[gmx-users] Mutation-induced binding affinity change using BAR

[Oliver Kuhn]

Hi Gromacs People,

we want to do a BAR calculation to calculate the binding affinity change
induced by a mutation.

That means:
- We want to transform a residue from aminoacid1 (aa1) to aminoacid2 (aa2)
in a single-transformation approach (transforming both vdw and coulomb at
the same time)
lambda=0 :  aa1 switched on and aa2 switched off
lambda=1 :  aa1 switched off and aa2 switched on
- to calculate binding free energy, we would calculate this transformation
process once in the complex and once in the receptor and calculate the
difference

However, I can find no information how to have both aminoacids switched on
half, for example for lambda=0.5
Any tutorials I found decouple atoms from its environment but not with
switching on different atoms at the same time.
In Amber, this is possible by using two topologies.
Is this possible with Gromacs and where can I read how to do it?

A simpler example illustration the problem would be:
Transforming Toluene (Benzol-CH3) to Phenole (Benzol-OH) in water (to
calculate the change in solvation energy)
from lambda 0 to 1, -CH3 interactions should be gradually switched off and
-OH switched on (with both switched on half at lambda=0.5)

For further disambiguition:
We want to perform equilibrium simulations at discrete lambda steps (with
delta_lambda=0 for each simulation) - no slow growth

Any help is greatly appreciated.
Greets
Oliver