Import PATO:'chronological age'

[timalamenciak]

Per import process here: https://github.com/EnvironmentOntology/envo/wiki/Import-terms-from-other-ontologies - closes #1587

[pbuttigieg]

Hi @timalamenciak this sets up the import target, but you'll have to trigger the import process in the Makefile, in the docker container.

you'll see that the pato import owl file will change too

[pbuttigieg]

Thanks @timalamenciak

@matentzn @cmungall

I'm not sure why the make process for PATO generates an OWL file with so much CHEBI in it.

The PATO terms we need are in there.

Is this why there's a merge process downstream ?

[cmungall]

I'm not sure why the make process for PATO generates an OWL file with so much CHEBI in it.

See: https://oboacademy.github.io/obook/tutorial/project-ontology-development/?h=slme#extracting-modules
For the most accessible description.

I updated this to document the new subset method, which is what ENVO should use: https://robot.obolibrary.org/extract#subset

[pbuttigieg]

I'm not sure why the make process for PATO generates an OWL file with so much CHEBI in it.

See: https://oboacademy.github.io/obook/tutorial/project-ontology-development/?h=slme#extracting-modules
For the most accessible description.

It's still unclear why so many CHEBI terms are in the PATO import. But if this is expected behaviour, then @timalamenciak we can disregard

I updated this to document the new subset method, which is what ENVO should use: https://robot.obolibrary.org/extract#subset

We won't use OBO format anymore - it's a lock in and anti-FAIR. I assume robot can process other formats for the subset definition ?

[matentzn]

The reason you have Uberon classes in the PATO import is because Description Logic based SLME module extraction demands it (an ontology is the union of all its base axioms plus all imported axioms, and given a signature, you have to import all relevant axioms, including imported ones). However, we have long overcome this limitation in our build process - we just need some funding for ENVO to fix all of this.

Chris solution has nothing to do with OBO format, maybe this is a bit misleading in the description in the PR we merged. Its just a sane way to extract a less lossy MIREOT style module which is very prevalent in many OBO ontology development processes. But there is another way, which is using base files. This allows us to extract PATO terms from PATO while preserving any logical links to other ontologies as well. However, we need some amount of funding to implement this in ENVO.

This is unrelated though to this PR here and should be attacked separately.

[matentzn]

I left a comment to explain the unrelated axioms, but this should not be handled as part of this PR here!

[pbuttigieg]

The reason you have Uberon classes in the PATO import is because Description Logic based SLME module extraction demands it (an ontology is the union of all its base axioms plus all imported axioms, and given a signature, you have to import all relevant axioms, including imported ones). However, we have long overcome this limitation in our build process - we just need some funding for ENVO to fix all of this.

If you have overcome this somewhere in OBO, yet need ontology specific funding to deploy it over the hundreds of ontologies that are OBO Library assets, something is deeply wrong. Is there documentation you can point to that describes how to get from the current ENVO build process (which was OBO compliant) to the new one?

Chris solution has nothing to do with OBO format, maybe this is a bit misleading in the description in the PR we merged.

Yes it does. The command specifies OBO format input.

Its just a sane way to extract a less lossy MIREOT style module which is very prevalent in many OBO ontology development processes.

That's parallel - the process is fine, the apparent requirement for OBO files isn't.

But there is another way, which is using base files. This allows us to extract PATO terms from PATO while preserving any logical links to other ontologies as well. However, we need some amount of funding to implement this in ENVO.

Where is the documentation explaining how to do this? If it's standard OBO approaches, why does each ontology need dedicated funding to remain current? Something is not working.