Simulation breaks the side chain of some residue gromacs

User discussions
,
1

GROMACS version:
GROMACS modification: Yes/No

I have performed 3 replicates of 10 ns protein simulation using the same npt.gro files. replicate 1 seems to be fine but in the other 2 replicates when I visualize the .gro and .xtc file produced after simulation in VMD I can see that for a small number of frames side chains of some residues are detached from the backbone. and in some frames, the Beta sheet becomes so thin like a loop. I cannot understand why it is happening and how to resolve this issue.
i have used gromos54a8 forcefield

2

The visualization don’t necessarily represent the topology since it only uses distance-based heuristics to place bonds when atoms are close enough (Bond Dissappearing in VMD - #2 by rpsingh), so weird artefacts like this often appear. As long as the topology file is correct, your simulation should be correct (Gromacs can’t add or remove bonds during the simulation)

3

thank you for the response.
I hjave to make a movie of the trajectory but it seems too fuzzy when one seees such drastic changes in the trajectory. is there a ways that I can see the correct visualization.

4

VMD should provide a method for adjusting the cutoff distance for bonds which you can adjust, as is suggested here ([gmx-users] Bonds missing in VMD visualization) which should solve issues with bonds disappearing in the visualization

5

It must be due to periodic boundary condition issues. You can fix this with the GROMACS suggested workflow: https://manual.gromacs.org/current/user-guide/terminology.html#suggested-workflow

If you are using VMD, check the trajectory smoothing window size to be zero (Graphics > Representations > Trajectory tab) as well.