System getting out of box upon simulation

User discussions
,
1

GROMACS version: 2024.1
GROMACS modification: Yes/No
Here post your question:

I want to study the aggregation of a small molecule through classical MD simulation. For which I initially generated the topology of a single molecule using CGenFF and used them further in GOMACS. In GROMACS, the first thing I did was to create a box around the single molecule (editconf) and then included some more molecules into the system using gmx_insert-molecule. It was followed by solvation, energy minimization, equilibration and final production run.

However, once I checked the structure files after 100 ns of simulation, I found out that the small molecules and solvents have gone out of the PBC box. Even some of my small molecules have lost their intra-atomic interactions, and turned into individual atoms. I am assuming that after some more time-scale there would be no trace of them.

Can someone tell me what could have gone wrong. All I can think of is that the PBC parameters have been somehow misguided, but again, the .mdp files used are the same as I use for protein-ligand systems. I only made changes to tc-grps as per my system.

I would really appreciate any lead. Thank you in advance.

2

A bit confused on what you mean by ‘gone out of the PBC box’, PBC means atoms passing through one side should just reappear on the other side, is the issue that the molecules are breaking apart when they shouldn’t be? Could you share a visualization of the system to show what the issue is?

3

Hello! Thank you for replying and sorry for the delayed response.

So when I extracted the structure, the molecules seemed to have moved away from each other a lot and upon super-imposing with the initial structure, it was obvious that they were out of the box and as you said, the atoms were breaking apart. Upon simulating further, there was no trace of the molecules at all. So when I extracted a structure of the entire system, I could see the dispersed water molecule only.

I got in touch with someone who had previously worked with a similar setup and they recommended me to use -pbc mol instead of -pbc nojump which surprisingly solved the issue.