RE: AMBER: Post-MD minimization

From: Ross Walker <>
Date: Fri, 23 Jun 2006 05:19:14 -0700

> I did carry out the entire minimization with the above parameters
> but the .rst file never got created! I am unable to
> understand why. I got
> only the *.out file.
> But, it is mentioned in the manual that setting IMIN = 5,
> Sander expects a
> trajectory (.crd) file as an input and also that -y flag can
> be used to
> specify this inpcrd file (page 87 as well as page 90 of the manual).
> However, if the -y flag is used, it results in an error that
> the flag is
> not a valid one. (This has been discussed in some of the earlier
> discussions on the AMBER reflector as well. At that time it
> was suggested
> that -x should be used for specifying the imput trajectory
> coordinates.)

HHHmmm, I did not realise you were using imin=5. I have never ever tried
this myself so don't know for sure how it works. According to the manual it
takes a trajectory file as input and performs the options in mdin on all of
the structures and writes the minimised structures as frames of a standard
mdcrd file. Hence it seems you should not get an rst file from this type of
minimisation but instead will get an mdcrd file as specified by the -x
option that contains the minimised frames. I'm not sure about the -y option
this maybe a typo in the manual. Perhaps others can comment. The -x option
it seems woul dbe for output only here as this is where the minimised
coordinates will be written to.

Note, this probably isn't the most efficient way to achieve what you want to
do and is not the way I would do it. I would take the trajectory you have
and use ptraj to split this into individual restart files. I would then
script a run with sander and imin=1 to minimise each of the restart files in
turn and create a new minimised restart file. This gives you much more
flexibility as you can do the minimisation in any order, you can skip say
every other structure and you can also do the minimisation much more
efficiently. I.e. you could create 16 runs that stride through the list of
restart files in steps of 16. Then you could run each one as a single cpu
job on 16 individual machines in a cluster. This will be much more efficient
that running a single 16 processor job.

> Moreover, there should be a provision with these settings to
> save **all**
> the minimized coordinates as this option may be used like a batch
> minimization of several captures structures. Right now only
> -r option is
> available for the output coordinates. Is there any such
> provision which I have missed?

According to the manual the **all** you are refering to above should end up
in the output trajectory file as specified by -x. As I say though I have not
tested this option so can't say for sure. I still think you would be better
off doing it manually via ptraj and then running 10,000 independent

Good luck,

|\oss Walker

| HPC Consultant and Staff Scientist |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- |
| | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read 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 Sun Jun 25 2006 - 06:07:18 PDT
Custom Search