Instantaneous positional shifts of protein trajectory

GROMACS version:
GROMACS modification: Yes/No
Here post your question

I performed a 440ns molecular dynamics simulation of a protein homotrimer using GROMACS. After conducting RMSF analysis, I found that the RMSF values of the three monomers fell within different ranges. When I viewed the protein trajectory with VMD, I observed that the entire trimer underwent instantaneous positional shifts in certain frames. Subsequently, I extracted the trajectory file for the first 100ns, and the RMSF values returned to normal. So, I would like to ask what might have caused this movement, whether it led to the abnormal RMSF values, and if there is indeed an issue, how can it be resolved?

This sounds like a PBC issue, have you corrected for that in the trajectory?