AMBER GAFF FF available inside moltemplate builder (v1.12), but testing needed

The preliminary version of moltemplate v1.12 supports canned
force-field(s) and "automagic" force-field assignment for simple
organic molecules.

It should now be possible to prepare molecule simulations for LAMMPS
using the simple AMBER force fields entirely using moltemplate (ie
without using ambertools or amber2lmp). This is a experimental
feature. The resulting LAMMPS files need testing for accuracy (by
you, see below).

Currently only the GAFF force-field (January 2013 version) is has been
converted into moltemplate format, but it should not be difficult to
add support for the other (lammps-compatible) AMBER force-field
variants, once I am sure that the current GAFF conversion is behaving

To try this out, users must have a very recent version of of
moltemplate (version 1.12 or later). This can be downloaded at You also currently need to compile LAMMPS with
the USER-MISC package enabled. (It uses dihedral_style fourier.)

  --- example ---

There is currently only one example which uses this force field:
To find it, download the examples at, and go to this

I would not normally post here, but a serious bug in moltemplate was
discovered which was preventing it from generating GAFF improper
interactions correctly.

I believe the newest version (v1.21) fixes this bug. If you use
AMBER/GAFF with moltemplate, please upgrade to version 1.21 or later.

I also simplified the examples in

Hopefully, it should be much now be easier to understand how these
examples work.

This software won't get better without bug reports.
Please post if you notice a discrepancy between the topology or
force-field files generated by moltemplate, with the files generated
by other molecule builders which support GAFF!