[gmx-users] mdrun blowing up with PME node error in 1 window of umbrella sampling

[Agnivo Gosai]

Dear Users

I am using umbrella sampling to calculate free energy of binding between
two biomolecules , in certain different environments.
This is my standard work flow for all the cases.

1. Topology preparation, solvation etc.
2. Energy Minimization using steepest descent and 50,000 steps.
3. 1 ns position restrained NVT equilibration.
4. 1 ns position restrained NPT equilibration.
5. 500 ps SMD run of the equilibrated structure keeping one molecule fixed
and pulling the other.
6. Extracting around 40 - 50 windows from the SMD trajectory.
7. 100 ps NPT equilibration and 10 ns umbrella sampling for each window.
8. Generate PMF.

For 3 cases I did the above and had no issues. However , recently for "only
one" window each in 2 cases my umbrella sampling runs crashed with the
following message :

2 particles communicated to PME node 10 are more than 2/3 times the cut off
of the domain decomposition cell of their charge groups in dimension x.

So , I read previous forum mails and the manual and I guess I can do the
following :-

1. Another energy minimization and a longer NPT equilibration of the
particular window before doing the umbrella sampling run. But, is this
advisable ?

Kindly suggest.

Thanks & Regards
Agnivo Gosai
Grad Student, Iowa State University.