Disease (MONDO:002117 - Lung Cancer) <risk_affected_by> Named Thing

sstemann commented 2 years ago

Query: risk.json PK: f357829d-7d4f-4046-af62-2d497b277651 Control: Tobacco, smoking, asbestos, radon, benzene, etc Results Tracking Sheet

andrewsu commented 2 years ago

It's possible we were deploying right when this query was submitted, but in any case, BTE is returning as expected now: https://arax.ncats.io/?r=939d598b-918e-42eb-bc29-840697365140

In terms of positive controls:

Tobacco/smoking: BTE did not find UMLS:C0040329 or any of its 25 narrower concepts
asbestos: BTE found UMLS:C0003947 as result 90 (prod node normalizer doesn't resolve it to a human-readable label, but dev does)
radon: BTE did not find UMLS:C0034629
benzene: BTE did not find UMLS:C0005036

dkoslicki commented 2 years ago

@sstemann Just a small note: if you want the results to be hyperlinked, you can make the PK value point to: https://arax.ncats.io/?r=f357829d-7d4f-4046-af62-2d497b277651 so it will be clickable just like the query JSON link.

cbizon commented 2 years ago

We found the issue. It was related to an over-aggressive traffic filter trying to block log4j attacks.

Here's the results: https://arax.ncats.io/?source=ARS&id=e2800952-aae8-4605-97ce-4cfbc596934e

Looks like ARAGORN's ranking prefers finding Genes that predispose for Lung Cancer and then finding what they have in common.

I reran using "ChemicalEntity" instead of NamedThing. ChemicalEntity still includes protein/genes, but it is a little smaller than everything, and so the enrichment in Aragorn performs better. Those results are here: https://arax.ncats.io/?source=ARS&id=bf04c388-b4d2-482e-9ddc-abb92c6c81c8, and the groupings are related to gene ontology terms. So the top hits are proteins related to "positive regulation of phosphorylation" and "positive regulation of immune system"

NCATSTranslator / testing

Disease (MONDO:002117 - Lung Cancer) <risk_affected_by> Named Thing #161