[gmx-users] CHARMM36 and energygroup exclusions

[Daniel Bauer]

Hello,

I need to run a simulation of a membrane system with CHARMM36 and
energygrp exclusions. Therefore, I used the mdp settings suggested by
the gromacs page but a lower rvdw-switch (0.8 instead of 1.0) to better
reproduce DPPC lipid params:

constraints = h-bonds
cutoff-scheme = Group
vdwtype = cutoff
vdw-modifier = force-switch
rlist = 1.2
rvdw = 1.2
rvdw-switch = 0.8
coulombtype = PME
rcoulomb = 1.2
DispCorr = no


Since I need to exclude energy groups, I cannot use the Verlet cufoff
scheme but have to use group scheme instead. However, It seems like this
incompatible with the cutoff and rlist settings since I get a warning:

WARNING 2 [file md0.mdp]:
  The sum of the two largest charge group radii (19.291325) is larger than
  rlist (1.200000)

I think this is because rlist is the same size as rcoulomb and rvdw.
However, setting a larger rlist is also not possible because of the PME
coulombtype (Error: With coulombtype = PME, rcoulomb must be equal to
rlist).

My question is is it safe to ignore this warning if the simulation is
only runs for a very limited time (2ns) or not? And if not, what changes
to my settings would you suggest?

To give you some background: I try to add a VdW sphere in my system that
only interacts with the protein but allows water to pass through it
freely (hence settin energygrp-excl between that particle and everything
but the protein). After 2 ns, the sphere will be removed and the system
can be simulated with the Verlet scheme and original settings again.

Best regards,
Daniel B