID:
- There is the practical argument of simply defining a "concept base URI" of "http://purl.obolibrary.org/obo/DOID_", or, in Wikidata terms, a formatter URI for RDF resource (P1921) of "http://purl.obolibrary.org/obo/DOID_$1".
- The argument has been made the on some websites (example), one sees "DOID:769". However, such websites also things like "MESH:D009447", "NCI:C3270", ...; in general, this looks like a listing of "property:value" pairs. This just means that a number is only useful in a given context, given by the property ("identifier type") to which this number pertains.
With over 7000 exceptions already for each the Single and Distinct constraints, I wonder if these should simply be removed. For example, I would like to annotate biological pathways with DOIDs, using the "mapping relation type" of "related match" to express the fundamental relationship between particular pathways and particular diseases. This is often not a one-to-one mapping however (breaking Single) and multiple pathway items will relate to the same DOID (breaking Distinct). Should I ignore this oft violated constraint? Should the constraint be removed? Or should I not annotate Wikidata items with this information? Thanks for feedback and advice. - AlexanderPico (talk) 19:55, 27 September 2019 (UTC)
- After further consideration and discussion with other wikidata curators, I'm going to try annotating biological pathways with wikidata items representing diseases which are themselves identified by DIODs. This way we can help fix the exceptions and not contribute further to the problem. This requires a new property, e.g., something like "pathway annotation" to capture many-to-many associations between pathway instances and diseases (see proposal for new property). --AlexanderPico (talk) 17:04, 9 October 2019 (UTC)