[gmx-users] Re: [gmx-developers] AMBER (parm96) force field & hydrogen/termini database

[Erik Lindahl]

Hi,

Actually we probably have this covered already - I've been cooperating 
with Vijay Pande here at Stanford to get Gromacs into folding at home, and 
they are willing to do the work to implement at least the all-atom 
protein versions of OPLS-aa, CHARMM and AMBER.

I've modified the topology format the last couple of days to support 
separate bonded and non-bonded atom types, and it is also possible to 
define dihedraltypes from all for atoms, or even with wildcards, so it 
should be quite straightforward - I've asked a couple of questions about 
wildcarded types in OPLS to Bill Jorgensen, but once that's settled they 
will get to work on the other forcefields.

The only problem with the termini database as far as I see it is to 
determine the atom types to use for the changed atoms, right?

If you're interested in hacking pdb2gmx it would be very nice if we 
could implement automatic generation of improper dihedrals!

Cheers,

Erik