Dual topology file for Alchemical transformation of water to ions

GROMACS version: 2019
GROMACS modification: No

I am using the PMX tool to generate a hybrid (dual) topology file for proteins. When I do mutation of uncharged amino acid to the charged amino acid that time net charge on the system becomes non-zero. To keep the net charge of the system constrained to zero, I want to mutate water molecules to Na or Cl ions with protein amino acid mutation.
I want to know, how can i make a dual topology file for water ion transformation and use it in Gromacs.
Can you please help me with this?

Thanks and regards
Brijesh Saraswat

With finite sampling, your proposed approach may be worse than just accepting the neutralizing plasma solution. The concerted change of a water to a charged ion is likely to introduce large statistical sampling errors as the water/ion would need to obtain Boltzmann sampling of all locations in your simulation system unless you use a clever approach to keep that water localized to bulk away from your solute (much like the sampling errors associated with alchemical decoupling in the absence of position/orientation restraints). And in many applications I can imagine you’re anyway going to have to compute the water/ion conversion free energy separately in order to remove its contribution. As for how to do it, many years ago you just needed to ensure you had the appropriate A and B states for any molecules you wanted to change, though that may not be the case anymore. You could certainly test with some 0-step energy evaluations under different Hamiltonians.

Have you checked out the solution that Aldeghi, De Groot and Gapsys came up with for charge-changing mutations?

we suggest using a double-system/single- box setup [24, 26]. In this approach both legs of a thermodynamic cycle, e.g., mutation in the folded and unfolded states in Fig. 1, are placed in the same simulation box. […] In this way, the charge of the system will be conserved during the transformation, and the free energy difference calculated already refers to the ΔΔG across the thermodynamic cycle of interest.

It is very elegantly explained in Aldeghi et al, Methods Mol Biol (2018) DOI:10.1007/978-1-4939-8736-8_2

I hope it works well!

1 Like