LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] GAFF Preparation for LAMMPS
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lammps-users] GAFF Preparation for LAMMPS


From: Andrew Jewett <jewett@...1937...>
Date: Sun, 5 Nov 2017 00:02:26 -0700

On Sat, Nov 4, 2017 at 12:13 AM, Axel Kohlmeyer <akohlmey@...24...> wrote:
> On Fri, Oct 27, 2017 at 8:49 PM, MD Simulation <mdsimulationgroup@...24...>
> wrote:
>> Does anyone have much experience preparing LAMMPS input using the General
>> Amber Force Field?  Does anyone have a few commands, to get ne started,
>> using Antechamber that would assign atomtypes and force field parameters for
>> an XYZ file (Axel?)?
>
>
> the first part of your question is a question to the Amber community and for
> the documentation of the Amber/GAFF force field.
> as for converting from Amber format to LAMMPS, please have a look at ParmEd
> or rather InterMol ( https://github.com/shirtsgroup/InterMol ,  which uses
> ParmEd)
> axel.

Hey, this looks like a really useful tool.  Thanks Axel!

Steve, can we include a link to the InterMol at
http://lammps.sandia.gov/prepost.html ?
(https://github.com/shirtsgroup/InterMol)

This looks like a much better choice than either of the suggestions I
made (amber2lmp, moltemplate).

There's no way to compete with the original molecule building software
(such as AmberTools) which was designed specifically for the
particular force field you want to use (eg. AMBER).  And newer force
fields increasingly rely on quantum chemistry calculations to generate
the atomic partial charges.  Neither moltemplate or EMC (to my
knowledge) can do that.  And unlike amber2lmp, both AmberTools and
InterMol appear to be in active development.  (And I have a high
opinion of Michael Shirts, and I think software coming from his lab
will be solid.)  If you use a more generic tool like moltemplate, you
currently run a risk of choosing the precise atom types (eg, SP2 vs.
SP3 carbon, "c1", "c2", "c3", "ca", "cp"...) incorrectly because
moltemplate does not yet infer these details automatically from PDB
files.  (And I have occasionally guessed atom types incorrectly in the
past...)

So if you've decided you want to use one of the AMBER force fields,
then learn AmberTools, and learn InterMol(ParmEd).
If you've decided you want to use the GROMOS force field, learn the
gromacs molecule builder tools (pdb2gmx, grompp,?  etc), and use
InterMol, ...
...

Cool!

andrew