Unqualified edges?

[cbizon]

For this query:
Match (a:biolink:ChemicalEntity {id:"PUBCHEM.COMPOUND:2083"})-[x]-(b:biolink:Gene {id:"NCBIGene:154"}) return *

There are two edges that I think should have been converted to qualified edges.

One is an "affects_abundance_of" from hmdb and one is "decreases molecular interaction" from ctd.

This may be coming down to a bl_lookup problem?

[EvanDietzMorris]

These came from deprecated predicate mappings that existed in biolink 3.0.3 (but were removed in 3.1.0):
CTD:decreases_molecular_interaction_with -> decreases molecular interaction
CTD:affects_abundance_of -> affects abundance of

Below are other predicates with CTD prefixes that need to be updated in Data Services, the biolink model, or bl lookup. Most of them are not in biolink model 3.0.3 or 3.1.0.

HMDB:
increases_transport_of

CTD:
affects_activity_of, affects_degradation_of, affects_localization_of, affects_metabolic_processing_of, affects_molecular_modification_of, affects_response_to, affects_secretion_of, affects_stability_of, affects_synthesis_of, affects_transport_of, affects_uptake_of, decreases_molecular_interaction_with, increases_molecular_interaction_with, molecularly_interacts_with

[cbizon]

So were these built with 3.1.0? I think we were going to use 3.0.3 to get around this problem. When I use bl-lookup with 3.0.3, CTD:increases_transport_of resolves.

Once BMT is updated to handle these mappings, we should be able to move to versions > 3.0.3

[EvanDietzMorris]

The graphs those edges came from were built with bl lookup 3.0.3. I just also checked the biolink model 3.0.3 and 3.1.0. The issue I foresee with the others is that many of the CTD:xxx mappings, such as CTD:increases_transport_of, were not explicitly mapped in the biolink model 3.0.3. Bl-lookup or bmt must have had a special rule for those, or something of that nature. In biolink model 3.1.0 increases_transport_of is not a predicate at all - so these special mappings will likely break if they are not handled.