[gmx-users] How to compute covariance?

[Semen Esilevsky]

Dear gmx users,
I'm sorry for stupid question, but I can't find the
answer for two days :(
The problem is how to compute the correlation (or
covariance) of two particular atoms using the
MDtrajectory. I just need a correct equation for this
(not the program, which does this).
If I have the atoms i and j I'd rather compute the
covariance between them as

cov_ij = <(xi-xi_mean)*(xj-xj_mean)> + 
         <(yi-yi_mean)*(yj-yj_mean)> +
         <(zi-zi_mean)*(zj-zj_mean)>

But I'm confused about the "off-diagonal" averages
like
<(xi-xi_mean)*(zj-zj_mean)>. Do I have to include them
or not and how? 

Thank you in advance for help!

Semen


		
__________________________________ 
Do you Yahoo!? 
Jazz up your holiday email with celebrity designs. Learn more. 
http://celebrity.mail.yahoo.com