[gmx-users] Covariance analysis of X-ray ensemble

[James Starlight]

Tsjerk, thanks for suggestion!

By the way I've found in the pca.log ( that time pca was done on
trajectory as well as TRP ( not pdb) files as the inputs ):

Read 30 frames from cam_xray_coors_bb.pdb (time 1 to 30 ps)
Read reference structure for fit from x_ray.tpr

Analysis group is 'System' (575 atoms)
Fit group is 'System' (575 atoms)
Analysis is non-mass weighted
Fit is non-mass weighted
Diagonalized the 1725x1725 covariance matrix
Trace of the covariance matrix before diagonalizing: 350.847
Trace of the covariance matrix after diagonalizing: 350.847

Wrote 1725 eigenvalues to PCA_eigenval.xvg
Wrote reference structure to PCA_eigenvec.trr
Wrote average structure to PCA_average.pdb and PCA_eigenvec.trr
Wrote eigenvectors 1 to 29 to PCA_eigenvec.trr

It seems that cov.analysis was done also as  non-mass weighted.
Actualy I extract eigenvectors for further essential dynamics sampling
runs. In the literature I've found that for EDA eigenvectors from only
non-mass weighted PCA are suitable. So if its true in what cases mass
weighted eigenvectors might be used?

James

2013/1/11 Tsjerk Wassenaar <tsjerkw at gmail.com>:
> Hi James,
>
> You don't need a .tpr file. A .gro or .pdb file will do as well. Unless you
> insist on doing mass-weighted PCA.
>
> Cheers,
>
> Tsjerk
>
> On Fri, Jan 11, 2013 at 4:04 PM, James Starlight <jmsstarlight at gmail.com>wrote:
>
>> Dear Gromacs users!
>>
>>
>> I want to perform Covariance analysis of my x-ray data sets via
>>
>> g_covar -f trr.pdb -s x_ray.tpr -n index -o PCA_eigenval.xvg -v
>> PCA_eigenvec.trr
>>
>> where trr.pdb is the trajectory made from the 30 x-ray structures
>> (only c-alpha atoms were selected. How I could obtain x_ray.tpr file
>> for such analysis?
>>
>> I've tried to use tpr from full atomic model of the same protein which
>> I';ve simulated recently. For that p[rotein tpr file have been already
>> created.  From that model via make_ndx I've defined only backbone
>> atoms corresponded to the atoms in my pdb ensemble. On the next step
>> I've used tpbconv on protein.tpr with index file to make reduced tpr
>> file with atoms corresponded to the backbone of the pdb-ensemble. Now
>> when I run g_covar with my ensemble and reduced tpr file I obtain
>> error
>> *** glibc detected *** g_covar: malloc(): memory corruption:
>> 0x00007f27df53e018 ***
>>
>>
>> On other hand if I specify for g_covar full-atomic tpr file with the
>> index ( where atoms corresponded to the trr.pdb are mentioned)
>>
>> g_covar -f trr.pdb -s protein.tpr -n index -o PCA_eigenval.xvg -v
>> PCA_eigenvec.trr
>>
>> I obtain another error
>>
>> Calculating the average structure ...
>> Reading frame       0 time    1.000
>> WARNING: number of atoms in tpx (575) and trajectory (575) do not match
>>
>> -------------------------------------------------------
>> Program g_covar, VERSION 4.5.5
>> Source code file: /tmp/build/gromacs-4.5.5/src/gmxlib/mshift.c, line: 102
>>
>> Fatal error:
>> Molecule in topology has atom numbers below and above natoms (575).
>> You are probably trying to use a trajectory which does not match the
>> first 575 atoms of the run input file.
>>
>>
>>
>> What I've done wrong and how I can define tpr file for trajectory made
>> from several pdb files ?
>>
>>
>> Thanks for help
>>
>> James
>> --
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> * Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>
>
>
> --
> Tsjerk A. Wassenaar, Ph.D.
>
> post-doctoral researcher
> Biocomputing Group
> Department of Biological Sciences
> 2500 University Drive NW
> Calgary, AB T2N 1N4
> Canada
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> * Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists