[gmx-users] how to obtain error bar for plotting?

[Tsjerk Wassenaar]

Hey :)

Just some side notes. Make sure to distinguish between 'statistical
error' or 'standard error of the mean', which is a measure of how
'correct' your estimate of the average is, and the standard deviation
or the variance, which is a measure of the natural spread. The latter
is not an error! In addition, there are correlations that make it
difficult to assess the standard error of the mean. These correlations
can be time correlations, which cause a false decrease of the standard
error. But simulations that start from similar starting points are
also correlated. If the simulations are too short, and 1 ns seems
quite short, then the observations will not be statistically
independent, and the standard error will again turn out smaller than
it should.

Cheers,

Tsjerk

On Thu, Dec 29, 2011 at 2:39 AM, Mark Abraham <mark.abraham at anu.edu.au> wrote:
>
>
> On 28/12/11, leila karami <karami.leila1 at gmail.com> wrote:
>
> Dear Mark
>
>
> can I obtain error bar twice the standard error of the average
>
>
> If you've used g_energy -av on multiple sets of data, then it will plot
> error bars corresponding to various quantities. Since twice the standard
> error isn't an option it supports (as you know from reading g_energy -h)
> then you'll have the option of accepting one it does support or looking at
> the output file with error bars of (say) one standard error and
> post-processing to double the size.
>
>
>
> distance between the donor atom of protein and acceptor atom of dna over 1
> simulation?
>
>
> You've made one observation from one simulation, so there is no statistical
> error to report.
>
>
>
>
> should I do several simulation. Is 1 simulation enough?
>
>
> Very likely you should do several simulations each much longer than 1ns. You
> should also examine the literature for people who've tried to observe
> similar things on similar systems to use their approach as a guide for your
> own.
>
> Mark
>
>
>
>
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists



-- 
Tsjerk A. Wassenaar, Ph.D.

post-doctoral researcher
Molecular Dynamics Group
* Groningen Institute for Biomolecular Research and Biotechnology
* Zernike Institute for Advanced Materials
University of Groningen
The Netherlands