AMBER Archive (2004)

Subject: Re: AMBER: General Question about AMBER's Free Energy Perturbation Method

From: David A. Case (case_at_scripps.edu)
Date: Thu Dec 16 2004 - 01:03:18 CST

On Wed, Dec 15, 2004, Ilyas Yildirim wrote:
>
> My question is as following: Let's say that I have an organic molecule,
> like Guanine, and want to change the amino group (-NH2) to a carbonyl
> group (-O) (Or vice verse). [NH2 -> O or O -> CH2] How can this be
> specified in xleap? Or should I do a [N -> O] or [O -> N] perturbation,
> and then include the hydrogen atoms?

The procedure for specifying which atoms change is analogous to that shown in the
crown ether tutorial. There is no restriction to changing just one atom at a time.

The people who are developing free energy methods in Amber (mostly me) generally prefer
to use thermodynamic integration rather than free energy perturbation calculations, and
this is what is supported in sander in Amber 7 and 8. Gibbs can still be used, but not
with PME or with the generalized Born model, nor with potentials that use "extra points".
For these reasons, I believe that gibbs is no longer the program of choice for carrying
out the sorts of simulations you describe.

...good luck...dac

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu