Re: namd: ABF query

From: Jerome Henin (
Date: Thu Mar 20 2008 - 11:29:30 CDT

As you say yourself, your output indicates that the ABF coordinate has
been sampled only partially. As a rule, it is difficult to predict how
long it will take for a given range of a coordinate to be sampled -
maybe you just need to run a bit longer, maybe there are trickier
issues involved.
One common cause for slow sampling with ABF is the presence of slow
degrees of freedom that are not your chosen coordinate, which may
cause the system to become "trapped". Often, a good way to diagnose
this is to rely on your intuitive understanding of the system.

In the ABF dist file, the abscissa is the value of the instantaneous
force component, and the ordinate is the statistical distribution (or
probability of occurrence) of force values.

On Thu, Mar 20, 2008 at 6:12 AM, <> wrote:
> Dear Namd users
> I am seeking your help regarding free energy calculation in Namd using
> ABF. In the abf output .dat file, the sampling seems to stop after a
> certain point, the force becomes zero and the A(x) value remains constant
> beyond that point. Could you explain what is causing this. Has it got
> something to do with dxi? I've included the config file as well as the
> ABF output file.
> Also, i am unable to understand how to interpret abf dist file. i dont
> know what is the column next to instantaneous force for each bin between
> xiMin and xiMax.
> Best
> Gurvisha
> Gurvisha Sehgal
> LeadInvent Technologies
> Technology Business Incubation Unit(TBIU)
> IIT Delhi
> Hauz Khas, New Delhi 110 016

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:47:39 CST