Write BEP for special EvidenceType annotation about bep HOT 9 CLOSED

Do we need this? The Evidence is helpful, but it's not a requirement in either BELScript or BEL Nanopubs (only requires: type, citation, assertions). If the Evidence isn't available, one would naturally look to the citation. I see the Evidence as an aid to creating and reviewing Nanopubs, but in a case where the Nanopub is created from a database entry, there is no need for an Evidence statement.

from bep.

I am also not sure on the use case for this, my understanding is that evidence is optional (but strongly recommended)

I do see a potential use case to more obviously distinguish whether or not evidence is a direct quote from the text.

from bep.

Interesting - I was under the impression that the Evidence is required for all statements in a BEL Script. Currently, PyBEL enforces this assumption and I've never encountered BEL Script where this wasn't the case.

However, ontological statements using relations like hasVariant, hasComponent, hasReactant, etc. (e.g., p(X) hasVariant p(X, pmod(Y))) don't have evidences assigned to them.

As a follow up, I'm curious what citation (and/or evidence) do you think is appropriate for statements generated by orthologization? Or ontological statements?

from bep.

from bep.

The old OpenBEL BEL Framework treated the statement expansions (hasVariant, etc.) as edges associated with the same citation, evidence, annotations as the primary statement. I found this valuable since the existence of a variant, modification, activity, etc. could be traced back to a supporting citation. At the same time, I found it frustrating since once any of these entities appeared in more than a few statements (e.g., the kinase activity of AKT1), it was difficult to wade through all of the support.

from bep.

from bep.

@wshayes that sounds fine to me - but the only point to clarify is when a user is writing BEL statement (not inferred) should there always be an evidence except for the has* relations? Are users even allowed to write has* relations themselves?

from bep.

from bep.

I think I'm content with dismissing this issue. We can close it after the meeting today.

from bep.