Re: alchemical free energy perturbation

From: by way of Jérôme Hénin (
Date: Mon Jan 05 2004 - 09:10:52 CST


In your case, it will probably be easier to make the particle vanish than to
make it appear, because when you simulate at lambda=0, the particle may
overlap with water. To make it grow or shrink would require using the "single
topology" method, wich is not supported by the FEP code in NAMD. However,
there is no good reason why the dual-topology version should fail to describe
your transformation.

In your script, the following line :
while {$step <= 1.00} {

should be :
while {$step < 1.00} {

so that the last lambda window is [0.99, 1.0] and not [1.0, 1.01].

I checked the alchemical FEP tutorial : it contains the same error. I suppose
you took it from there - sorry about that. I'm going to correct it as soon as

About the accuracy of your results, you should monitor the convergence of dG
(the last column of the FEP output) in every window. 6400 fs is rather short,
especially if half of it is not used for statistics. If the convergence is
not reached, increase the sampling time (it may be wise to use larger times
only for certain slow-converging windows). With a system as simple as yours,
however, you shouldn't need a huge amount of sampling to obtain good results.


> Hi,
> I want to calculate the free energy change of growing a 3.0 ang. radius LJ
> particle in water. I tried dual-topolgy method outlined in tutorial on NAMD
> site but did not give expected free energy values. Is there any way to
> 'grow/shrink' rather than going from lambda=0 (nothing) to lambda =1 (full
> partcle)? Are they the same? I have also attached the config file, part of
> the fep file (system.fep). The system is a LJ particle with 1000 water
> molecules. thanks,
> Amit.
> PS: CMT is the atom type of the LJ particle that I have defined.

Jérôme Hénin
Equipe Dynamique des Assemblages Membranaires
Université Henri Poincaré / CNRS    UMR 7565
B.P. 239	54506 Vandoeuvre-lès-Nancy Cedex
Tel : (33) 3 83 68 43 95	Fax : (33) 3 83 68 43 71

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:37:16 CST