AMBER: <dV/dL> fluctuations in TI

From: Nelson Fonseca <nfonseca.dq.ua.pt>
Date: Wed, 17 Nov 2004 11:39:01 +0000

Experience is what you get when you don't get what you
want.

                      -- Dan Stanford--


Dear all,


Im using TI within sander in order to compute the free
energy between two enantiomeric complexes, via an
intermediate molecule that is transformed into R or S by
mutation of the apropriate group into dummy atoms.
Im using 12 points for Gaussian Quadrature, the same
points recomended in the manual, to perform data
collection for <dV_dL>. I also set klambda=6 and use 20 ps
for equilibration period and 30 ps for collect averages.
When I look at the results of dV_dL for the fisrt 5 points
the behavior is good, having relatively small fluctuation
dor dV_dL, but for the last points I have very larges
fluctuations for <dV_dL>.
My question is: How can I reduce the large fluctuation for
the last points? Should I use more equilibration time? Or
changing the value of klambda to 4 or 5?



Thanks to all in advance.

Regards

Nelson Fonseca
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Nov 17 2004 - 11:53:01 PST
Custom Search