Dear Dr. Hannes Loeffler,
thank you very much for your help. Following your description, I open the PCA.pev file that I generated with the command line as follows in ptraj:
================================
trajin ../MDproduction.mdcrd.gz
rms first :1-720.CA
matrix covar name mcovar :1-720.CA
analyze matrix mcovar out PCA.pev vecs 20
==================================================
Than I found the files containt several sections.
=============================================
Eigenvector file: COVAR
1563 1563
.....................................................
****
1 790.12599
..................................................
****
2 71.09028
....................................................
so on and on till the 20th section
===========================================================
Is the 790. and 71. are eigenvalue of the first two eigenvectors?
I guess to calculate the percentage for the first three eigenvalues, i should add the number of the ****1 (i.e. 790.12599) and ****2. (ie. 71.09028) number and over all the sum of the rest's eignevalues. am I correct? Did I misunderstand how to interprete the results? Many thanks again for your help and guildance.
Best regards,
Catherine
> Date: Sat, 24 Jul 2010 08:26:27 +0100
> From: hannes.loeffler.stfc.ac.uk
> To: amber.ambermd.org
> Subject: RE: [AMBER] Where can I find a tutorial/manual on the IED in VMD for visualizing PCA data from ABMER?
>
> > In my suitation, I have 20 eigenvalues, when I open the PCA data with IED in VMD.
> > I found the cursors originally stop at the following values (20, 10, 5, 3, -1, -2,
> > -3, -5...) do you mean I should sum up all these 20 values and divide by 35 (i.e.
> > the first three eigenvalue (20+10+5), to see if the value larger than 90%.
>
> No, you misunderstood. The values above are some arbitrary (possible just
> interpolated) values from the projections that you have obtained from ptraj's
> 'projection' command. What you want is the eigenvalues from the file which you write
> with the 'analyze matrix out' command. For the sum you will have to either compute all eigenvectors/eigenvalues or directly analyze the matrix (sum of the diagonal elements)
> which you can write with the 'matrix' command.
>
>
>
> --. Hannes Loeffler e
> Drmail: h.loeffler.dl.ac.uk
> STFC Daresbury Laboratory, tel: +44-1925-603367
> Warrington, fax: +44-1925-603634
> WA4 4AD, UK
>
_________________________________________________________________
Hotmail: Powerful Free email with security by Microsoft.
https://signup.live.com/signup.aspx?id=60969
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Jul 25 2010 - 01:30:03 PDT
