switch branching_part_of use in the vasculature to connecting branch of

[dosumis]

branching_part_of is a subproperty of part_of, so use of it leads to distantly connected parts of the vasculature appearing to be parts of each other. the replacement is a subroperty of connected_to, so does not cause this problem

TBC: change the name of connecting_branch_of in RO to branches_from?

[aleixpuigb]

These are the blood vessels that are branching_part_of another blood vessel: https://api.triplydb.com/s/ISg2WL4YA
Need to do sorting of which ones are correct and which ones need to be changed.

[aleixpuigb]

This is the list of terms. The column connected to indicates that the object property needs to be changed. The column Other indicates if AS does not have the object property or if it has multiple and one needs to be removed.

[dosumis]

I guess this table just has the subject terms? I think it would be easier to automate this if we have a table of triples (subject , relation, object) that need to be changed. It should be easy to do this via UberGraph queries.

subject_label	subject_id	relation	object_label	object_id

This can easily be turned into a robot template or used as source for a SPARQL update to remove the old relations.

[dosumis]

@anitacaron we need a query like this

SELECT * WHERE {
  ?subject_vessel rdfs:subClassOf vessel: .
  ?object_vessel rdfs:subClassOf vessel: .
  ?subject_vessel branching_part_of: ?object_vessel .
  ?subject_vessel rdfs:label ?s_label .
  ?subject_vessel rdfs:label ?o_label .
}

But where the first 2 clauses query the redundant graph the 3rd queries the non-redundant graph and the last two the ontology. (Might also work if the only the third is restricted to nonredundant)

[aleixpuigb]

As there are some grouping classes of branches 'branch of ...' , making ?subject_vessel branching_part_of: ?object_vessel . non-redundant misses all the subclasses, which also have the object property 'branching part of'. However, as the grouping class already has this relation and it should be inherited, I would favour to delete them (manually?).
Here is the SPARQL query from Anita

I think it would be easier to automate this if we have a table of triples (subject , relation, object) that need to be changed.

Here is the table

[dosumis]

The safest way to identify what's already in the ontology would be to have this clause run on the ontology graph:

?subject_vessel branching_part_of: ?object_vessel .

as this will identify OWL axioms you need to target.

@anitacaron should be able to help you write SPARQL that deals with blank nodes.

[aleixpuigb]

Update uberon-editor-sop.md file accordingly

[aleixpuigb]

So far I have:

• Curated all vasculature that had to change branching part of to connecting branch of
• Using SPARQL update, change the object properties

However, we also want to create a pattern for vasculature. @dosumis would it be on the scope of this ticket?

[dosumis]

However, we also want to create a pattern for vasculature. @dosumis would it be on the scope of this ticket?

I think so. Go for it.