From: Axel Kohlmeyer (
Date: Sat Jun 06 2015 - 13:28:17 CDT

On Sat, Jun 6, 2015 at 5:31 AM, Eric Smoll <> wrote:
> Hello VMD users,
> I am working with reactive QM/MM trajectories computed in GROMACS (.gro
> structure file + .xtc trajectory file). VMD appears to compute the displayed
> connectivity from the structure file and this connectivity is not dynamic
> unless the "dynamics bonds" drawing method is used.

correct. for performance and storage efficiency reasons VMD assumes
trajectories are from a classical MD simulation without changes.

> I am interested in computing kinematic properties of dynamically defined
> molecules. Is there a way to do this in VMD? Is there any other tool that
> might be capable of this?

that depends on what kind of heuristics you want to use to determine
the bonding topology. the heuristics in VMD are fairly simple. it is
based on the atomic radii that VMD guesses from the names of the atoms
(unless you provide diameter information to overwrite it) with the
additional restriction to not have bonds between two atoms that are
detected as hydrogens. that gets a decent result for most people that
use VMD. if that is sufficient for you, you can trigger a recomputing
of the bond topology from the script interface through the "mol
bondsrecalc" command.

> I find that the "dynamics bonds" drawing method works well in most
> situations but many of my trajectories evolve hydrogen gas from organics and
> the H2 bond is never rendered over the H2 vibrational period. Is there a
> reason this might be?

yes. the simple distance based heuristic would determine a lot of
unphysical H-H bonds where both H atoms would be connected to another
atom. H2 molecule bonds could only be reliably detected by adding
another pass to the bond detection scheme.


> Best,
> Eric

Dr. Axel Kohlmeyer
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.