AMBER Archive (2002)

Subject: Re: MM-PGBSA Question!!

From: David A. Case (case_at_scripps.edu)
Date: Tue Oct 22 2002 - 09:45:20 CDT

On Tue, Oct 22, 2002, Bernd Wellenzohn wrote:
>
> 1.)We performed several mm-pbsa calculations of DNA and
> although two simulations starting from different
> conformations exhibit a similar mean value (+- 1kcal/mol)
> of the free energy
> the standard deviations are extremly large (about 20 kcal/mol).
> The standard deviations of similar calculations (Srinivasan et al. 1998 )
> is much smaller while Wu et al. 2002 had comparable standard
> deviations. Does anyone have an idea why this happens, and is the
> calculation of small (about 10 kcal/mol) delta G still allowed.

Be sure to distinguish between the standard deviation of the distribution,
(which will be something like what you describe above), and the estimated
standard error of the mean, which will depend on how much sampling you
do. Roughly, the estimated standard deviation in the mean is smaller
than the std. dev. of the distribution by a factor of the square root of the
number of independent samples you have. So, if you have 100 snapshots that
are sampled far enough apart to be independent, the standard error of the
mean is 10 times smaller than that of the sample standard deviation.

In the 1998 paper, we reported the standard errors in the mean values (see
Footnote to Table 1), since it is the mean value that is the interesting
quantity.

..hope this helps...dac 

-- 

==================================================================
David A. Case                     |  e-mail:      case_at_scripps.edu
Dept. of Molecular Biology, TPC15 |  fax:          +1-858-784-8896
The Scripps Research Institute    |  phone:        +1-858-784-9768
10550 N. Torrey Pines Rd.         |  home page:                   
La Jolla CA 92037  USA            |    http://www.scripps.edu/case
==================================================================