AMBER Archive (2007)

Subject: Re: AMBER: cutoff & speed-up of GBSA

From: Julien Michel (j.michel_at_soton.ac.uk)
Date: Thu May 31 2007 - 16:24:11 CDT


Hi Wei,

You could check on the same snapshot how much the effective Born radii
of the protein atoms change with rgbmax and see if that gives you clues
as to which lower value of rgbmax might be suitable. I think there is
keyword that allows you to plot these Born radii.

Julien

Wei Chen wrote:
> I haven't got a chance to use pmemd since I am not accessible to Amber9
> now.
> I am wondering what value of rgbmax should be safe for the big protein
> (25,000 atoms).
>
> Thanks,
>
> Wei
>
>
> On 5/31/07, *David A. Case* <case_at_scripps.edu <mailto:case_at_scripps.edu>>
> wrote:
>
> On Thu, May 31, 2007, Wei Chen wrote:
> >
> >
> > In order to speed up, I have to make cutoff and rgbmax as small
> as possible.
> > >From my single point calculation, I see that a 16 Angstrom
> cutoff should be
> > good. But I am not sure whether I have to use a large value for
> rgbmax,
> > since there is a difference of 500 kcal/mol between rgbmax=16 and
> > rgbmax=999, which is 1% of total elec. energy. Using large rgbmax
> really
> > slows down the simulation.
>
> Changing rgbmax changes the physical approximations made in gb
> theory: the
> effects of long-distance atoms on the local effective radii are
> ignored.
> You can't tell whether this is good or bad just by looking at the total
> energy. What is important is the way in which the potential energy
> surface is
> changed. And I don't think much is really know about this aspect.
>
> ...dac
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu <mailto:amber_at_scripps.edu>
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
> <mailto:majordomo_at_scripps.edu>
>
>

-- 
Julien Michel
http://www.julienmichel.net
Dr J. W. Essex research group
University of Southampton
United Kingdom

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