From: Ashar Malik (
Date: Wed Nov 26 2014 - 17:31:30 CST


When you save snapshots in a trajectory they are at intervals of say 250 to
500 ps, where as the integration time step that we use is 1 to 2fs. What
that means is that if you are just looking at the trajectory (after the
simulation has finished) there is no way of recreating the intermediate
information (the coordinate information).

What one can do (in theory) is run a simulation and analyze it to find
possible region if interest. Lets say that in a 100ns run, your region of
interest is between 38.0ns and 38.5np. So what you can do is use a restart
point immediately before 38ns only now your script saves snapshots every
integration time step. This way you will be able to capture all the
information and once that information is logged in your (new trajectory
which you ran for a time frame of 500ps) trajectory, you have the
coordinates - what is left is just standard physics - you know the mass of
water - you have the water moving a distance (1 integration time step, from
2 consecutive time points), you should be able to calculate the forces.
However saving a snapshot of a system over 500ps @ 2fs intervals = 250,000
states in the new trajectory. You could reduce this to a lesser time span,
since you probably need a couple of frames - for before - an impact frame -
and a couple of them after, for force calculations (in theory even three
should be enough - if you calculate for a couple of waters and its not too
different, you can assume it to be a rough average).

However a trajectory file - is just coordinate information, forces are
calculated as internal steps of the MD code, you could also look at that -


On Thu, Nov 27, 2014 at 11:51 AM, John Stone <> wrote:

> Hi,
> This kind of question falls into the domain of the particular
> MD simulation tool you're using, and asking that program to save
> the velocities or forces in a trajectory file. VMD doesn't have
> the ability to compute these values, so if they aren't available
> in a trajectory, then there's really no way to compute them
> after-the-fact, at least in VMD. You might be able to feed a
> simulation trajectory into an MD engine as input and have it
> compute the quantities of interest, but again, this is totally
> outside the scope of VMD currently.
> Cheers,
> John Stone
> On Mon, Nov 24, 2014 at 09:29:50AM +0800, sunyeping wrote:
> > Dear all,
> > Do you have an idea how to monitor the velocity of the water molecules
> > that interact with a particular residues of the protein of the force
> that
> > these water impose on that residue in the protein in a MD simulation
> > trajectory?
> > Thanks in advance.
> >
> > Yeping Sun
> >
> > Institute of Microbiology, Chinese Academy of Sciences
> --
> NIH Center for Macromolecular Modeling and Bioinformatics
> Beckman Institute for Advanced Science and Technology
> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
> Phone: 217-244-3349