Standalone App For Geometry Optimization

Question

[Aloctavodia]

I am looking for a standalone application to perform geometry optimization on proteins and other macromolecules. It should support parameters for the AMBER force field (and probably others) and the possibility to add/modify those parameters. I will be using this app as part of a Python script to perform a lot of geometry optimizations. Tinker (http://dasher.wustl.edu/ffe/) could be a solution but have some drawbacks, 1) I can´t re-distribute the code 2) I would prefer to use pdb files directly (not to convert them to xyz files) 3) i would prefer to use amber parameters in this format http://ambermd.org/#ff

Thanks in advance.

Answer 1

[João Rodrigues]

So if I understood correctly, you want something that does what AMBER does, with support for AMBER forcefields, using AMBER formats, but it's not AMBER :) Why not GROMACS? It's free, it can be modified, it supports AMBER, the format is not the same but not so hard to learn neither, and it can be scripted with this (look in the website for other alternatives).

[Aloctavodia]

Nop, I just want a geometry optimization routine that support the AMBER force field. I don´t want something that does what AMBER (the suite) does. AMBER, GROMACS NAMD etc are packages to perform molecular dynamics, MD involves a lot of different routines and geometry optimization is just a tiny subset of those routines. Probably I could take the GROMACS code and try to "isolate" the energy minimization routine, but I think it would be easier to start from some application smaller than GROMACS. I have been thinking on using Gabedit.

[João Rodrigues]

What about MMTK? It's in Python, free and open-source, and has the AMBER forcefield in it. I would still go with GROMACS though, it's small, fast, and easy to install.

http://dirac.cnrs-orleans.fr/MMTK/

[Aloctavodia]

I want to (eventually) re-distribute the code. It make not sense to make my code depends on whole gromacs if I only going to use the minimization routine. MMTK is an interesting workaround for my problem MMTK is very modular so it would be relatively easy to isolate the minimization routine.