Re: ABF question

From: Chris Chipot (
Date: Wed Sep 19 2007 - 08:17:12 CDT

Dear Matt,

your assumption is correct. In the example of two interacting particles
that you provide in your message, a repulsive force arises from the
difference in the availability of phase space at short and at large
separations. The Jacobian correction guarantees that all separations
will be sampled uniformly. In our implementation, the applied biasing
force is such that, upon normalization, the free energy tends, indeed,
towards zero at quasi-infinite separation, which is the expected result
when the particles are sufficiently far apart to no longer interact.

Hopefully this answers your question.

Chris Chipot

Matthew Davies a écrit :
> Hi
> I have a question regarding the application of an ABF simulation.
> In the ABF application of NAMD, the average force along the reaction
> co-ordinate xi is
> F(xi) = -dA(xi)/dxi = - <dV(x)/dxi> + 2RT/xi
> The latter term in the repulsive force is due to the greater availability
> of phase space as xi increases.
> The question is whether, in the integration of the above to give A(xi),
> the term 2RT/xi is removed in NAMD-ABF so that A(xi) tends to a constant
> as xi -> infinity. Otherwise A(xi) -> minus infinity as the resulting 2RT
> ln(xi) -> infinity (entropy increasing indefinitely) as xi -> infinity, .
> Best wishes
> Matt Davies

Chris Chipot, Ph.D.
Equipe de dynamique des assemblages membranaires
Unité mixte de recherche CNRS/UHP No 7565
Nancy Université                            Phone: +33 (0)3-83-68-40-97
B.P. 239                                    Fax:   +33 (0)3-83-68-43-87
54506 Vandœuvre-lès-Nancy Cedex
     Science without management is worse than management without science
                                                        N. G. van Kampen

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:45:16 CST