RE: ABF question

From: Xu, Jiancong (xuj1_at_ornl.gov)
Date: Mon Jan 14 2008 - 11:19:27 CST

Jerome,

I am trying with smaller RC, for example, 1-2 A, and it seems the results are getting better. I'm also wondering how you prepare a well-equilibrated starting configuration for different ABF windows. Do you use SMD instead? Thanks.

Jiancong

-----Original Message-----
From: heninj_at_gmail.com on behalf of Jerome Henin
Sent: Mon 1/14/2008 12:02 PM
To: Xu, Jiancong
Cc: namd-l_at_ks.uiuc.edu
Subject: Re: namd-l: ABF question
 
Jiancong,
I'm afraid I cannot tell you much that has not been discussed on this
list before with other users, mostly to check for non-equilibrium
pulling effects and the influence of the fullSamples parameter. It
might be that your new starting configurations are not yet
well-equilibrated. It might be that there are other very slow degrees
of freedom in the system, interfering with your calculation.
Jerome

On Jan 14, 2008 11:14 AM, Xu, Jiancong <xuj1_at_ornl.gov> wrote:
>
> Hi Jerome and NAMD users,
>
> I started over with better configurations and it seems working now. However, the new problem is that the RC is not equally sampled. I've tried with different interval, initial configurations and bin width, and one of the borders is always oversampled than the rest. I would greatly appreciate any help and advice. Thanks much.
>
> Jiancong
>
>
>
>
>
> -----Original Message-----
> From: heninj_at_gmail.com on behalf of Jerome Henin
> Sent: Wed 1/9/2008 12:26 PM
> To: Xu, Jiancong
> Cc: namd-l_at_ks.uiuc.edu
> Subject: Re: namd-l: ABF question
>
>
> Hi Jiancong,
>
> It is best to prepare an equilibrated system with the RC in the
> desired range, using any kind of appropriate restraints. The boundary
> biases of ABF are harmonic - if you start far away from the min or max
> value, the excessive pulling on the system may have side-effects.
>
> An ABF tutorial is in preparation. We will announce it on this mailing
> list when it goes online - this should not take very long.
>
> Jerome
>
> On Jan 9, 2008 11:04 AM, Xu, Jiancong <xuj1_at_ornl.gov> wrote:
> >
> > Hi All,
> >
> > I'm performing an ABF calculations with the reaction coordinate spanning between 20 and 30 A defined by "distance-com". The reaction coordinate was split into 2 windows (1) 20 -25 A, (2) 25-30 A. From a short simulation, about 100ps long, no single data point fell into the 25-30 A range. Does this mean the default force constant applied at the borders of the reaction coordinate (10 kcal/mol/A2) is too small?
> >
> > I'm also wondering if there's any ABF tutorial available on NAMD website with all the input files for a model system. That'd be greatly helpful.
> >
> > Jiancong
> >
> >
>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:11 CST