Position Restraint and Center of Mass Removal

Hi everyone

I am trying to run a NVT simulation using positions restraints using Gromacs 2020.5. I am holding a graphene sheet fixed in the x and z-axis. In my simulations, I am using Nose-Hoover thermostat.

However, I am having several warnings, and looking to the GROMACS forum I found several messages but I did not get what I really should do.

So, if I use position restraints I get the warning:
"Removing center of mass motion in the presence of position restraints
might cause artifacts. When you are using position restraints to
equilibrate a macro-molecule, the artifacts are usually negligible."

Them I am using comm-mode = None. From this “old” discussion Re: [gmx-users] com motion and position restraints may cause artifacts seems that NVT without center of mass remotion is OK.

However, with comm-mode = none, I get this warning:

"You are not using center of mass motion removal (mdp option comm-mode), numerical rounding errors can lead to build up of kinetic energy of the center of mass"

So what I should do? Use or not use the center of mass removal?

It looks like GROMACS does not have proper support for the case where coordinates are fixed in some but not all dimensions by position restraints or freezing.

Can’t you also restrain along y?