Re: double-wide sampling with soft-core

From: Chris Chipot (chipot_at_ks.uiuc.edu)
Date: Thu Jan 22 2009 - 12:23:55 CST

Floris,

> One minor point, there's an explicit alchDoubleWideSampling parameter documented, maybe this could be dropped and inferred from the use of both lambdaPlusDelta and MinusDelta?

alchDoubleWideSampling < Will an FEP alchemical transformation be
                          performed in the forward and the backward
                          directions concomitantly? >
Acceptable Values: on or off
Default Value: off
Description: Simultaneous alchemical transformations in the forward and
              the backward directions constitute a straightforward route
              towards the estimation of the error associated to the
              corresponding free energy calculations through their
              hysteresis. When set to on, this option will request the
              concomitant evaluation of the ensemble averages over
              configurations representative of state λ, going (i) from 
              to λ+δλ (controlled by keyword alchLambdaPlusDelta), and
              (ii) from λ to λ-δλ (controlled by keyword
              alchLambdaMinusDelta). Accessing from a single free energy
              calculation the statistical information of both
              transformations is particularly useful for data
              post-processing, e.g.BAR and SOS, which have proven to
              improve the accuracy of the free{energy estimates over
              brute-force application of the FEP formula. It should be
              emphasized, however, that forward and backward
              transformations do not necessarily share the same
              convergence properties, so that, over finite-length
              simulations, free-energy differences will not be strictly
              equal.

As you correctly pointed out yesterday, and though everyone probably
knew what was implied in the Jorgensen-Ravimohan λ+δλ notation, one
should really think in terms of differences between λ and λ2, which
subsumes different δλ's along the reaction path.

Chris

_______________________________________________________________________

Chris Chipot, Ph.D.
on leave from Nancy Université, CNRS
Theoretical and Computational Biophysics Group
Beckman Institute
University of Illinois at Urbana-Champaign
405 North Mathews Phone: (217) 244-5711
Urbana, Illinois 61801 Fax: (217) 244-6078

                             E-mail: chipot_at_ks.uiuc.edu
                                     Christophe.Chipot_at_edam.uhp-nancy.fr
                             Web: http://www.ks.uiuc.edu
                                     http://www.edam.uhp-nancy.fr

       To sin by silence when we should protest makes cowards out of men
                                                     Ella Wheeler Wilcox
_______________________________________________________________________

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:52:17 CST