search

NCATSTranslator / Feedback

A repo for tracking gaps in Translator data and finding ways to fill them.
7 stars 0 forks source link

Does this support graph support the question? #799

Closed sstemann closed 1 month ago

sstemann commented 3 months ago

In Prod, Christine ran MVP2 > what chemicals may downregulate SCN9A https://ui.transltr.io/main/results?l=SCN9A%20(Human)&i=NCBIGene:6335&t=2&r=0&q=4dc7c935-ab1e-4dc7-bf67-ee6bc6c12955 > search Urokinase Prod PK: 4dc7c935-ab1e-4dc7-bf67-ee6bc6c12955

This is reproduced in Test:
https://ui.test.transltr.io/main/results?l=SCN9A%20(Human)&i=NCBIGene:6335&t=2&r=0&q=6e309577-ab46-4599-835f-d99d46051c98 > search Sodium Chloride Test PK: 6e309577-ab46-4599-835f-d99d46051c98

There are quite a few support paths that read: Chemical of a gene of input gene

The result would be an increase of input gene.

It looks to be coming from Unsecret Agent, in both Test and Prod. Asking @kaiwenho and @webyrd to take a look.

Prod image

Test image

kaiwenho commented 3 months ago

Thank you, @sstemann, for pointing it out. Unsecret identified the logic error in the MVP2 query results. We've corrected the issue, and the update is now in CI and Test. The test result you highlighted appears correct to me: Sodium Chloride increases SCN2A and SCN2A decreases SCN9A. This implies that SCN9A would be further decreased due to the increase in SCN2A. [Please excuse my simplified expression of the directional terms.]

sstemann commented 3 months ago

i think this is fixed in Eel, but it's looking like i didnt actually replicate it in Test, so I guess we will need to re-look at Prod once Eel is in Prod.

sstemann commented 1 month ago

retesting in Prod and looks good