[gmx-users] couple-moltype in FEP

[Hannes Loeffler]

On Tue, 12 Jul 2016 10:16:24 +0200
Alexander Alexander <alexanderwien2k at gmail.com> wrote:

> Hi,
> Thanks for your response.
> I want to calculate the binding free energy of a single amino acid to
> a solid surface in aqueous solution by FEP, alchemical
> transformation, where perturbations applying on different species in
> ordered steps. For instance: Step.1 applying harmonic restraints  to
> the O and H of the amino acid. Step 2. Charges removal from the amino
> acid and two water molecules and an Ion, happen.

I think the key figure here is actually Fig. 1.  What that paper
attempts to do is to replace a molecule on a surface with two water
molecules through a decoupling scheme.  I guess with Gromacs you would
have to manipulate the topology files to some extent because the
decharging and recharging steps involve different species (the sodium
ion is there only to offset any changes in total charge).
couple-lamnda0/1 may be helpful for some steps.  There are also both
insertions and deletions so I think step 3 must be split into two vdW
(deletion) steps.  Probably best to work this out by going through
steps 1 to 6 an pay careful attention to what species have zeroed out
charges.

It's unclear to me why you need to append any data.  What coordinates
you use doesn't really matter because each run is independent from
every other run (unless you see a specific need to take over a
coordinate from an adjacent lambda).