Need to decide on valid entity types and corresponding shapes

[vanaukenk]

I am combining tickets #26 #31 #50 into this one, new ticket.

Models fail validation, I believe, because the entity types do not match what is in the ShEx.

In the gpi 2.0 specs, we propose using SO terms to define entity types, although we know that this doesn't cover protein-containing complexes and we need a GO term for that.

I am wondering if we need to add to, or replace, ChEBI's information biomacromolecule in the ShEx with SO term(s).

@balhoff @goodb

See also: https://github.com/geneontology/go-annotation/issues/2740

[goodb]

@vanaukenk could you share the .ttl file for the model shown there? Or just send me a link?

That looks like a problem with the validator. In the json there, I see an inferred type of CHEBI_33695 which should line up with InformationBiomacromolecule in the shex schema and validate. This sounds familiar - its possible I have something locally that fixes it that hasn't been made live.

[vanaukenk]

Thanks @goodb

This is the test model I made: http://noctua-dev.berkeleybop.org/editor/graph/gomodel:5df94d0c00000005

[goodb]

@vanaukenk I believe this PR should fix the problem: https://github.com/geneontology/go-shapes/pull/174

@kltm once the PR is accepted, we will need a noctua-dev restart. (Any of us can accept the PR any time, all tests are good. but I hoped @balhoff might have a look in case it raises other questions).

[goodb]

Okay merged. ready for noctua-dev restart.

[goodb]

Didn't mean to close.. opening up for @vanaukenk to close if she chooses.
It looks like the restart blew up the test model. I made another one and things are looking good: http://noctua-dev.berkeleybop.org/editor/graph/gomodel:5e18e83800000000

Note that the chemokine binding node is red because a function is not allowed to regulate a process - (a control to show that the shex test is running). Take away that negatively regulates relation and the model should validate.

[vanaukenk]

Thanks for the fix @goodb The validation for entity types looks good. I added a ncRNA and a protein-containing complex to the test model for good measure, and they validate, as well.