merge DOID:2410 'skin granular cell tumor' + DOID:2411 'granular cell tumor'

[cmungall]

DOID:2411 'granular cell tumor' is undefined; the label is very vague, 'granular cell' is not v specific

However, it's already classified as a 'skin benign neoplasm', which suggests it is more specific than the label suggests

Further more this would make it equiv to DOID:2410

[elviram]

We will look over all the cancer related terms you pointed out, while also looking at them from a higher level. We are also working with NCI along those lines.

[cmungall]

Do you have any design pattern docs from your work with NCI? I'd be interested in helping review these and automating checking / term generation

[elviram]

No, there are no design pattern docs. For now we have a mapping file.

[cmungall]

is the mapping file in VCS?

it's a good start but I think we should move away from n*(n-1) disease term mapping files and take a more proactive integration approach at a higher level. If we use the same axiomatic pattern for defining cancers, e.g. using site of origin, and we have our tissues and cells integrated, the integration at the disease level becomes trivial.

[elviram]

It's a classic xls file: DOID, term/synonym, concept name and concept code.
Your suggestion is interesting, but we need to discuss this more. We'll have a chat about it soon.