Re: About RMSD based PMF's

From: Jérôme Hénin (jhenin_at_ifr88.cnrs-mrs.fr)
Date: Fri May 14 2010 - 06:05:28 CDT

Hi Felipe,

You've got it exactly right: the "RMSD = 0" state is extremely
unlikely, and more generally, the phase space volume corresponding to
a given RMSD value decreases very rapidly as RMSD approaches zero
(precisely as RMSD to the power 3N-4, where N is the number of atoms
involved), so does entropy. As a direct result, the PMF goes to
infinity there, even if the reference conformation is a very stable
one! In that case, the stable basin should be visible at small values
of the RMSD.

Cheers,
Jerome

On 14 May 2010 06:03, Felipe Merino <felmerino_at_uchile.cl> wrote:
> Hi,
>
> Maybe i am a little lost with this, but i have been performing some PMF
> calculations using RMSD as the order parameter. As the simulations
> approaches zero RMSD the free energy becomes very high i guess caused by a
> "conformational stretching".
> If we consider the number of microstates (really the volume of the phase
> space compatible with the E + dE energies given by the hamiltonian -
> including the harmonic RMSD constraint) that the system can show as a
> function of RMSD they seem to have an inverse relation, Indeed it is obvious
> when you consider that for the extreme case where RMSD is 0 there is just
> one microstate possible (considering only the constrained atoms). This
> should put an extra entropic component in the reconstructed PMF that i do
> not know if NAMD accounts for. I think that could explain some crazy things
> that i see on my PMF's.
>
> Any hints would be appreciated
>
> best
>
> Felipe
>
> Laboratorio de Bioquímica y Biología Molecular
> Facultad de Ciencias
> Universidad de Chile
>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 05:22:59 CST