MVP2 - Gene Results by Organism

NCATSTranslator / Feedback

A repo for tracking gaps in Translator data and finding ways to fill them.

7 stars 0 forks source link

MVP2 - Gene Results by Organism #781

Open sstemann opened 3 months ago

sstemann commented 3 months ago

I ran MVP2 what genes may be downregulated by Benazepril

Prod: https://ui.transltr.io/main/results?l=Benazepril&i=PUBCHEM.COMPOUND:5362124&t=4&r=0&q=75b35688-894b-4b8c-91c2-ab2ec256c394

Test: https://ui.test.transltr.io/main/results?l=Benazepril&i=PUBCHEM.COMPOUND:5362124&t=4&r=0&q=fd6faebf-131c-4b69-a29a-469a5643e7e0

ACE is returned as three results

34 paths for humans https://www.ncbi.nlm.nih.gov/gene/1636
4 paths for rabbits https://www.ncbi.nlm.nih.gov/gene/100009274
5 paths for rats https://www.ncbi.nlm.nih.gov/gene/100009274

i'm not sure if the UI should indicate this more? or should there be a way to merge/seperate?

dnsmith124 commented 3 months ago

@Genomewide what's your take here? I would think merging these might not be a good idea, since fact that they're associated with different species seems pretty important to me, at least at first glance. We could add the species to the label, like we do in the autocomplete

sstemann commented 3 months ago

i was thinking like a collapse/expand but not really sure how that would be possible

dnsmith124 commented 3 months ago

@sstemann I'll discuss with Andy and Reese and let you know what we come up with!

cbizon commented 3 months ago

This might be an opportunity for the kind of grouping that O&O has been discussing?

Genomewide commented 3 months ago

I wonder how many times this happens. I will do some spreadsheeting and see what I get back. I will only detect the ones that have the same gene symbol for homologs, though.

My first instinct is to have them merged, but that could get weird when homologs don't have the same name.

The simplest thing would be to have the species displayed in parenthesis and keep them separate.

OR we could have an inferred human gene with these as support graph edges. I.e. the drug regulates the rat homolog, there for it may regulate the human one? The more species that show the same thing the more likely it will work for the human homolog.

sstemann commented 3 weeks ago

seems related to #537

this still happens https://ui.test.transltr.io/main/results?l=Benazepril&i=CHEBI:3011&t=4&r=0&q=4c4debfe-6e1b-4f1b-9386-a63023a80225

if this is a grouping thing - should we check if annotator has species? @newgene