[gmx-users] Re. normalized overlap

[Lovas, Sandor]

Hello Mark,
Thanks for the clarification, I have checked the trr file with gmx dump
and, indeed, the eigenvalues are are written in them as time, for
example 

"gmx dump -f eigenvec_5.trr
...
eigenvec_5.trr frame 2:
natoms=       443  step=         1  time=1.5521152e+01  lambda=    0
..." 
and 1.5521152e+01 is the first eigenvalue in the eigenval_5.xvg file.

I think the manual, however, needs to be clarified since it implies
that you need to use the eig and eig2 inputs to get the normalized
overlap:

"When -v, -eig, -v2 and -eig2 are given, a single number for the
overlap between the covariance matrices is generated. The formulas
are:..."  
Thanks again,
Sandor



>Hi,
>
>GROMACS tools write the eigenvalue to the time field of trr files, so
>both of your runs are presumably computing the same quantity from the
>same input data (which you can check by comparing the .xvg files to
>the output of gmx dump on the trr files). I will amend the docs of gmx
>anaeig to note that the -eig and -eig2 options permit you to supply
>eigenvalues, which will replace those read from the trr file. So, if
>what you want is to do the analysis with eigenvalues all zero, then
>make copies of your .xvg files and zero the values suitably.

Mar
-- 
______________________________________
Dr. Sandor Lovas, Professor            
Department of Biomedical Sciences      
Creighton University                   
2500 California Plaza,                 
Omaha, NE 68178, USA 

Phone: 402-280-5753
Fax:   402-280-2690