Correlation function can be negative and even can increase (not above 1
though). However, I would not trust the reliability of the function up to
1/1000 of the total sampling time or at most 1/100. It means you have 1000
ps data and your correlation function may be reliable only up to 1 ps - 10
ps. If you want to be more careful, try to measure standard deviation of the
correlation function.
On Wed, Sep 7, 2011 at 1:51 PM, Sindrila Dutta banik <
sindrila.duttabanik.yahoo.com> wrote:
> Dear All
>
> I have computed atomic fluctuation correlation function for a particular
> atom of a protein system based on the trajectory file generated
> using AMBER. I have trajectory file for 1000ps with time step 4fs. I
> considered elapsed time 20ps and averaging the correlation function over
> different time origin. The result shows an unusual feature. As I am
> averaging over 200ps it decreases and become negative within 20ps. However,
> the nature of the plot is changed as I am averaging over 1000ps. The rate of
> decrease of correlation function gradually decrease as the time upto which
> time average performed is increased from 200ps to 300ps ...1000ps.
>
> If you kindly suggest me what could be the possible reason for this, it
> will be very helpful for me. I use a in-house program to compute the
> correlation function. Is this can be a possible source of error? Then how to
> varify it? Please let me know,
>
> With best regards
> Sindrila
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Best,
InSuk Joung
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Sep 07 2011 - 11:30:06 PDT
