Update groups can not be used for this system because atoms that are (in)directly constrained togeth

User discussions
1

GROMACS version: 2023.3
GROMACS modification: No

I have coarse-grained an atomistic system (CHARMM36 FF) and using MARTINI 2.2 FF. I often encounter the following error after mdrun:
Update groups can not be used for this system because atoms that are (in)directly constrained together are interdispersed with other atoms

Going through the Gromacs mailing list, this error appears when hydrogens bonded to a heavy atom do not appear after the heavy atom in the topology file (link1 and link2).

I have two questions here:

  1. I do not have any hydrogens in the CG model and several beads are attached to multiple other beads (upto 3 beads). Most of the bonds are defined as type 1 bonds while some are defined either as type 10 bonds or as type 1 constraints. The molecule is periodic in nature and periodic-molecules=yes is entered in mdp file. How can I circumvent this error?
  2. My system usually blows up within 100 ns simulation. I perform minimization and then step-wise equilibration using Berendsen and then Parinello-Rahman barostat, increasing timestep from 2 fs to 5, 10 with Berendsen and finally 15 fs with P-R. Then, the production run with P-R barostat and 15 fs time step. I was wondering if the system is blowing up due to the error in point 1 above?

Thanks for your help.

Regards,
Raman

2

You shouldn’t interpret that message as an error - I assume that the simulation continues. That is just to inform you that some performance optimizations might not be possible, see https://manual.gromacs.org/current/release-notes/2019/major/performance.html#implemented-update-groups and https://manual.gromacs.org/current/reference-manual/algorithms/parallelization-domain-decomp.html#coordinate-and-force-communication for more information.

Regarding your simulation crashes, have you tried switching to the c-rescale barostat? That is often more stable than Parinello-Rahman.