The largest distance between excluded atoms is 1.725 nm, which is larger than the cut-off distance

GROMACS version: 2022.1
GROMACS modification: No
When using editconf to make a box from a crystal lattice (4x4x4 unit cells) and using a gro file which is fine, I get a new gro file where there are some bad H-H interactions. Moreover, when I try to run grompp in order to do an energy minimization, I get the error:
The largest distance between excluded atoms is 1.725 nm, which is larger than the cut-off distance. This will lead to missing long-range corrections in the forces and energies.
I have never encountered this error before and I am wondering what has caused it. I am using a cut-off distance of 1.0 which I can’t increase without also increasing the box, but I don’t think that this is the problem. Any help would be appreciated, Thanks, Peter

1 Like

Hello Peter!

That’s a new check that was introduced in GROMACS 2022, see Miscellaneous — GROMACS 2022 documentation.

It is there to prevent you from getting mildly incorrect energies, but I suppose that’s not a concern if you only want to do minimization.

As a workaround now, you can use an older version of GROMACS to run grompp. But, if you want to run proper MD later, make sure the error does not appear with minimized structure.

We will probably add a way to circumvent this warning in the future releases of GROMACS.

1 Like

OK. I’ll try what you suggested. T

Could it be that molecules are broken over PBC in the structure you used for stacking?

First of all, I followed Andrey’s suggestion which solved the problem. I only had to use a previous version of grompp once and after I succeeded to minimize the structure the errors no longer appeared in the equilibration and production phases.
As to Berk’s question, I looked at the gro file generated by editconf upon making the box in PyMol and I didn’t see any broken molecules. Moreover, I was under the impression that broken molecules aren’t a problem for Gromacs and that they just affect the visualization of the system.
Thanks again for your help.

If Andrey’s suggestion solved your issue, then we should probably allow such distances during minimization.

But you can use -maxwarn 1 and then grompp should also work with GROMACS 2022, right?

We could consider changing this to a note when doing energy minimization.

I agree.

No. That’s the problem. It’s an “error” and not a “warning.”

I uploaded a change to the release-2022 to change this to a warning.

1 Like

Thank you.

What I did was I increased the cut-off and minimized the structure. After that I reduced the cut-off and proceeded with MD.