Skip to content

Replace chebi slim URL and migrate to ODK v1.6#41

Open
aleixpuigb wants to merge 4 commits intomasterfrom
replace_chebi_slim
Open

Replace chebi slim URL and migrate to ODK v1.6#41
aleixpuigb wants to merge 4 commits intomasterfrom
replace_chebi_slim

Conversation

@aleixpuigb
Copy link
Copy Markdown

@aleixpuigb aleixpuigb requested a review from matentzn January 19, 2026 14:57
@matentzn
Copy link
Copy Markdown
Collaborator

@aleixpuigb can you handle this one? Once QC passes, I think you can merge this but best check with @jamesamcl

@aleixpuigb
Copy link
Copy Markdown
Author

Will do, thanks!

@aleixpuigb
Copy link
Copy Markdown
Author

The ontology is inconsistent because the object property derives from is declared reflexive while also having a range of independent continuant. Reflexivity implies that every entity derives from itself, so anything participating in derives from is inferred to be an independent continuant. This becomes contradictory when roles are involved, since roles are subclasses of specifically dependent continuant, which is explicitly disjoint with independent continuant in BFO. As soon as a role is instantiated, it is inferred to be both dependent and independent, collapsing the ontology to inconsistency. The correct resolution is to remove the reflexivity axiom from derives from, as the relation is not reflexive in BFO/RO and is intended to relate distinct continuants; keeping the range restriction and BFO disjointness axioms intact preserves the intended semantics.

image

@matentzn
Copy link
Copy Markdown
Collaborator

I dont think the scatlas folks care about that axiom - feel free to remove it

@anitacaron
Copy link
Copy Markdown
Contributor

This might be coming from an import, scatlas ontology is only made of imports, so it would be better fix it on the source? 👀

@aleixpuigb
Copy link
Copy Markdown
Author

This might be coming from an import, scatlas ontology is only made of imports, so it would be better fix it on the source? 👀

Yes, it is coming from the Plant Trait Ontology (TO) import. I'm trying to find the best way to deal with, it, but even removing this problem, the majority of imported classes are unsatisfiables.

@matentzn
Copy link
Copy Markdown
Collaborator

If you tell me what needs doing I can fix it in https://github.com/Planteome/plant-trait-ontology

@aleixpuigb
Copy link
Copy Markdown
Author

aleixpuigb commented Jan 19, 2026

Thank you, it might be solved in 'to' but the imports haven't been refreshed in a long time, so I'm checking if it can be solved this way. I had a look in https://github.com/Planteome/plant-trait-ontology and I could not find the reflexivity axiom.

EDIT: The axiom was removed after refreshing imports.

Reflexivity axiom has been removed
@aleixpuigb
Copy link
Copy Markdown
Author

Currently there are 15471 unsatisfiable classes in the ontology. They should be solved in different ontologies, but I am not sure who would be responsible for this work. What should we do with this @jamesamcl?

@matentzn
Copy link
Copy Markdown
Collaborator

Can you run sh run.sh make explain_unsat and post the results here?

@aleixpuigb
Copy link
Copy Markdown
Author

aleixpuigb commented Jan 19, 2026

Since the file is a bit long to paste here (259 lines), I have attached the file:
explain_unsat.md

The default command only shows 10 random unsats I think:

robot --catalog catalog-v001.xml explain -i tmp/scatlas-preprocess.owl -M unsatisfiability --unsatisfiable random:10 --explanation tmp/explain_unsat.md

@matentzn
Copy link
Copy Markdown
Collaborator

lung cancer cell SubClassOf Nothing

lung cancer cell line SubClassOf Nothing

Burkitt lymphoma cell SubClassOf Nothing

melanoma cell line SubClassOf Nothing

blood serum SubClassOf Nothing

osteosarcoma cell line SubClassOf Nothing

myoblast cell line SubClassOf Nothing

A-549 cell SubClassOf Nothing

cell suspension culture SubClassOf Nothing

colonic cancer cell line SubClassOf Nothing

Axiom Impact

Axioms used 5 times

Axioms used 4 times

Axioms used 3 times

Axioms used 2 times

Axioms used 1 times

Ontologies used:

@matentzn
Copy link
Copy Markdown
Collaborator

Hmm.. Get clarity on whether this is something your team is obligated to work on; if scatlas ontology is dead, just abandon otherwise.. Cant see a super obvious fix.

@aleixpuigb
Copy link
Copy Markdown
Author

Thanks Nico, there are some things that can be fixed in EFO (EBISPOT/efo#2612) and I need to fix anyway, but it looks like there are many more problems.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants