Unsatisfiable classes revealed by ELK 0.5

[balhoff]

Classifying with ELK 0.5, which supports property ranges, reveals a pretty large number of unsatisfiable classes (5590) when the full RO is imported into Uberon (release http://purl.obolibrary.org/obo/uberon/releases/2018-07-30/ext.owl). You've got to import RO to see the problems.

[matentzn]

This is bad. Do you see any regularities in terms of what is going on? Always part of involved?

[balhoff]

I think this one is fairly typical:

[balhoff]

Deleting the axiom 'material entity' DisjointWith 'immaterial entity' eliminates all but 4 unsatisfiable classes.

[matentzn]

Yeah, I thought that might have been the cause. I found a bunch of those where anatomical lines and anatomical structures were mixed and matched together... Solving this will be very important.

[balhoff]

I guess the pivotal axiom is really 'has developmental contribution from' Range 'anatomical structure'.

[cmungall]

Yes, I think this should be generalized. Lexically this sounds strange to talk about spaces 'contributing', but the intent of the relation is literally the property chain hp o df

…

On 7 Sep 2018, at 8:46, Jim Balhoff wrote: I guess the pivotal axiom is really `'has developmental contribution from' Range 'anatomical structure'`. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: #1455 (comment)

[balhoff]

The range of 'has developmental contribution from' in RO has been loosened to 'anatomical entity'. This seems to have resolved nearly all of these problems. There are 23 remaining unsatisfiable classes (when you import both RO and uberon-caro-bridge into Uberon). The issue is things attached to a border of a bone, where the border is defined as an anatomical line (immaterial entity):

[cmungall]

OK, I think borders should be material, will change

…

On Mon, Mar 25, 2019 at 11:31 AM Jim Balhoff ***@***.***> wrote: The range of 'has developmental contribution from' in RO has been loosened to 'anatomical entity'. This seems to have resolved nearly all of these problems. There are 23 remaining unsatisfiable classes (when you import both RO and uberon-caro-bridge into Uberon). The issue is things attached to a border of a bone, where the border is defined as an anatomical line (immaterial entity): [image: deltoid] <https://user-images.githubusercontent.com/210210/54944786-a9e2ef80-4f0a-11e9-9875-e41680548f1a.png> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1455 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AADGOUSzh4FEy6LkW8ZcBAjpaXZieJ42ks5vaRYIgaJpZM4WbLjh> .

[balhoff]

Only one problem class right now:

[cmungall]

I think for now, we change the attachment site to be the scapula, and add a note that it's in the fossa

Removing incoherencies is number one priority, fine grained axiomatization can be added iteratively.

I made a ticket for the broader issues #2017. but this may take a while to resolve

[matentzn]

I neither see this showing up in uberon.owl nor uberon-edit.obo.. Which file/reasoner are you using?

[balhoff]

You need to import the full ro.owl and also uberon-bridge-to-caro.

[matentzn]

#2023 indeed fixes the issue..