From: Aron Broom (broomsday_at_gmail.com)
Date: Thu May 16 2013 - 07:55:46 CDT
I've always just used a series of simulation "windows" with their own
harmonic restraint and not the "targetNumStages" method, but it looks like
each "window" is essentially only 100,000 steps or ~200 ps? If this is
true, it's possibly too short, and your just seeing the classic problem of
trying to analyze data that isn't equilibrated. Maybe up the window length
to 10 ns per window to get really nice data (obviously just a guess).
Having said the above, the PMFs do look reasonably smooth, so maybe
equilibration isn't the issue.
Maybe run a couple of classical harmonic restraint windows and see if the
trajectories match?
On Thu, May 16, 2013 at 3:06 AM, 서용일 <seoyi86_at_gmail.com> wrote:
> Dear NAMD users
>
> We perform MD simultaions for potential of mean forece calculation using
> colvars module in NAMD_2.9_Linux-x86_64-multicore package.
>
> The system is one potassium ion in 2135 waters on NVT esemble (V = 40.0 x
> 40.0 x 40.0 A^3, T = 298.15 K).
>
> Colvars parameters are below:
>
> colvarsTrajFrequency 2
>
> colvar {
>
> name ProjectionXY_ion
>
> width 2.5
>
> distanceXY {
>
> main {
>
> atomNumbersRange { 1-1 }
>
> }
>
> ref {
>
> dummyAtom ( 0.0, 0.0, 0.0 )
>
> }
>
> axis (0.0, 0.0, 1.0)
>
> }
>
> }
>
> colvar {
> name ProjectionZ_ion
>
> width 0.4
>
> lowerboundary -15.0
> upperboundary 15.0
>
> lowerwallconstant 10.0
> upperwallconstant 10.0
>
> distanceZ {
> main {
> atomNumbersRange { 1-1 }
> }
> ref {
> dummyAtom ( 0.0, 0.0, 0.0 )
> }
> axis (0.0, 0.0, 1.0)
> }
> }
>
> harmonic {
> colvars ProjectionXY_ion
> forceConstant 1.0
> centers 0.0
> }
>
> harmonic {
> colvars ProjectionZ_ion
> forceConstant 1.0
> centers -10.0
> targetCenters 10.0
> targetNumSteps 100000
> targetNumStages 50
> }
>
> Later simulations, i calculate potentail of mean forece using umbrella
> sampling with WHAM.
>
> We expect that the free energy is changed little.
>
> But, The defference of free enegy between the initial window and final
> window is 25.336992 to target centers from centers.
>
> So, I performed the same simulations using namd2.8_64-multicore with the
> same module and parameter sets.
>
> In this case, the difference of free energy is 0.095781.
>
> Link to the PMF profiles<http://4.bp.blogspot.com/-qJnJclWf6uw/UYmxTjftt6I/AAAAAAAAAAc/5cUXbWe1s6s/s1600/pmf.png>
>
> I'm wondering why there is some differences between NAMD 2.8 and 2.9?
> Do i need to add other options or paramters about this colvars in 2.9
> package?
>
> Yongil Seo
>
>
-- Aron Broom M.Sc PhD Student Department of Chemistry University of Waterloo
This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:12 CST
