AMBER Archive (2007)

Subject: AMBER: Reply regarding PCA with ptraj and IED

From: Steve Spronk (spronk_at_umich.edu)
Date: Tue Jun 19 2007 - 16:00:37 CDT

I ran into a little trouble today preparing my trajectory for essential
dynamics visualization with IED. I came across these messages that had an
identical error to mine, but there wasn't really any discussion of how to
fix it. I think I figured out the issue, so I thought I should put it in
the written record. No need to reply-my stuff seems to be working now.

 

As both Vlad and Miguel did (see below), I ran the following two ptraj files
in succession:

 

** evec.in **

trajin <filename>

matrix covar name mcovar

analyze matrix mcovar out evec.pev vecs 3

 

** proj.in **

trajin <filename>

projection modes evec.pev out proj.ppj beg 1 end 3

 

The first ptraj run using evec.in ran fine, and the output file had no
problems. However, running the second file with ptraj resulted in the
output:

PTRAJ: projection modes evec.pev out proj.ppj beg 1 end 3

FYI: Number of read evecs is 1, number of requested evecs is 3

Mask [*] represents 6021 atoms

WARNING in ptraj(), transformProjection: no. of atom coords does not agree
with no. of vect elementsignoring command

 

And of course, my proj.ppj file was empty because the command was ignored.
I double checked evec.pev, and it did have the correct number of items in
each of the three vectors. What was going wrong?

 

Vlad's work (below) indicated that it may be a mask issue, but I didn't
actually specify a mask. Even so, I tried to specify a mask of @1-6021 and
the warning message still appeared. However, when I started using fewer
atoms, such as 2000 (by using mask @1-2000) the calculation ran perfectly
fine. It seems likely that there's some sort of memory error, and if there
are too many atoms in the mask, this error is the result.

 

The magic number for my system was 3333. When I specified only atoms
@1-3333 in the masks, the calculation ran fine, but specifying a mask of
@1-3334 led to the warning message.

 

If you see this error, try again using only part of the system, say, only
heavy atoms or only the backbone atoms.

 

Steve

 

 

******

> From: Vlad Cojocaru <Vlad.Cojocaru.eml-r.villa-bosch.de>

> Date: Tue Jan 24 2006 - 16:02:20 GMT

>

> Dear ambers,

> When using the inputs below to perform PCA on a protein trajectory

> (file.mdcrd containing a full protein but no ions or water), ptraj

> always fails to create the projection file (${fname}.ppj -- second

> step), the error being listed below. Obviously, the error has something

> to do with the mask because if I replace @~H* with @CA, the projection

> file is properly created. However I would like to be able to use all the

> heavy atoms so that I can visualize the results in IED using a pdb file

> with all the heavy atoms in the protein (I believe, IED doesn't allow me

> to use a pdb of the entire protein but eigenvectors calculated only for

> the CA atoms).

> If somebody has experienced such a problem, could you give me an

> advice on how to deal with this?

> I have to mention also that the modesfile ${fname}.pev created in the

> first step looks OK with 25 eigenvectors output and in the ptraj log

> file of the 2nd step the mask seems to be properly recognized (PTRAJ:

> projection modes ${fname}.pev out ${fname}.ppj beg 1 end 25 @~H* \

> Mask [~@ H*] represents 3736 atoms), with 3736 being the appropriate

> number of heavy atoms in my protein. The error comes on STDOUT output.

>

> Thanks in advance

> vlad

>

> /input1:/

>

> trajin file.mdcrd

> rms first @~H*

> matrix covar name mc @~H*

> analyze matrix mc out ${fname}.pev vecs 25

>

> /input2:/

>

> trajin file.mdcrd

> projection modes ${fname}.pev out ${fname}.ppj beg 1 end 25 @~H*

>

> /ptraj error (file ${fname}.ppj is blank -- only the header present)/

>

> FYI: Number of read evecs is 1, number of requested evecs is 25

> FYI: Number of read evecs is 1, number of requested evecs is 25

> WARNING in ptraj(), transformProjection: no. of atom coords does not

> agree with no. of vect elementsignoring command

>

> --

> Dr. Vlad Cojocaru

> EML Research gGmbH

> Molecular and Cellular Modeling Group

> Schloss-Wolfsbrunnenweg 33

> 69118 Heidelberg, Germany

> Phone: +49-6221-533266

> Fax: +49-6221-533298

> e-mail: Vlad.Cojocaru_at_eml-r.villa-bosch.de

> http://projects.villa-bosch.de/mcm/people/cojocaru/

>

*****

> From: Vlad Cojocaru <Vlad.Cojocaru.eml-r.villa-bosch.de>

> Date: Wed Jan 25 2006 - 13:28:42 GMT

>

> Dear amber users,

>

> Regarding my last email about the PCA in ptraj, I did some tests and

> I realized that the problem is indeed the mask in the ptraj projection

> command, more precisely it seems to be the wild card * which creates the

> problems.

>

> So in summary:

>

> projection modes ${fname}.pev out ${fname}.ppj beg 1 end 25

> @CA -- WORKS FINE (PPJ file created properly)

> projection modes ${fname}.pev out ${fname}.ppj beg 1 end 25

> @CA,N,C,O -- WORKS FINE (PPJ file created properly)

> projection modes ${fname}.pev out ${fname}.ppj beg 1 end 25

> @~H* -- DOESNT WORK (ppj file not created)

> projection modes ${fname}.pev out ${fname}.ppj beg 1 end 25

> @C*,N*,O*,S*,P*,F* -- DOESN'T WORK (ppj file not created)

>

> This problem seems to affect only the projection command, the matrix and

> analyze matrix commands work fine with all the above mentioned masks!

>

> Hope this helps fixing this issue!

>

> Best regards

> vlad

>

> --

> Dr. Vlad Cojocaru

> EML Research gGmbH

> Molecular and Cellular Modeling Group

> Schloss-Wolfsbrunnenweg 33

> 69118 Heidelberg, Germany

> Phone: +49-6221-533266

> Fax: +49-6221-533298

> e-mail: Vlad.Cojocaru_at_eml-r.villa-bosch.de

> http://projects.villa-bosch.de/mcm/people/cojocaru/

>

****

> From: Miguel Ferreira <tetrass.hotmail.com>

> Date: Wed Nov 01 2006 - 14:11:57 GMT

>

> Dear amber users,

>

> I want to perform essential dynamics using ptraj and IED software. After

> strip the waters and ions of my trajectory I used the following scripts to

> calculate the eigenvectors and the corresponding projections:

>

> trajin mds_semH2O.mdcrd.gz

> rms first *

> matrix covar name mcovar

> analyse matrix mcovar out evec.pev vecs 25

> go

>

> trajin mds_semH2O.mdcrd.gz

> rms first *

> projection modes evec.pev out proj.ppj beg 1 end 25

> go

>

> The first script ran ok and I obtained the evec.pev file. However, the

> second script gave me the following error at command line:

>

> --- > FYI: Number of read evecs is 1, number of requested evecs is 25

> WARNING in ptraj(), transformProjection: no. of atom coords does not agree

> with no. of vect elementsignoring command

>

> Does anyone knows what this means? Any suggestions to solve this problem?

>

> Thanks in advance

> Miguel

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu