FW: [gmx-users] troubles with eigenvectors

[Graham Smith]

> you are right. I did covariance analysis just on the ligand and then I 
> tried to filter all atoms (protein, water and ligand).
> Right now, I tried to run covariance analysis on all atoms, but I got 
> this message:
> 
> *** malloc_zone_calloc[749]: arguments too large: -679109872,8
> Fatal error: calloc for matd (nelem=-679109872, elsize=8, file 
> g_covar.c, line 210): No such file or directory
> 
> Is the matrix too big (system has 20044 elements) or is it another kind 
> of problem

It probably is too big, since the matrix will have 9N^2 elements for N
atoms. And I'm not sure it's a good idea to include the water anyway.

I think that you should make a new .tpr file for just the ligand with
tpbconv, 
tpbconv -s full.tpr -o ligand.tpr -n ligand.ndx

Then you can use this with the filtered .xtc in ngmx. 

If you do this with e.g.just the Ca atoms in a protein then you still see
nothing because they are too far apart for ngmx to draw bonds, but it
sounds like you have all atoms of your ligand, so it should be OK. 

-G

########################################################################

Dr. Graham R. Smith, 
Biomolecular Modelling Laboratory, 
Imperial Cancer Research Fund, 
44 Lincoln's Inn Fields, 
London WC2A 3PX,
U.K. 
Tel: +44-(0)20 7269 3348 
email: graham.smith at icrf.icnet.uk
URL: http://www.bmm.icnet.uk/~smithgr

On Mon, 14 Jan 2002, Ivano Eberini wrote:

> > Have you done covariance analysis on all atoms? Otherwise your f
> > iltered
> > trajectory will contain a subset of the atoms only, and your tpr does
> >  not
> > match your filtered.xtc.
> 
?
> Is there any solution?
> Thanks for your help.
> 
> Best regards,
> Ivano
> 
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
>