[gmx-users] Eigenvalues relative cumulative positional fluctuation

[Berk Hess]

>From: Anthony Cruz <acb15885 at uprm.edu>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>Subject: Re: [gmx-users] Eigenvalues relative cumulative positional 
>fluctuation
>Date: Wed, 17 Aug 2005 09:34:14 -0600
>
>Thank you for your responce. The rmsf option of g_anaeig gives the plot the
>RMS fluctuation per atom of eigenvectors the plot that I try to do is the
>relative cumulative positional fluctuation (with respect to
>the total positional fluctuation) vs the eigenvector index, because as I 
>see
>from the papers I have this plot is used to identify the eigenvector that
>contribute more with the overall positional fluctuation. How I can do this?
>
>See the insert in the attachment.

That looks like a graph from a paper of mine.

I don't remember how I have done it.
But this will work:

awk '{if (substr($1,1,1) != "#" && substr($1,1,1) != "@") { s+=$2; 
printf("%d %f\n",$1,s/12.3); } }' eigenval.xvg

where you should replace 12.3 by the trace given in covar.log.

Berk.