BioLink API "GET /bioentity/phenotype/{id}/anatomy/" Not Functioning

[kshefchek]

Original ticket by @Superraptor in monarch-app

Hi! First, I'd like to say amazing work here, I'm a huge fan of the Monarch Initiative application and the API has been an amazing resource as of late.

Second, my apologies if this is the wrong place to post this. I'm specifically referring to the BioLink API documentation as seen here.

Essentially, there are two provided examples, "ZP:0004204" and "MP:0008521". The first provides the following output:

[
  {
    "label": "muscle pioneer",
    "categories": null,
    "types": null,
    "synonyms": null,
    "description": null,
    "consider": null,
    "deprecated": null,
    "id": "ZFA:0001086",
    "replaced_by": null
  }
]

While the second produces:

[]

I thought it might work with another Mammalian Phenotype identifier, so I tried MP:0003279 (aneurysm), but it still provided:

[]

Further, I tried a Human Phenotype Ontology identifier (HP:0100790, hernia) and got the same result. All of these calls resulted in a 200 response.

Is this how the function should be working? Or is it still in development? Thanks so much!

[cmungall]

we have this in scigraph_util

        objs = self.traverse_chain(id, [HAS_PART, INHERES_IN], "anatomical entity")

the actual path we need is has_part/{inheres_in|inheres_in_part_of}.

Currently the traverse_chain method doesn't support for path expressions... nor should it! No sense reinventing cypher / sparql property paths here. We should just pass this in as a direct cypher query

@kshefchek and @deepakunni3 should we take the opportunity to go with BOLT? Or just do this via /cypher/execute in the existing SciGraph API?

[kshefchek]

In the short term I think the only option is /cypher/execute until someone finishes https://github.com/SciGraph/SciGraph/pull/238

[deepakunni3]

@cmungall Do you want me to take a look at SciGraph/SciGraph#238?

It looks like the bulk of the work is done. The only thing pending is making it configurable.

@kshefchek Let me know if there is more to that PR than what I assume.

EDIT: nevermind, I misunderstood what was going on in that PR. Its still a WIP then.

[kshefchek]

@deepakunni3 it might not be too much work, I keep meaning to take a closer look.

@cmungall do we aim to support complex owl restrictions, such as http://compbio.charite.de/hpoweb/showterm?id=HP:0012792? I'm not sure how this is translated in scigraph.

[cmungall]

Hmm, I think I prefer it being outside the scope of SG and having ontologies produce pre-reasoned versions with shortcut relations connecting the phenotype directory with the a anatomy term.

But in the interim it may be easier to replicate the pattern in a path query

[kshefchek]

bump, @realmarcin was asking about this today.

See also https://github.com/biolink/biolink-api/issues/115