> On Feb 16, 2015, at 10:46 AM, Asmita Gupta <asmita4des.gmail.com> wrote:
>
> Dear users,
>
> I have run an AMD simulation on a 900 amino acid long protein structure.
> I wanted to reweigh 2D histogram data from PCA, using the link given on
> amber tutorial:
>
> http://mccammon.ucsd.edu/computing/amdReweighting/
>
>
> But the note at the end says that the current reweighing scheme will work
> only for proteins that are 10-40 residues long...for larger structures the
> profiles will be too noisy...
>
> Can you please suggest any other scheme which i can use for reweighing?
The fact that the profile will be too noisy is a result of the nature of aMD itself, not the standard way it is analyzed. A larger boost means you will get to sample more conformations, but since it biases away from the low energy structures, you spend more time sampling high energy structures which don't contribute much when reweighted.
This is highly related to the idea of importance sampling, which you should be able to learn about through Google if you are unfamiliar with that term.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Feb 16 2015 - 10:30:02 PST
