Sudden spike in ligand RMSD during protein-ligand simulation

[ShintY]

GROMACS version:2023.5
GROMACS modification: No

I recently performed a protein-ligand simulation for 25ns and during the post simulation analysis, I saw a sudden spike in ligand rmsd value which lasted only for about 20-30ps.

Above is the image from ligand-rmsd.xvg file and just below it is a graphical representation of the data. I used gromos54a7ff and used https://atb.uq.edu.au/index.py to build ligand topology. I’m fairly new into gromacs simulations and from my short experience, never encountered this. I even observed the md-simulation movie on the exact time frame but didn’t observe any abnormal ligand movements.

Any help is appreciated,
Thank you

[Karis]

Could you share the pictures of the system for the time frames directly before, at and after the spike?

Also, did you remove the pbc effects and include the -nopbc flag in gmx rms?

[ShintY]

Firstly I didn’t remove the PBC while calculating the rmsd. I thought since the protein never touched the pbc boundaried during the entire simulation period, it was unnecessary (please correct me if im wrong).

Below are the pictures of the system, they are in order from 21570-21600ps.

The ligand is in RED colour
Thank you for your response

[Karis]

Is there any improvement once you do remove pbc?

[ShintY]

Yes there is an improvement

But i dont understand why the values spike up that high when neither the ligand nor the protein contact the PBC. Anyways thank you very much sir