AMBER Archive (2008)

Subject: RE: AMBER: Correlation functions from iRED analysis

From: Samuel Genheden (a03samge) (a03samge_at_student.his.se)
Date: Thu Jun 19 2008 - 08:10:51 CDT

Helo,

I still get identical correlation functions for all the N-H vectors, even though I do a RMS fit. And I thought one of the advantages of doing an iRED analysis was that I should not do an RMS fit. Isn't that correct?

/ Samuel

-----Original Message-----
From: owner-amber_at_scripps.edu on behalf of Myunggi Yi
Sent: Thu 6/19/2008 2:52 PM
To: amber_at_scripps.edu
Subject: Re: AMBER: Correlation functions from iRED analysis
 
On Thu, Jun 19, 2008 at 6:22 AM, Samuel Genheden (a03samge) <
a03samge_at_student.his.se> wrote:

>
> Hello, Amber users
>
> I'm studying a protein using MD and would like to calculate correlation
> functions with the iRED method in order to compare order parameters from
> NMR. My protein is 138 residues long and contains 10 prolines, and therefore
> I have 128 N-H vectors. I'm using Amber10. My input file looks like this:
>
> trajin ../mdcrd5.gz
> trajin ../mdcrd6.gz
> vector v2 :2_at_N ired :2_at_H
> vector v3 :3_at_N ired :3_at_H
> ..
> vector v138 :138_at_N ired :138_at_H
> matrix ired name matired order 2
> analyze matrix matired vecs 128 out ired.vec
> vector v2 :2_at_N corrired :2_at_H order 2 modes ired.vec beg 1 end 128 npair 1
> vector v3 :3_at_N corrired :3_at_H order 2 modes ired.vec beg 1 end 128 npair 2
> ..
> vector v138 :138_at_N corrired :138_at_H order 2 modes ired.vec beg 1 end 128
> npair 128
> analyze timecorr vec1 v2 tstep 10.0 tcorr 10000 out Ired/v2.out
> analyze timecorr vec1 v3 tstep 10.0 tcorr 10000 out Ired/v3.out
> ..
> analyze timecorr vec1 v138 tstep 10.0 tcorr 10000 out Ired/v138.out
>
> (I've have also tried to break it up in two ptraj scripts, since the manual
> is a little bit vague if this is neccessary.) The problem is that all the
> correlation functions calculated, v2.out, v3.out, .. v138.out is the same,
> i.e. all the output files contains the same numbers. What am I doing wrong?
> I can hardly believe that all the correlation functions should be identical.
>
> And when I'm at writing - what is the best way to obtain the order
> parameters from the correlation functions?
>

To get the generalized order parameters, you need to calculate
auto-correlation functions after "rms fitting".
Then fit your graph with single or double exponential functions.
The plateau corresponds to the S^2.

>
> Best regards, Samuel
> 

-- 
Best wishes,

Myunggi Yi PhD
==================================
KLB 419
Institute of Molecular Biophysics
Florida State University
Tallahassee, FL 32306

Office: (850) 645-1334
http://www.scs.fsu.edu/~myunggi

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu