[gmx-developers] Force Fields

[Erik Lindahl]

Graham Smith wrote:
> It's great to have opls now (thanks, Erik!) but this gives me a guilty
> pang about amber. How are we getting on with that? Have you had a
> chance to look at your conversion scripts again Anton? I think that
> all that was required was just to sort out the order in which some
> parameters were being printed - we could surely live without automatic
> generation of impropers for chiral centres. Sorry to be pathetic but I
> don't know awk at all so I don't feel able to make even the relatively
> small changes I think are required. Then I can have another go with
> the terminal residues if you want.
> 

Hi,

Eric Sorin in Vijay Pande's group here at Stanford is looking into it, 
and I might be able to put some time into it too - it shouldn't take too 
long.

Since OPLS is a derivative of Amber (it still uses Amber bonds & anges) 
we can use a lot of things. I introduced a new dummy database that 
contains equilibrium geometry angles and bond lengths for dummy 
constructions (note that this is not always the values in the forcefield
file if the ring is strained!), but since these are the same in Amber we
can just copy that file.

I don't think that Amber uses improper dihedrals for chiral centers - in
that case the only improper dihedrals you need those to keep groups 
planar, and they are all listed in the OPLS rtp file.

Be careful with the sigma/epsilon parameters, by the way - I think Amber
uses r_min/2 instead of sigma (which is r_0).



Cheers,

Erik