ncats-tangerine / biolink-beacon Goto Github PK

In the current implementation, data without a relation is filtered out. Would it be possible to include this data and allow it to be empty or null?
https://github.com/NCATS-Tangerine/biolink-beacon/blob/de04cf1/main.py#L121

Also, phenotype should be lower case when accessing the category in solr and/or biolink.
https://github.com/NCATS-Tangerine/biolink-beacon/blob/de04cf1/main.py#L335

Right now the Biolink beacon is just reporting back on 500 errors that it receives from Monarch. This might be useful for now for debugging, but in the future it should just return an empty response and hide the Monarch errors.

The /kmap is implemented based on direct query of Biolink but times out. We may need a caching strategy here(?).

Concept FlyBase:FBgn0000527 from statement biolink:f6b8d12d-04e9-4d2b-a376-bdf17932591c does not have the same name and type as result from concept search:
https://biolink-kb.ncats.io/concepts?keywords=e&size=1
https://biolink-kb.ncats.io/statements?s=FlyBase:FBgn0000527&size=100

Detail type is not the same as concept type:
https://biolink-kb.ncats.io/concepts?keywords=e&size=100
https://biolink-kb.ncats.io/concepts/FlyBase:FBgn0000527

This is probably because biolink allows for multiple concept types, whereas the beacon only allows for one. I suggest updating the beacon API to allow for multiple concept types.

Predicate now has edge_label and relation, and I don't think I've implemented this correctly.

At present there seem to be a significant number of statements from Biolink that are not making it into the aggregator. One issue is the fact that associations lacking a relation are not pulled, but this should be resolved on the Biolink end soon (see #7).

This is likely not the only cause of the problem, as there are examples of statements that have relations in Monarch. e.g. ALS phenotype annotations have asserted relationship (has_phenotype) but are not returned by tk.bio.

@kshefchek has made a preliminary fix at level of Monarch solr index that is not yet in production - it may be worth setting the beacon aggregator to pull from beta to see if fix propagates from solr through biolink and into the aggregator. We should evaluate once the fix is propagated to tk.bio.

Underlying Biolink API may not yet support this properly... need to fix? See.(monarch-initiative/biolink-api#114)

This works: https://biolink-kb.ncats.io/statements?s=NCBIGene:14571
This does not: https://biolink-kb.ncats.io/statements?s=NCBIGENE:14571

Monarch is expected to have HP term annotation (e.g. "Disease has_phenotype HP_term?") but for some reason, these terms are not accessed by the current beacon?