Re: [NAMD] Using ABF to study separation of two molecules

From: Giacomo Fiorin (
Date: Mon Jan 23 2017 - 08:43:26 CST

Hi Sebastian, ABF, metadynamics and umbrella sampling are all methods that
allow you to compute one- and multi-dimensional PMFs. They rely on
different free energy estimators (ABF measures the internal, or "system"
forces acting on the variables, metadynamics targets a flat probability
distribution so that the PMF is calculated directly from the biasing
potential, umbrella sampling does the same but with a Gaussian distribution
instead of a flat one).

You can use any free energy estimator on an equilibrium MD simulation
(internal forces or probability distribution), but you won't see much
beyond the local minimum you're in. This is the reason why most free
energy calculation methods use a biasing force or potential. The effects
of this bias are automatically removed from the free energy estimate.

To use methods based on collective variables, you need to decide which
variables are accurate descriptors of your system. You may use similarity
with a system previously published, but a general rule for the choice of
collective variables simply does not exist.

The choice of variable / reaction coordinate is much more important than
the method you will use on it.


On Mon, Jan 23, 2017 at 2:00 AM, The Cromicus Productions <> wrote:

> Hi everyone,
> I'm new to free energy calculations using NAMD but currently I'm studying
> how two molecules separate.
> To do so, in simulations I untwist and separate them (under normal
> conditions they separate in milliseconds).
> If I want to obtain the free energy through the reaction coordinate, which
> is the most convenient thing to use,
> ABF, metadynamics, umbrella sampling?
> Also, if I want to use for example ABF, do I need to run it during a
> simulation where I do this mechanism or do
> I just have to sample during equilibrium steps of the molecules? If I
> sample during equilibrium for a few nanoseconds
> It doesn't seem natural to be able to get to the barrier, but If I
> simulate during my mechanism aren't I putting external
> Forces that will modify the real free energy?
> Thank you very much,
> Sebastian

Giacomo Fiorin
Associate Professor of Research
Institute for Computational Molecular Science (ICMS)
College of Science and Technology, Temple University
1925 North 12th Street (035-07), Room 704D
Philadelphia, PA 19122-1801
Phone: +1-215-204-4213
Lab page:
*"As computer programmers we have a responsibility to make sure that we run
the computers instead of the computers running us."*  - Steve Oualline

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:01 CST