RE: AMBER: unstable MD velocities

From: Ross Walker <>
Date: Thu, 1 Sep 2005 11:06:08 -0700

Dear Hayden,

> I have created a lipid chain bound to 2 carbohydrate moieties
> using gaff
> parameters. The partial charges were generated using a gaussian/resp
> procedure.
> The problem is this:
> Dynamics runs of the molecule (in gas phase and igb=1)
> rapidly give rise to
> huge velocities that quickly cause the system to crash. The
> crash can only
> be avoided by resetting the velocities after every dynamics cycle.

If you created the structure yourself then you will need to be much more
careful with setting things up. Start off by doing minimisation on the
system, do around 2000 steps or so and see what the final structure looks
like. If this looks okay and the component energies are not huge (especially
bonds) then you can try dynamics on it. In this case I would start at zero
kelvin and then heat very slowly to 50 K or so. Use weight restraints to
increase the temperature over 50-100ps. During this time you should also use
a very short timestep (0.5fs or shorter). I would also look to record
snapshots to the mdcrd file very frequently, every 5 steps or so if you have
the disk space. This will allow you to watch the system undergo initial MD
so you can locate where the instabilities are and fix them. Once you have
the system stabole at 50K you can then slowly heat it up to your desired
temperature. Once it has begun to equilibrate at the target temperature you
can then up the timestep to 1fs (no shake) or 2fs (shake).

This sort of procedure (careful heating and equilibration) is often required
with 'model' structures. Note, you may also need to restrain parts of the
molecule to allow the more strained parts to relax without destroying the
entire structure.

All the best

|\oss Walker

| Department of Molecular Biology TPC15 |
| The Scripps Research Institute |
| Tel: +1 858 784 8889 | EMail:- |
| | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be ready every day, and should not be used for urgent or sensitive issues.

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Thu Sep 01 2005 - 19:53:00 PDT
Custom Search