PBC centering Issue

[Anamika]

GROMACS version: 2022.3
GROMACS modification: Yes/No
Deal all,

md.mdp (2.6 KB)

I am Anamika Kumari, doing PhD in IIPE, Visakhapatnam. I am working with a dimeric and tetrameric protein system containing one and two ligands, respectively, and I have performed a 100 ns molecular dynamics simulation. After completion of the simulation, I observed that the protein-ligand complex was drifting outside the simulation box during visualization, and therefore attempted to remove periodic boundary condition (PBC) artifacts.

To address this issue, I applied multiple PBC correction options available in GROMACS, including molecule-based reconstruction, jump removal, and centering. Specifically, I tested combinations of -pbc mol, -pbc nojump, -pbc whole, and -center to ensure the integrity and proper positioning of the complex. I also applied PBC corrections using molecule-based and jump-removal options, followed by centering. Specifically, I used the following command:

gmx trjconv -s md_0_100.tpr -f md_0_100.xtc -o md_noPBC.xtc -pbc nojump -center

After applying this correction, the overall protein–ligand complex appeared intact and centered. However, upon visualization, I observed that certain residues appeared to be missing or discontinuous, particularly within the oligomeric protein chains. I have attached md.mdp file and a part of protein where residues disappeared.

I would appreciate guidance on the correct PBC correction workflow for multimeric protein–ligand systems, and advice on how to prevent residue discontinuities while maintaining structural integrity during post-processing and visualization.

[Karis]

If the protein is properly centered without any contact with the boundaries, this is likely due to an issue with the visualization software rather than a PBC issue, the residues are still there but at such an orientation that VMD isn’t visualizing it properly. You could double check by visualizing those residues with lines or CPK rather than NewCartoon.

[neilthomson]

Are you using pymol to visualise a .gro or .pdb output here? Pymol guesses bonding, so it will usually make proteins look fragmented like this. Try viewing in vmd with a .tpr that contains bond info, or load your .tpr into MDAnalysis to create a pdb file in python.

PBC corrections vary depending on system sadly. But, I usually give mine a first attempt with this template - a four step trjconv protocol

GROMACS trjconv post-processing pipeline (4 steps):

Make molecules whole (-pbc whole)

Fixes molecules broken across periodic boundaries

Center protein in box (-center -pbc mol -ur compact)

Centers the protein and uses compact unit cell representation

Cluster & remove jumps (-pbc cluster)

Clusters molecules around the centered protein, removing PBC jumps

Fit to reference (-fit rot+trans)

Removes rotational and translational motion by fitting to the reference structure

This works for some systems, but inspect the trajectory afterwards, because sometime the solvent doesn’t get corrected fully

[Anamika]

Thank you so much for your suggestions.

Anamika Kumari

UGC-JRF
23RS9P10
Molecular Modelling & Protein Engineering Lab
Indian Institute of Petroleum & Energy (IIPE),Visakhapatnam-530003