Re: Re: How to set up a window distance of 2 angstroms in Umbrella sampling

From: Giacomo Fiorin (
Date: Mon Mar 05 2018 - 07:38:59 CST

Hi Lutimba, this tutorial covers a few of the features of NAMD and in
particular the Colvars module, but not all of them.

One useful feature is embedding some of the configuration for the
collective variables and/or restraints in the NAMD input, which is a Tcl
script as you know:

The example NAMD input would be:
colvars on
colvarsConfig ; # *Defines just the variable*
set us_center *$env(US_CENTER)*
cv config "
harmonic { # *Defines the restraint*
  name us
  colvars <name of your variable>
  forceConstant <your K>
  centers *${us_center}*

The trick is defining the bias via the cv config <...> command in the NAMD
main script: this command will accept a string, which you can make as long
as you want (e.g. multiple lines) and contain the values of Tcl variables.
In this example, the Tcl variable us_center is a copy of the environment
variable US_CENTER, which you can set in your bash script before calling
NAMD, so that the above configuration will be valid for *all* replicas.
(Tip: if you use Slurm, look into the job array options).

As for generating the initial configurations, it is usually assumed that
you take snapshots from a prior simulation (e.g. SMD), and manually assign
them to be the starting points for each replica. But for something like
small-molecule/protein binding this usually takes more of the scientist's
time than the computer's time. My advice would be generating these initial
snapshots automatically, for example turning the restraint on gradually by
setting forceConstant to zero and targetForceConstant to the value needed
for umbrella sampling:
This way, no matter how far is your starting point from the umbrella
center, you can gently drive it there.


On Mon, Mar 5, 2018 at 4:40 AM, Lutimba Stuart <
> wrote:

> Dear NAMD user
> Am running umbrella sampling after obtaining a reaction coordinate from
> Targeted molecular dynamics that applied a harmonic force on the sugar
> driving it to its active site. I would like to run umbrella sampling with a
> window distance of 2 angstroms but am not sure of how to set this. AM
> following the Umbrella sampling tutorial but it does not describe this. I
> am very great full for your help please. Thank you very much.

Giacomo Fiorin
Associate Professor of Research, Temple University, Philadelphia, PA
Contractor, National Institutes of Health, Bethesda, MD

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:53 CST