Re: FEP implementation in NAMD

From: Jérôme Hénin (
Date: Wed Aug 02 2006 - 21:16:17 CDT

Hi Naveen,

> I am currently trying to implement the dual-topology method in LAMMPS
> (another parallel MD code) and was wondering how the method is implemented
> in NAMD (or if there is a reference that would be terrific). Does NAMD
> currently use a soft-core approach to avoid the end-point catastrophe? I
> found an email from last year stating that it doesn't, and nothing in the
> current documentation or tutorial stating otherwise, so I assume not.
That's correct, although some options are currently being tested, so future
versions of NAMD should have that feature.

> Does
> NAMD take the approach that Charmm takes, applying a correction to the
> potential postpriori? Or is the potential modified before the beginning of
> the simulation (say, applying the lambda scaling directly to the charges
> and lennard jones coefficients of each group).
The scaling is not applied to charges or LJ coefficients, but directly to the
forces and energies.

> Also, are within group
> interactions scaled as well (ie, say you mutate an alanine to a lysine -
> does the terminal amine group interact with lysine beta carbon at the
> scaled level or normal level?)
No, within-group interactions are preserved. The scaling only affects
nonbonded interactions between one atom of the initial state and one atom of
the final state.


This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:43:53 CST