AMBER Archive (2004)

Subject: Re: AMBER: average structure using ptraj

From: Thomas E. Cheatham, III (
Date: Thu Jul 01 2004 - 12:40:25 CDT

> I tried to use ptraj in AMBER7 to average the last 200 frame of my MD run.
> The script I used:
> traj << EOF
> trajin AR.md7.crd 300 500
> center :1-250
> image center familiar
> average average.rst restrt
> Then I did one more miminization for 5000 steps. It didn't go through.
> The EEL term became ********.

It is always a good idea to "look" at the structure. Note that rotating
groups (such as methyls) spin rapidly and this can lead to overlapping
atoms upon coordinate averaging.

Also note that the average command does not do any fitting of the
structure prior to averaging, so if the whole molecule is rotating, this
may lead to overlapping atoms.

Finally, averaging water positions or mobile ions will also lead to
overlapping; if you have ions or water in the system, as they move across
the box rapidly, this will lead to a big blob of water/ions in the center
overlapping the solute (and hence high VDW and EEL energies).

So, I would modify your script as follows:

 traj << EOF
 trajin AR.md7.crd 300 500
 center :1-250
 image center familiar
 rms first mass out rms.dat :1-250
 average average.rst restrt
 average average.pdb pdb

...and take a look at the average.pdb.

Likely, if you want the average structure for minimization, you will have
to strip the water/ions and then (optionally) add them back in.

Good luck.

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to