Hi Jason,
Thanks for your tips one month ago. A few weeks ago, I was able to hack
two source files in AMBER to enable the functionality I wanted (looping
over a trajectory and computing the total force at each conformation).
The result is that we can now optimize AMBER force fields using
energies, forces and ESP values taken from quantum calculations as
reference data. :)
The changes to the source are small, and hopefully they'll be useful for
somebody else at some point.
https://dl.dropbox.com/u/5381783/Research/debug.F90
https://dl.dropbox.com/u/5381783/Research/trajene.F90
I also made patch files, but I couldn't figure out how to make them
compliant with the rest of the AMBER patches. Is there a webpage that
explains how to do this?
https://dl.dropbox.com/u/5381783/Research/debug.patch
https://dl.dropbox.com/u/5381783/Research/trajene.patch
I also created an example calculation with data and input files, which
allows you to compute the forces over a set of conformations for a molecule.
https://dl.dropbox.com/u/5381783/Research/amberforce.tar.bz2
After extracting the package, run:
tleap -f stage.leap
sander -i force.mdin -o force.mdout -p prmtop -c inpcrd -y all.mdcrd -O
It should generate a file called "forcedump.dat" with energies and
forces from all 1,000 conformations. :)
Thanks,
- Lee-Ping
On 05/29/2012 11:41 AM, Jason Swails wrote:
> On Tue, May 29, 2012 at 11:18 AM, Lee-Ping Wang<leeping.stanford.edu>wrote:
>
>> Hi Jason,
>>
>> Thanks. The variable "open_mode" sounds like a good way to go, and I'll
>> try it out.
>>
>> I'm writing an interface between AMBER and my force field parameterization
>> program (https://simtk.org/home/forcebalance). I'm pretty much the only
>> person using my program at the moment. If someone would like to use my
>> program to parameterize force fields in AMBER, they would have to compile
>> AMBER with the modified debug_frc.F90 and trajene.F90 files as I did.
>
>> Because of this, I'm trying to modify the code in a way that doesn't break
>> other stuff for the possibility that others might use it in the future.
>>
> Then my suggestion is to release your changes as patches rather than your
> version of the files, since it's more extensible that way. This is how
> plumed releases their Amber plugins. I also suggest using the "amopen"
> wrapper for the "open" intrinsic.
>
> HTH,
> Jason
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jul 04 2012 - 13:00:03 PDT
