there is no spatial disjointness between pancreas and intestine but even so you don't mean that every instance of a delta cell is found in the spatial overlap of these.
Another way to think about this: the true path rule in GO
The correct way to do this is to make 3 subtypes as in CL
is_a CL:0000066 ! epithelial cell
is_a CL:0000150 ! glandular epithelial cell
is_a CL:0000164 ! enteroendocrine cell
is_a CL:0000502 ! type D enteroendocrine cell *** [SYNONYM: "D cell" (exact)]
is_a CL:0000173 ! pancreatic D cell
is_a CL:0002265 ! type D cell of colon
is_a CL:0002266 ! type D cell of small intestine
is_a CL:0002267 ! type D cell of stomach
I think you should also xref the generic CL class if your delta cell is generic
The relationship to pancreas is redundant with the islets relationship; let's ignore that for now
If you say
Then think of the venn diagram with the spatial extent of A, B and C. X is in the intersection of all 3. I don't think this is what you mean.
In uberon, stomach is spatially disjoint with intestine so XAO will cause inconsistencies when merged. See: https://github.com/obophenotype/uberon/wiki/Part-disjointness-Design-Pattern
there is no spatial disjointness between pancreas and intestine but even so you don't mean that every instance of a delta cell is found in the spatial overlap of these.
Another way to think about this: the true path rule in GO
The correct way to do this is to make 3 subtypes as in CL
I think you should also xref the generic CL class if your delta cell is generic