duplicate results returned for pk_anatomical_class

[wdahdul]

pk_anatomical_class() returns duplicate superClassOf terms for post-composed terms. In the example below, see rows 3+4 and 5+6:

> pk_anatomical_class("pectoral fin")

                                         label
1                           archipterygial fin
2                     pectoral fin (zebrafish)
3        pectoral fin and (part of some larva)
4        pectoral fin and (part of some larva)
5 pectoral fin and (part of some larval stage)
6 pectoral fin and (part of some larval stage)

@id
1                                                                                                                                                                                                                                             http://purl.obolibrary.org/obo/UBERON_4200003
2                                                                                                                                                                                                                                                http://purl.obolibrary.org/obo/ZFA_0001161
3 http://purl.org/phenoscape/subclassof?value=%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0000151%3E%0A+and+%28%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FBFO_0000050%3E+some+%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0002548%3E%29#c9f2c0d3-6dcb-4551-ab76-3e23b0bddc4e
4 http://purl.org/phenoscape/subclassof?value=%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0000151%3E%0A+and+%28%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FBFO_0000050%3E+some+%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0002548%3E%29#d6516e85-2134-42bf-a2fe-e23c8cdfada4
5 http://purl.org/phenoscape/subclassof?value=%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0000151%3E%0A+and+%28%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FBFO_0000050%3E+some+%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0000069%3E%29#44d06b02-a775-4cd9-bcb5-66f6fc923972
6 http://purl.org/phenoscape/subclassof?value=%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0000151%3E%0A+and+%28%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FBFO_0000050%3E+some+%3Chttp%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0000069%3E%29#db58593b-6f70-4dcc-b5f8-7f22aa623949

[hlapp]

@balhoff the question really is why there are duplicate terms in the database that have identical semantics (as post-compositions) but slightly different IDs. So presumably this is the wrong tracker; which one should it go to? (There's a "Transfer Issue" feature now.)

[balhoff]

Thanks, I transferred! The issue is that post-compositions are defined as subclasses of the entered expression. They all have different IDs because they shouldn't be equivalent. One reason for this is that curators have put in post-compositions to describe an "absent" structure, and it would be too strong to say that there are no instances at all of the entered expression.

Perhaps this should be treated as a curation issue instead of creating this complicated data model.

[hlapp]

Perhaps discuss at next team call? I'm not sure I'm following. I.e., I'm not following why the two classes pectoral fin and (part of some larva) have different semantics and thus shouldn't have the same ID.

[hlapp]

@balhoff and @wdahdul we should discuss this at an upcoming team call so we have time for designing and implementing a solution if needed for the SSB workshop. If it's not needed for that, then remove the project assignment.

[balhoff]

The annotation issue is much more widespread than I realized. Here is a query showing all the Q=absent annotations with a post-composed entity: https://api.triplydb.com/s/OVzITJ63

There are 1855 to look at!

PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX ps: <http://purl.org/phenoscape/vocab.owl#>
PREFIX absent: <http://purl.obolibrary.org/obo/PATO_0000462>
PREFIX has_character: <http://purl.obolibrary.org/obo/CDAO_0000142>
SELECT DISTINCT ?matrix_label ?matrix_char_label ?state_label ?entity_label WHERE {
?phenotype ps:quality_term absent: .
?phenotype ps:entity_term ?entity .
?entity rdfs:label ?entity_label .
FILTER(CONTAINS(?entity_label, " some "))
?state ps:describes_phenotype ?phenotype .
?matrix_char ps:may_have_state_value ?state .
?matrix has_character: ?matrix_char .
?matrix rdfs:label ?matrix_label .
?matrix_char rdfs:label ?matrix_char_label .
?state rdfs:label ?state_label .
}
LIMIT 5000

[hlapp]

The terms with redundant labels do cluster pretty closely, though not more closely than other siblings that aren't redundant:

[wdahdul]

related to phenoscape/phenoscape-kb-web-ui#18