Re: [AMBER] P2 correlation functions questions

From: David Case <>
Date: Wed, 8 Jan 2020 02:29:27 +0000

On Tue, Jan 07, 2020, erik zuiderweg wrote:
>cpptraj ../ << EOF
>trajin ../BETA_APO_MIN_HEAT_EQUIL_MD.x 1 800
>vector v2 :2.H :2.N
>analyze timecorr vec1 v2 out corr_2_NH

How many frames are in the trajectory file? Generally, only the first
5-10% of the correlation function is worth much. Internal correlation
functions for NH groups in proteins generally reach plateaus in about 50
ps or less, suggesting that you want 1 nsec or so of sampling at a

To get internal motion, you would want to remove overall rotation first,
usually by doing an rms fit on (say) the secondary structure parts of
the protein. If you want to study overall tumbling as well as internal
motion, then you should not do the RMS fit. (But then you will need
long-long trajectories, and no thermostats, etc.)

A dated, but still pretty relevant paper is this one:

%A V. Wong
%A D.A. Case
%T Evaluating Rotational Diffusion from Protein MD Simulations
%J J. Phys. Chem. B
%V 112
%P 6013-6024
%D 2008

There have been improvements in force fields since 2008, but the basic
ideas are little changed.

>In real life, the P2 is taken between the direction of the NH vector and
>the magnetic field.

The calculation computes P2 of the angle between the vector at time 0
and the same vector at time tau later. This computation is invariant to
what the PDB "Z" direction is. There is no notion of a reference axis.

...hope this helps....dac

AMBER mailing list
Received on Tue Jan 07 2020 - 18:30:02 PST
Custom Search