Dear Hossein
The moltemplate implementation of the GAFF forcefield has been tested by several people and they say it is accurate now. (It wasn’t at first. The GAFF2 force field probably works as well because it was automatically converted from a file with an identical file format as GAFF. But it has not yet been tested.)
The main issues with using GAFF/GAFF2 in moltemplate are:
-
If my understanding is correct, it was not intended for proteins and nucleic acids. It was intended for small organic molecules. Hopefully that’s not a problem for you.
-
Moltemplate does not calculate the partial charge on each atom correctly for the AMBER force fields. Moltemplate is a (relatively) simple database lookup program. It does not perform quantum-chemistry calculations to determine atomic charges (AmberTools does this). You will have to use AmberTools or some other quantum code to calculate them (or lookup the partial charges from “oplsaa.lt”. That’s what the AMBER examples which come with moltemplate do.) . Afterwards, you can use commands like
set atom 70 charge -0.212
to assign partial charges to each atom.
For details, see http://lammps.sandia.gov/doc/set.html)
and
http://ambermd.org/antechamber/gaff.html
Cheers
-andrew