Ligand gets parameterised twice

[fjclark]

If the user supplies unparameterised input, the ligand gets parameterised during both the bound leg and the free leg preparation. Note that the conformations used are identical, so this shouldn't be an issue in terms of substantially varying partial charges (https://onlinelibrary.wiley.com/doi/full/10.1002/jcc.70112), but it is a waste of time and it would be preferable to have exactly the same charges.

This is a sloppy error on my part, which I didn't notice as I have mostly performed parameterisation outside of a3fe in the past. Sorry.

This should be easy to fix - just have the free and bound stages both rely on a ligand_param.prm7/ rst7 instead of writing to free_param and bound_param.

I will try and get round to this soon (I'll probably just fix in the feature decouple somd 3 branch as this should be merged soon) but feel free to have a shot @Roy-Haolin-Du if you have time (no worries at all if not - I'm sure you're very busy too).

Thanks.

[Roy-Haolin-Du]

Hi Finlay,

Thanks for your valuable suggestion!
You're absolutely right, the ligand does get parameterised twice, as both slurm_parameterise_bound() and slurm_parameterise_free() call the parameterisation separately.

This should be easy to fix by adding a check like:
“Check if the ligand is already parameterised; if not, parameterise it, and then saving the parameterised ligand for reuse."

I can take care of this bug fix as a follow-up, probably right after the decouple_somd_3 merge.

Thanks again!

Best,

[fjclark]

Amazing, thanks Roy!