'Triclinic box too skewed' - will this affect the way my system behaves?

User discussions
1

GROMACS version: 2021
GROMACS modification: No

I’m running a simulation with a unit cell with parameters: -box 3.9970 2.6950 67.7900 -angles 89.24 94.59 105.58. This creates an issue with ‘Triclinic box too skewed’ as |zx| > 0.5 |xx| (|-5.423| > 0.5 * 3.997). As I understand it, this shouldn’t be an issue as we could add the a vector to the c vector to then reduce the value of |zx|, and Gromacs seems to do this after using genion.

However after viewing trajectories in vmd, the system does not seem to follow the correct unit cell and gmx solvate does not seem to place the solvent inside the correct box (instead it seems to follow a unit cell in which zx is equal to 0). Is this just an issue with the way gromacs is outputting the trajectories and during the simulation it will follow the correct unit cell? Or is the triclinic box being too skewed mean the simulation is using the wrong unit cell?

Thanks for any help.

2

Which program gives the error?

mdrun will adjust the box as you say and there is usually no issue.

In what unit cell the atoms are in the output, e.g. a rectangular or a skewed one, is irrelevant for all calculation purposes.

3

I’m looking at the output from gmx solvate and the trajectory produced after mdrun. So is this just due to the way gromacs stores the box and its calculating it correctly during the run?

4

mdrun can store the atoms in many different ways, depending on what is most convenient for the particular calculation setup. The output you got is not incorrect (I presume), you can stack periodic images without clashes or gaps.

https://manual.gromacs.org/current/reference-manual/algorithms/periodic-boundary-conditions.html