search

UCL-CCS / TIES

Topology Superimposition based on joint graph traversal
MIT License
5 stars 1 forks source link

Charge assignments with antechamber #330

Closed AlexW12344 closed 1 year ago

AlexW12344 commented 1 year ago

Hey,

If we pass un parameterized mol2s to TIES it will use antechamber to add the charges. Since the input ligands are in the docked pose will antechamber parameterize these exact coordinates? I think since antechamber was designed to reproduce charges on low energy conformers this might not be perfect?

Is my thinking correct here? Is there a better approach to what I'm doing that would avoid this problem?

Cheers, Alex

bieniekmateusz commented 1 year ago

Hi Alex,

AFAIK yes that's how it works and it is certainly a limitation.

I believe OpenFF takes this into account and generates different conformations before running antechamber. Also openff-nagl now works and I think is an interesting approach, but there is no publication yet so I can't speak about how good it is at the moment.

Best, Mat

AlexW12344 commented 1 year ago

Hey Mat,

Okay good to know. I'm looking at the effect of conformers and I think that partial charges might confound the result. Maybe I'll have a think about adding a way to pass the low energy conformer as input also, the user might have it from pre docking stages. Then you could use the charges from this and the coords from the docked molecule.

Does seem like a limitation of antechamber overall really, I'll give openff-nagl a look.

Thanks :)

best, Alex

bieniekmateusz commented 1 year ago

It might be easier to precompute them and pass the .mol2 with charges, and it will just use that. I am hoping to transition to openff as the backend in the near future, if that changes anything! :-)