Use case: Drug repurposing for ADCY5

[andrewsu]

Gain-of-function ADCY5 mutations were described in this paper. Recently a ReFRAME screen was run in an in vitro assay to identify potential therapeutics, and apparently some reasonable hits have been identified.

We could use this as a Translator use case in one of two ways. First, we could blind ourselves to the screening results, run some repurposing queries using BTE to PREDICT potential therapies, and compare how those results compare to the screening results. Second, we could look at the screening hits and ask BTE to EXPLAIN potential mechanisms.

[kevinxin90]

@andrewsu Anywhere we can find the screening results from ReFRAME for this case?

[andrewsu]

I'm 99% sure we can get the screening results/hits, but do we want them now? We could blind ourselves to the screening results and attempt to predict compounds that would hit (the first use case I described above). Granted this is the much harder use case, but might be a fun challenging exercise... (Having said that, I can work on getting the results now if you think that's a better plan.)

[kevinxin90]

@andrewsu Initial results:

1. Path: ADCY5 -> (entity_negatively_regulated_by_entity) -> ChemicalSubstance

• Calcium (CHEBI:29320)
• CARBAMAZEPINE (CHEBI:3387)
• METOPROLOL (CHEBI:6904)
• CARVEDILOL PHOSPHATE (CHEBI:3441)
• Guanosine 5'-O-(3-Thiotriphosphate) (MESH:D016244)

2. Path with intermediates nodes: ADCY5 -> (entity_positively_regulated_by_entity) -> Gene -> (entity_negatively_regulated_by_entity) -> ChemicalSubstance

• CHLORAL HYDRATE (CHEBI:28142)
• path: ADCY5 -> positively_regulated_by -> KIF3A -> negativel_regulated_by -> CHLORAL HYDRATE

Will continue to try paths involving Pathways and BiologicalProcess involving cAMP biosysnthesis. But above are initial (and probably most confident) results from BTE.

[andrewsu]

Hmm, one thing we could do would be to constrain your final list to compounds that are in the reframe collection (but without any information on the compound activity). If I give you that list as a mix of DrugBank, ChEMBL, and ChEBI identifiers, would that be sufficient to filter your list?

[kevinxin90]

Yes. That would be sufficient.

[andrewsu]

@mmayers12 I keep forgetting to ask you about this. I think given the complexity of those reframe annotation files, can I ask you to do this please? Can you generate a list of reframe compounds with mappings to DrugBank, ChEMBL, or ChEBI identifiers? Don't post the file here though -- on slack or one of the su0x servers...