OntologyClass metadata additions: alternative_names, alternative_identifiers, relations #2239
Conversation
…the OntologyClass class
…erm as a first step
…xact mappings and examples
|
src/schema/basic_classes.yaml
Outdated
| biolink: https://w3id.org/biolink/vocab/ | ||
| owl: http://www.w3.org/2002/07/owl# | ||
|
|
||
| default_curi_maps: |
There was a problem hiding this comment.
I stopped using default curie maps under @cmungall 's guidance a long time ago. I'd like the three of us to discuss this before merging, but that would push it back to the first week of December.
src/schema/basic_classes.yaml
Outdated
| OntologyClass: | ||
| aliases: ["OntologyClass"] |
There was a problem hiding this comment.
If I ask a blunt question it doesn't mean I think your idea was crazy. Just want to succinctly start some conversations.
- Why are you asserting the class name an alias?
- why are you using [short list] notation? I don't think we have any precedent for that in nmdc-schema
There was a problem hiding this comment.
@turbomam - I toyed with the idea of changing the name of the Class ("Class" is redundant and inconsistent in the context of this schema, but after some research -- I checked into Biolink, and other LinkML schemas, and I see that its a common name for this kind of class). I will remove the alias for now.
short list notation is a convention I use in other models, happy to remove the list.
src/schema/basic_classes.yaml
Outdated
| class_uri: nmdc:OntologyClass | ||
| description: >- | ||
| This class is used to represent ontology terms. |
There was a problem hiding this comment.
I would like to stay away form self-referential definitions. Would something like "An optimized, local representation of class defined in an external ontology" be OK?
There was a problem hiding this comment.
sure, it currently has no definition.
| - owl:Class | ||
| - schema:Class | ||
|
|
||
| OntologyRelation: |
| @@ -550,7 +550,7 @@ slots: | |||
| multivalued: true | |||
| description: >- | |||
| A list of alternative identifiers for the entity. | |||
| pattern: '^[a-zA-Z0-9][a-zA-Z0-9_\.]+:[a-zA-Z0-9_][a-zA-Z0-9_\-\/\.,]*$' | |||
| pattern: '^[a-zA-Z0-9][a-zA-Z0-9_\.]+:[a-zA-Z0-9_][a-zA-Z0-9_\-\/\.,\(\)\=\#]*$' | |||
There was a problem hiding this comment.
what are some examples of alternative identifiers using those additional characters?
Adds metadata to
OntologyClassso that its collection in Mongo and its table in Postgres can hold cross-references (curie form), synonyms, and term descriptions available from external ontologies. It also adds a new class,OntologyRelation,to capture NMDC-search/display-relevant axioms/relationships (in particular,subclassOfandpart_ofrelations).In particular,r in the example data, @turbomam @cmungall - I'd like some feedback on the difference between "is_a" and "BFO:0000050" as the value for "predicate" in an OntologyRelation (using "is_a" follows the obographs convention, but I picked
rdfs:subclassOfas the predicate for "is_a" because it sort of enables consistent processing of this field by downstream software?).These changes support generic ontology metadata loading into NMDC.
Some related tickets:
microbiomedata/nmdc-server#1388 - support more facets for GO in search
microbiomedata/pilot#17 - support metadata search from ontologies
microbiomedata/pilot#12 - support ENVO search facets
microbiomedata/nmdc-field-notes#187 - field notes ENV triad term selection
We also want to provide users with more details about our existing terms. For example, we could extend the widget on the data portal search to include ontology metadata like cross-references, definitions, etc, to enable more informed selections.
Storing these metadata in Mongo and, subsequently, in Postgres might make this data more accessible for downstream tools.