Re: free energy pertubation,binding energy and constant

From: Chris Chipot (Christophe.Chipot_at_edam.uhp-nancy.fr)
Date: Wed Jun 29 2005 - 08:14:27 CDT

Sabri:

computation of binding free energies, say between a protein and
a ligand, can be achieved following a thermodynamic cycle:

                         ?Gbind,A
(protein)aq + (ligand A)aq ? (protein...ligand A)aq
     ?Gmut,1 ? ? ?Gmut,2
(protein)aq + (ligand B)aq ? (protein...ligand B)aq
                               ?Gbind,B

In general, the direct transformation (?Gbind,A/B) is not doable
on account of the extraordinary sampling issues it would involve.
One can, however, devise point mutations, in which the ligand is
altered into an alternate one (?Gmut,1/2). Closure of the cycle
gives: ?Gbind,B - ?Gbind,A = ?Gmut,2 - ?Gmut,1.

One possible route towards this goal is to use free energy
perturbation, which defines the change in free energy between
state a and state b as:

?Gmut = -1/? ln ? exp -?[Vb(x) - Va(x)] ?a

where Va(x) is the potential energy function for state a and
? ?a is an ensemble average over configurations representative of
state a. A trajectory is thus generated using state a as a
reference and the above difference in the potential energies is
evaluated on the fly. This methodology is available in NAMD.

Your only responsibility, as an end-user, is to define the point
mutation you want to carry out, namely define the topology of the
species being transformed. The strategy proposed in NAMD is the
dual-topology paradigm, in which the topologies for state a and
state b are defined concurrently. Note that the latter never
interact with one another, but the interaction with their
environment is scaled as you go from state a to state b, via some
ordering parameter, ?:

V(x,?) = ? Vb(x) + (1-?) Va(x)

This implies that a proper exclusion list must be defined to avoid
interactions between the alternative topologies. In practice, you
can set this list up in your Charmm topology file and then run
Charmm, if this is your favorite package for preparing the PSF files
for NAMD. Another route consists in using VMD and PSFGen, which will
build an "erroneous" PSF file, containing all possible interactions,
and subsequently correct for this by means of a little program,
alchemify, available on the ks.uiuc.edu website, which will clean
up the PSF file and remove extraneous interactions and internal
degrees of freedom.

Chris Chipot

s.bora erdemli wrote:

> Hi all;
>
> I am trying to use FEP mtehod with NAMD. I would like to calculate
> binding free energy change and maybe the binding constant.
> I wonder if there is anybody who tried it before. I need help about
> the usage of FEP and the principles of FEP calculation. the tutorial
> on the web is not easy to understand without basic knowledge about
> FEP. I did not have access to the some references behind this
> tutorial. Therefore, as I said before I need help....
>
> thank you very much in advance
>
> sincerely
>
>
> Sabri Bora Erdemli
> Research and Teaching Asistant
> KOC UNIVERSITY
> Computational Sciences and Engineering
> Koc Universitesi pk.218 34550
> sariyer/Istanbul TURKEY
> tel no: 02123381736
> 05326512523

_______________________________________________________________________

Chris Chipot, Ph.D.
Equipe de dynamique des assemblages membranaires
Unité mixte de recherche CNRS/UHP No 7565
Université Henri Poincaré - Nancy 1 Phone: (33) 3-83-68-40-97
B.P. 239 Fax: (33) 3-83-68-43-87
54506 Vandoeuvre-lès-Nancy Cedex

                            E-mail: Christophe.Chipot_at_edam.uhp-nancy.fr
                                           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 - 05:18:52 CST