[gmx-users] problem with NVT equilibration for membrane protein

[Floriane Montanari]

Dear all,
I am working with an homology model of a protein inserted in a POPC membrane.
I first ran minimization on the system, which ended up with I think
"reasonable" values for Fmax and Epot:

Steepest Descents converged to Fmax < 1000 in 1390 steps
Potential Energy  = -1.2822042e+06
Maximum force     =  8.4696790e+02 on atom 7447
Norm of force     =  2.2902491e+01

I am using the Amber99sb forcefield modified to include Berger parameters
for my lipids.

I am now trying to run NVT using the parameter file sent as attachment. I
tried to lower the temperature ref_t / different values of cutoffs
rcoulomb, etc. / a Fourier transform of 0.12 as seen in literature
but all this only ended up in the same error:

Depending on how many threads I launch mdrun with, I get
- either a particles error
Fatal error:
3 particles communicated to PME node 3 are more than 2/3 times the cut-off
out of the domain decomposition cell of their charge group in dimension x.
(more or less particles / more or less nodes depending on the runs)
- or a segmentation fault
In any cases, those errors occur right at the end of step 0.

Is there anything I am doing wrong? Can someone give me advices on how to
fix this?
I can send many other files if needed (my topol file, my minimized system,
my forcefield...).
I am running gromacs 4.5.3.

Thank you in advance for any help,
Best regards,
Floriane M.