From: Aron Broom (broomsday_at_gmail.com)
Date: Thu Nov 15 2012 - 14:52:48 CST
Since it can be challenging to get an error estimate from ABF or
Metadynamics, what you can do is to take the bias (i.e. your .colvars.state
file) and apply that bias without further updating (i.e. just a general
reaction coordinate dependent bias) and do some fairly short and well
spaced umbrella sampling.
This has a number of benefits:
1) You can use the umbrella sampling results to correct for errors in your
2) Umbrella sampling (at least in 1D) readily gives good error estimates.
On Thu, Nov 15, 2012 at 5:50 AM, Jérôme Hénin <jerome.henin_at_ibpc.fr> wrote:
> Hello, Insert Name Here!
> There have been a few discussions about this in the past. Here are
> some links (you might find some more):
> If you have a more specific question that has not been covered yet,
> I'll be happy to discuss it.
> On 15 November 2012 10:38, dream <dream3773_at_163.com> wrote:
> > Dear all,
> > I've met some problems when using the ABF method. Is there a standard to
> > judge whether the result is believable? For example, the PMF won't
> change a
> > lot(How long?How much?); the count density should be good(What about the
> > level?). Is there other ways to decide it?
> > Sometimes when using different way dividing windows or using different
> > starting structures, the result would be different. Is the force not
> > correct? Which parameter should be changed?
> > Could anybody please tell me how to solve these problems? Thank you so
-- Aron Broom M.Sc PhD Student Department of Chemistry University of Waterloo
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:45 CST