[gmx-users] Correlation of dynamics between pairs of residues

[Tsjerk Wassenaar]

Hi,


Perhaps "defeats the purpose" wasn't the right way to say it.  But simply
> plotting the covariance matrices with arbitrary min/max is just an
> aesthetic thing, so it's not really functional.  Hence why no one has coded
> that.
>
>
I disagree with that. If you have simulations of the same system under
different conditions, you want to be able to see differences in the
covariance matrices. For that it is necessary that these use the same color
scale, so use the same min/max. In addition, you could think of a situation
where the covariance in one loop is large, such that it obscures other
features in the covariance matrix. Then you'd like to use a minimum/maximum
(truncation) color to allow assessment of the features up to a certain
covariance where it just says 'larger/smaller than'.

Hence also why I have coded that :) I can send an XPM parsing python script
that allows rewriting xpm files to colore .ppm files, with a specified
min/max. It's based on my xpm to generic data file converter that is
floating around already, and that could be used for the same purpose, with
the aid of, e.g., R.

Cheers,

Tsjerk

-- 
Tsjerk A. Wassenaar, Ph.D.