Re: questions regrading ABF

From: Chris Chipot (Christophe.Chipot_at_edam.uhp-nancy.fr)
Date: Mon May 29 2006 - 16:50:46 CDT

Jan,

a) what are the atoms corresponding to abf1 4 and abf2 117 ? Keep
    in mind that the RC should be decoupled from constrained DOFs.

b) note that the free energy minimum of the alpha-helix emerges
    around 14 angstroms. In your input file, xiMin is 18.

c) the role of dxi is important and should be chosen as a function
    of how rapidly the free energy is changing. 0.5 angstroms is in
    the case of deca-alanine probably too large. 0.1-0.2 angstroms
    is typically what we used. dSmooth should set accordingly.

If you respect the above, I anticipate your sampling to be markedly
more uniform, and your free energy profile closer to what is expected.

Chris Chipot

Jan Saam wrote:

> Dear fellow NAMDers,
>
> I'm seeking your help regarding a PMF calculation in NAMD using ABF. I
> encountered some problems when I tried to do the decalanin unfolding
> example from the users guide:
>
> I prepared a decalanin helix, minimized it and then started an ABF
> calculation with the parameters from the ABF section of the NAMD user
> guide.
> My understanding of the method is that a force along the direction of
> the reaction coordinate is exerted on the system. The strength of the
> force is chosen so that the effect of the (estimated) PMF
> preventing the system to travel along the reaction coordinate is
> compensated. Thus the uniform sampling over the entire reaction
> coordinate is achieved. Correct?
>
> However, no matter how long I run the simulations, only the shorter
> distances (18-23 A) were sampled. This behaviour is completely
> independent of the abf parameters like the bin size or the smoothing
> window size. The helix is immediately unfolded but remains in a globular
> state until the end.
> It was my understanding that I don't have to use the 'moveBoundary'
> option in this example since the adaptive force should take care for the
> sampling of the complete reaction coordinate.
> Nevertheless, only when I'm using 'moveBoundary' I get resonably uniform
> sampling.
>
> Is there anybody who can give me some advice or even send me some
> example input?
> (I attached the NAMD input file and the abf output file.)
>
> Thanks and best wishes from Berlin,
>
> Jan
>
>
> ------------------------------------------------------------------------
>
> # NAMD CONFIGURATION FILE FOR DECALANIN
>
> # initial config
> coordinates alanin_min.coor
>
> # output params
> outputname alanin_abf
> binaryoutput no
> outputEnergies 100
>
> # integrator params
> timestep 1.0
>
> # force field params
> structure alanin.psf
> paraTypeCharmm on
> parameters /projects/namd/toppar/par_all27_prot_lipid.inp
> exclude scaled1-4
> 1-4scaling 1.0
> switching on
> switchdist 8.0
> cutoff 12.0
> pairlistdist 13.5
> stepspercycle 20
>
> # trajectories
> DCDfile alanin_abf.dcd
> DCDfreq 100
>
> # Temperature control
> set temperature 310
> temperature $temperature; # initial temperature
>
> # Langevin Dynamics
> langevin on; # do langevin dynamics
> langevinDamping 5; # damping coefficient (gamma) of 5/ps
> langevinTemp $temperature; # bath temperature
> langevinHydrogen no; # don't couple langevin bath to hydrogens
> seed 12345
>
> source /usr/local/NAMD_2.6b1_Linux-i686/lib/init.tcl
> package require abf
>
> abf coordinate distance
>
> abf abf1 4
> abf abf2 117
>
> abf dxi 0.5
> abf xiMin 18.0
> abf xiMax 32.0
> abf outFile deca-alanine.dat
> abf fullSamples 500
> abf inFiles {}
> abf distFile deca-alanine.dist
> abf dSmooth 1.0
> abf applyBias yes
> abf historyFile deca-alanine.history
> run 5000000

-- 
_______________________________________________________________________
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 Vand½uvre-lès-Nancy Cedex
                             E-mail: Christophe.Chipot_at_edam.uhp-nancy.fr
                                            http://www.edam.uhp-nancy.fr
     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 - 15:42:06 CST