search

EBISPOT / OLS

Ontology Lookup Service from SPOT at EBI
http://www.ebi.ac.uk/ols
Apache License 2.0
97 stars 40 forks source link

FoodOn ontology showing amphibian and fish under mammal #363

Closed ddooley closed 4 years ago

ddooley commented 4 years ago

This is an odd one. The main foodon.owl ontology file version date that is listed under OLS is 020-05-06, which is about a week old, and yet under mammal in foodon.owl (github versions) that go back at least a month or two, "amphibian" and "fish" are not under mammal. I'm wondering if one of the foodon files - perhaps the main one which would be responsible for the above problem, is months out of date, due to some loading problem. I have corrected a few things in the main file which may trigger a clean load of that. Could you check? Much obliged. Damion

jamesamcl commented 4 years ago

We had an issue recently with some ontologies not updating. We will ensure that foodon is re-indexed in the next run (early next week), and hopefully then the problem will be resolved going forward.

jamesamcl commented 4 years ago

Hi, just an update on this. FOODON seems to be updating regularly in OLS. So I loaded it manually from http://purl.obolibrary.org/obo/foodon.owl into an empty OLS instance, and I also see fish listed under mammal.

It's possible that it's an issue with how OLS loads ontologies, which will require further investigation. I will keep you informed!

cmungall commented 4 years ago

I think ENVO has the same issue cc @wdduncan @pbuttigieg

jamesamcl commented 4 years ago

I am not sure this is an OLS issue. Loading the latest version of FOODON into Protegé and running HermiT 1.4.3.456 also shows fish as a subclass of mammal. This seems to be because there is an equivalence between vertebrate animal and mammal?

Screenshot 2020-06-03 at 15 58 22 Screenshot 2020-06-03 at 15 57 49
ddooley commented 4 years ago

Hmm. Admittedly that equivalence needs to be fixed, thanks for pointing that out. First question - is OLS implementing hermit inferred classes in displayed ontologies? Secondly: I'm applying hermit 1.4.3.456 to my github mirror /foodon.owl & and /src/ontology/foodon-edit.owl and /src/ontology/foodon-merged.owl (it has no import files) and am still not seeing fish under mammal in any of those: image I'll fix the equivalency and post it asap. But unless I have some Protege reasoner setting different from your set, I fear the problem will remain.

ddooley commented 4 years ago

I'll add that our FoodOn "mammal" was meant to be a subclass of both vertebrate animal and vertebrate animal food product , and 'in taxon' some Mammalia.

ddooley commented 4 years ago

Actually, looking at fish equivalence, i see now you are right! My bad. I will fix ASAP. Much obliged. So this does teach me something - that you are running a reasoner to generate inferred classes for OLS. For some reason I didn't think OLS was doing that. That knowledge will help me going forward.

ddooley commented 4 years ago

I'll close this now, seeing that I've just released a fix to this. Thank you James for your help here.

jamesamcl commented 4 years ago

No problem! Glad it's sorted.