search

HumanCellAtlas / ontology

3 stars 1 forks source link

Unsatisfiable classes discovered while trying to build HCAO #123

Closed ghost closed 1 year ago

ghost commented 1 year ago

Whilst trying to make a release of HCAO, 1050 unsatisfiable classes (mostly in Uberon) were noted, preventing the build of the ontology.

After running make all, part of the output is:

import.owl --input ../imports/duo.owl --input clo_import.owl --input ncit_import.owl -o hcao_merged.owl &&\
    ../../bin/robot remove --input hcao_merged.owl -T terms_to_delete.txt -p false -o hcao_cleaned.owl &&\
    ../../bin/robot reason --input hcao_cleaned.owl -r ELK reduce -r ELK annotate -V http://ontology.data.humancellatlas.org/ontologies/hcao/releases/`date +%Y-%m-%d`/hcao.owl -o hcao.owl
ERROR There are 1050 unsatisfiable classes in the ontology.
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/UBERON_0016440
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/UBERON_0016442
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/GO_1902746
...
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/UBERON_0001570
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/CL_0000561
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/UBERON_0001577
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/UBERON_0002668
ERROR     unsatisfiable: http://purl.obolibrary.org/obo/UBERON_0001578
make: *** [hcao.owl] Error 1

Running a robot report for random unsats against the hcao_cleaned.owl file reveals the following (edited to fit in comment):

type 1 cone bipolar cell (sensu Mus) SubClassOf Nothing

amacrine cell SubClassOf Nothing

type 4 cone bipolar cell (sensu Mus) SubClassOf Nothing

retinal cone cell SubClassOf Nothing

OFF-bipolar cell SubClassOf Nothing

type 3 cone bipolar cell (sensu Mus) SubClassOf Nothing

choroidal cell of the eye SubClassOf Nothing

retina horizontal cell SubClassOf Nothing

scleral cell SubClassOf Nothing

retinal ganglion cell SubClassOf Nothing

cone retinal bipolar cell SubClassOf Nothing

type 5 cone bipolar cell (sensu Mus) SubClassOf Nothing

corneal epithelial cell SubClassOf Nothing

Mueller cell SubClassOf Nothing

Axiom Impact

Axioms used 20 times

Axioms used 19 times

Axioms used 12 times

Axioms used 9 times

Axioms used 5 times

Axioms used 3 times

Axioms used 2 times

Axioms used 1 times

Ontologies used:

ghost commented 1 year ago

FYI, @shawntanzk, @anitacaron ++ @dosumis, @cmungall

ghost commented 1 year ago

After discussing with @anitacaron and @shawntanzk, lens of camera-type eye was excluded from HCAO. Number of unsats has been reduced to 547.

Report (of random 20) below.

starburst amacrine cell SubClassOf Nothing

flag amacrine cell SubClassOf Nothing

flat bistratified amacrine cell SubClassOf Nothing

A2 amacrine cell SubClassOf Nothing

A2-like amacrine cell SubClassOf Nothing

retina horizontal cell SubClassOf Nothing

retinal pigment epithelial cell SubClassOf Nothing

H1 horizontal cell SubClassOf Nothing

broad diffuse amacrine cell SubClassOf Nothing

DAPI-3 amacrine cell SubClassOf Nothing

flag B amacrine cell SubClassOf Nothing

diffuse bistratified amacrine cell SubClassOf Nothing

H2 horizontal cell SubClassOf Nothing

flag A amacrine cell SubClassOf Nothing

internal epithelial cell of tympanic membrane SubClassOf Nothing

AB broad diffuse-2 amacrine cell SubClassOf Nothing

AB diffuse-1 amacrine cell SubClassOf Nothing

diffuse multistratified amacrine cell SubClassOf Nothing

monostratified amacrine cell SubClassOf Nothing

AB broad diffuse-1 amacrine cell SubClassOf Nothing

Axiom Impact

Axioms used 19 times

Axioms used 18 times

Axioms used 5 times

Axioms used 3 times

Axioms used 1 times

Ontologies used:

shawntanzk commented 1 year ago

Screenshot 2022-12-16 at 13 43 26

This has already been changed in uberon, I've just checked the new release and it is no longer material. This is probably a pipeline updating issue or something?

anitacaron commented 1 year ago

The class internal epithelial cell of tympanic membrane was unsat because an intermediate file in the HCAO pipeline was not updated with the new UBERON release.

ghost commented 1 year ago

After excluding the terms below and running make all -B, no unsats appeared.

http://purl.obolibrary.org/obo/UBERON_0000965 # lens of camera-type eye http://purl.obolibrary.org/obo/UBERON_0003902 # retinal neural layer http://purl.obolibrary.org/obo/CL_0000745 # retina horizontal cell

dosumis commented 1 year ago

Missed this. I'm pretty sure we excluded disjoints in the past. Presumably that step has been lost in editing the Makefile at some point. I think worth adding that step back before the next release. The excluded terms look useful.

On Fri, 16 Dec 2022, 17:13 bvarner-ebi, @.***> wrote:

Closed #123 https://github.com/HumanCellAtlas/ontology/issues/123 as completed.

— Reply to this email directly, view it on GitHub https://github.com/HumanCellAtlas/ontology/issues/123#event-8058452659, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAA3RRY64Y6N3F7WLSMTOFDWNSPJ7ANCNFSM6AAAAAATAX6JZY . You are receiving this because you were mentioned.Message ID: @.***>

dosumis commented 1 year ago

Although I guess it might fix itself with Uberon update.