[gmx-users] Re: Re: Re: Re: Re: umbrella potential

[Stefan Hoorman]

>
> Stefan Hoorman wrote:
>
> > Ok Justin, here are both profile and histogram files. Please let me know
> > if you can't get access or something like this. I have never used this
> > protobucket before.
> > "
> http://i784.photobucket.com/albums/yy123/stefhoor/wham_stefhoor/profile.jpg
> "
> > "
> http://i784.photobucket.com/albums/yy123/stefhoor/wham_stefhoor/histogram.jpg
> "
>
> The histogram shows only one peak, indicating only one window, just beyond
> 1.0
> nm of COM separation.  Is this an error in creating the plot, or is this
> really
> the result?  Like I've said before, you should have many windows, with
> overlapping distributions in order to correctly extract PMF using WHAM.
>
> > And here are my pull code stuff that is inside my mdp files. The first
> > one (Pull Code 1) is the one I used to separate the structures, and the
> > following (Pull Code 2) is the one used to simulate each window.
>
> I hadn't noticed this before, and I don't know if it's meaningful, but it's
> worth looking into.  You're pulling in two dimensions.  I'm not sure the
> status
> of the WHAM implementation in g_wham, but it is more common in the
> literature to
> extract one-dimensional PMF, and this is the most commonly-used method.  I
> would
> suggest orienting your system such that you are pulling directly along a
> single
> axis, and running things again.
>
> -Justin
>
> > ; Pull Code 1
> > pull  =  umbrella
> > pull_geometry  =  distance
> > pull_dim  =  Y Y N
> > pull_nstxout  =  10
> > pull_nstfout  =  1
> > pull_ngroups  =  1
> > pull_group0  = Protein
> > pull_group1 = SLC
> > pull_vec1  =  1 1 0
> > pull_init1  =  0
> > pull_rate1  =  0.001
> > pull_k1  =  2000
> > pull_constr_tol  =  1e-06
> > pull_pbcatom0  =  0
> > pull_pbcatom1  =  0
> >
> > ; Pull Code 2
> > pull  =  umbrella
> > pull_start  =  yes
> > pull_geometry  =  distance
> > pull_dim  =  Y Y N
> > pull_nstxout  =  10
> > pull_nstfout  =  1
> > pull_ngroups  =  1
> > pull_group0  = Protein
> > pull_group1 = SLC
> > pull_vec1  =  1 1 0
> > pull_init1  =  0
> > pull_rate1  =  0
> > pull_k1  =  2000
> > pull_constr_tol  =  1e-06
> > pull_pbcatom0  =  0
> > pull_pbcatom1  =  0
> >
> > The rest of my mdp stuff is pretty standard so, to save space I didn't
> > post it, but if you think it is necessary I would be glad to post it as
> > well. Sorry about the distance, I inverted the 2.41 (wrote 2.14). My
> > separation is from starting position (close to 1nm) till 2.5. But since
> > the structures always move a little in the beginning of each window, I
> > guess the final maximum distance is 2.41.
> >
>  <http://www.gromacs.org/search>
>

Ok. The histogram is the actual result. As I said, my windows are all there,
all the reaction coordinates I mentioned before are there to be analysed and
in the correct order, but the result comes always the same way. This
histogram is the actual result.
I will try pulling in only one direction then. Should have the results in a
few days.
Thank you
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090927/f65f979c/attachment.html>