Partial charge generation outside the protocol (user charges)

[hannahbaumann]

[IAlibay]

With the next release of gufe, ingesting SDF files with partial charges will be simple.

The main question I would have is; "do we want to get into the system preparation game". I.e. should it be our problem to offer a CLI tool that adds partial charges to a set of input ligands OR should that be left to the user.

I'm happy to take a neutral stance here, we do have tooling to assign partial charges here https://github.com/OpenFreeEnergy/openfe/blob/main/openfe/protocols/openmm_utils/charge_generation.py#L281 so the CLI tool is probably just a multiprocessing call on top of that method. However, the team needs to make sure they are happy with whatever maintenance burdens that this might lead to.

[jthorton]

My initial thought was that it was provided as part of the industry benchmarking script so industry partners might expect it, but this should be less of an issue when we have things like nagl.

[IAlibay]

less of an issue when we have things like nagl

I think we'll need to plan for folks using ambertools for a while. There's a longer discussion to be had though.

[jthorton]

I think it would make sense to expose this in two places 1) a standalone CLI which charges the given ligands provided in the following formats SDF, LigandNetwork, and AlchemicalNetwork 2) as an extension to the network planning CLI which will charge the ligands while generating the network.