[gmx-users] Inconsistent shifts over periodic boundaries error.

[Artem Shekhovtsov]

Hello!
    I performed simulations with the small molecules in water and faced
with the fact that in about 5% of cases, the simulation falls at the last
step with an error:

-------------------------------------------------------
Program:     gmx mdrun, version 2018.4
Source file: src/gromacs/pbcutil/mshift.cpp (line 906)

Fatal error:
There are inconsistent shifts over periodic boundaries in a molecule type
consisting of 47 atoms. The longest distance involved in such interactions
is
1.298 nm which is close to half the box length. This molecule type consists
of
multiple parts, e.g. monomers, that are connected by interactions that are
not
chemical bonds, e.g. restraints. Such systems can not be treated. The only
solution is increasing the box size.

For more information and tips for troubleshooting, please check the GROMACS
website at http://www.gromacs.org/Documentation/Errors
-------------------------------------------------------
   The problem occurs when simulating in a rhombic dodecahedron cell with
the standard distance between solute and the box 1nm. There is no such
error in the cubic cell with distance 1nm.
    Trying to establish the reason for this behavior, I found that it is
associated with the presence of virtual atoms in a small molecule. When I
exchange a virtual atom for an atom linked by a chemical bond to the rest
of the molecule, an error does not appear.

    I want to remove errors while preserving virtual atoms as well as the
size and shape of the cell. Maybe someone already encountered this behavior
and knows what my mistake is. I will be glad of any help.

    The link
https://drive.google.com/file/d/0B5NzD-LVrUalbDdSblhfTkowUDN1aWxER3hlLWhaRFROcjhj/view?usp=sharing
contains files for 1-step molecular dynamics that failed with my system of
interest. Gromacs version - 2018.4.

Thank you,
Artem Shekhovtsov