[gmx-users] Intermolecular Nonbonded Interactions

[Batuhan Kav]

Dear Al

lI have a question regarding the non-bonded LJ interactions between two molecules. In particular about how Gromacs generates the non bonded LJ interactions between the water and protein molecules.

In my system I am using a custom forcefield for the water molecules and Charmm36 for the protein. My intention is to change the epsilon/sigma LJ parameters between the water atoms and the protein atoms.

For the sake of a simple test I use spc water model (spc.itp) to solvate the protein. As far as I can see in the ffnonbonded.itp there is the correct atom type for the spc water (in the [ atomtypes ] directive) defined however it lacks the protein-water atom pairs in the [ pairtypes ] section. This makes me think that when I run grompp it will generate the non bonded LJ interaction values between the protein and water atoms following the combination rule in the forcefield.itp file provided gen-pairs is set to “yes”. When I set gen-pairs to “no” I expected that grompp will complain that the LJ pairs for the water atom and protein atoms are missing. However grompp runs without any warnings/errors and the simulation can be started.

Therefore my question is: how does Gromacs generate nonbonded interaction between the molecules if no specific atom-atom pairs are provided in the [pairtypes] and gen-pairs is “no”  (these atom-atom pairs are not defined anywhere in any part of the forcefield at all)?

I’m sorry if this is a trivial question but I am rather confused about this.

Best,
Batuhan