Potential trjconv issue? - dichloromethane molecule with broken carbon-chlorine bonds

GROMACS version: 2016.3-foss-2018a
GROMACS modification: No

Dear All,

I am currently working on establishing a solvent box of dichloromethane (DCM) for use with some of my ligand systems. For that I processed a single dichloromethane molecule using LigParGen (LigParGen Server) to obtain my .gro and .itp files, which I then use to form make up a solvent box of my desired density by adding several of these molecules to a box and running EM, NVT, NPT and a short MD run. I did not run into any errors with this process, but the only thing, which throws me off is that when visualized with ChimeraX, the chlorine atoms of my DCM molecules are not attached to the central carbon anymore. This is already the case for the single molecule .gro file directly after processing with LigParGen (screenshot1) and carries through to my solvent boxes. I am aware that in MD simulations no bonds can be broken, so when I first saw it, I thought I would be able to iron it out using the trjconv command after making the box, but I did not have any success with this yet.
Interestingly, the single molecule file looks perfectly fine when visualized via VMD instead of ChimeraX, but also here a small amount of DCM molecules “loses” their chlorine atoms after the NVT step. The problem seems way more pronounced in ChimeraX though.
I’ve since run some simulations with these boxes and despite my “problem” with the chlorine atoms, everything runs smoothly. I am sure I am missing something regarding the visualization here, so I would be grateful for any advice on this.

Thank you in advance!

Kind Regards,


gro files contain no information on bonds, mostly just the atom coordinates. The topology resides in the topology file.

When a visualization software show you bonds after loading a gro file, it is the result of a heuristic inside that program, that mostly considers a bond to exist when two atoms are close enough. This heuristic distance may vary between atom types, but more importantly, also between software.

If the topology correctly has a bond between these atoms, the simulation is okay and only the visualization is problematic.

Thank you for your reply!
Indeed, the issue could be fixed when I visualized my system with Chimera (Not “X”!) from the .tpr and .xtc file. Thank you for clarifying the issue with visualization from the .gro file.