[gmx-users] Question about switching off non-bonded interactions between two molecules for MD simulation

[Soomin Choi]

Dear GROMACS users,

I am new to GROMACS software and have a question about switching off
pair-wise non-bonded interactions between two molecules.
I have a simulation box composed of a protein with a substrate inside the
protein catalytic pocket and there are no covalent bonds between protein
amino acids and a substrate. I would like to run MD simulation of this
system after switching off non-bonded interactions between one of the amino
acids and a substrate (leaving all intra non-bonded interactions & inter
non-bonded interactions between given amino acids and other amino acids). I
would appreciate it a lot if you could guide me on how to do this. Below
are approaches I tried but got warning/error messages:

Approach 1) use [exlucsions] in .top file
If, let's say, atom # 1,2,3 are for the amino acid and atom# 4,5 is for a
substrate, I added:
[ exclusions]
1  4 5
2  4 5
3  4 5
I have successfully obtained .tpr file from grompp but got a fatal error
after mdrun:
'There is no domain decomposition for 448 ranks that is compatible with the
given box and a minimum cell size of 25.3862 nm. Change the number of ranks
or mdrun option -rdd or -dds'

Approach 2) use energygrps and energygrp-excl in .mdp file
If the name in the index for the amino acid and substrate is A and B
respectively, I added:
energygrps          = A B
energygrp-excl      =  A B
Also, I knew that energygrp-excl does not support Verlet cut-off scheme, I
used a 'group'. However, I have got an error message when I used grompp:
'The sum of the two largest charge group radii (33.381752) is larger than
rlist (1.200000)'

Just to be clear below is a summary of my system:
- a system composed of one protein and a small substrate
- the substrate and amino acid that I would like to switch-off are charged
(+1 and -1, respectively)
- protein-substrate complex are solvated in TIP3P waters, total # of atoms
in the system = ~450,000

First of all, I am not sure whether these approaches are correct to
switch-off non-bonded interactions between two molecules. Second, if so, I
would like to know how to correct these issues.

Sincerely,

Soomin