Date: Thu, 28 Jun 2001 17:27:00 -0700
From: David Case
Subject: Re: folding simulations

On Thu, Jun 28, 2001, Elizabeth K. White wrote:

> This is probably ambitious, but I'd like to use AMBER to simulate
> protein unfolding, and possibly refolding. Initially, I thought that it
> would be best to do this in (explicit) water, but now I am worried that
> the solvent will blow up if I heat it much too much. The generalized Born
> solvent model looks more attractive in terms of computational time, but
> may not be accurate enough for folding studies. Does anyone have
> experience in using AMBER for folding simulations? I'd appreciate any
> advice people might have about clever tricks to try and pitfalls to avoid.
>

You could read Valerie Daggett's papers to get a feeling for what happens
with high temperature explicit solvent unfolding simulations -- also for
how long things might take, etc.

I don't think anyone has very much evidence about the performance of GB
on folding/unfolding. We do know that the current parameters don't do
a good job of matching corresponding numerical PB results across the whole
range from folded to unfolded -- this may or may not raise a red flag for
you. (Some revisions that may make GB more like PB are under development.)

You mention computational time: depending on the size of your system, you
might be surprised at how little difference there can be between GB and
explicit solvent: PME does wonders for regular simulations, but is not
(yet) available for GB. So be sure to run a few short timing comparisons
before making up your mind.

...hope this helps....dac