Re: Re: namd-l digest V1 #2196

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Fri Jun 09 2017 - 17:07:05 CDT

Hi Rabeta,

It depends on the design of the experiment. In replica exchange umbrella sampling, the biases swap with one another changes over time. If you use umbrella-sampling like ideas for steered molecular dynamics, then the umbrellas also move. You could, if you like bookkeeping, use wall for simple umbrella sampling where the umbrellas don't move. However, this is not the recommended way of doing things. Instead, you would use a harmonic section in your configuration file, since you can do things like output the energy directly, or use the same bias to setup your initial umbrellas via a steered umbrella sampling simulation.

-Josh

On 06/09/2017 02:48 PM, Rabeta Yeasmin wrote:
Dear Josh,

This is my first time using Umbrella Sampling and I was unable to catch all of your information. I apologize for that. I am wondering how can I know my umbrella will never move before finishing the analysis.
Thanks.

Rabeta Yeasmin

On Fri, Jun 9, 2017 at 1:07 AM, namd-l digest <owner-namd-l-digest_at_ks.uiuc.edu<mailto:owner-namd-l-digest_at_ks.uiuc.edu>> wrote:

namd-l digest Friday, June 9 2017 Volume 01 : Number 2196

In this issue:

    Re: namd-l: Umbrella Sampling colvar file parameters
    namd-l: Colvars wall boundaries for CoordNum
    Re: namd-l: Colvars wall boundaries for CoordNum

----------------------------------------------------------------------

Date: Thu, 8 Jun 2017 17:32:44 +0000
From: "Vermaas, Joshua" <Joshua.Vermaas_at_nrel.gov<mailto:Joshua.Vermaas_at_nrel.gov>>
Subject: Re: namd-l: Umbrella Sampling colvar file parameters

Hi Rabeta,

They are not required to be set. Pay attention to the defaults in the
documentation, since those are what is used if nothing is set. Width
defaults to 1, which is just a scaling factor on the force constant you
are using. 1 is usually fine, except when you are combining two
different colvars into 1 reaction coordinate, and they are unequal. For
umbrella sampling, boundary and wall constants are usually NOT set,
since the harmonic section defines your umbrella. However, if you know
your umbrella never moves, I suppose you could use upper and lower walls
to define your potential instead of a single harmonic.

- -Josh

On 06/08/2017 11:26 AM, Rabeta Yeasmin wrote:
> Hi,
>
> I am wondering if defining the width, boundary and wall constant in
> the umbrella Sampling colvar file is a must. If so, how should I
> define them means what parameters should I consider while defining
> them. If these are not a must, Will using these parameters give any
> extra benefit?
>
> Thanks.
>
> Rabeta Yeasmin

------------------------------

Date: Fri, 9 Jun 2017 00:27:50 +0000
From: Wasut Pornpatcharapong <wpornpat_at_mail.ucsd.edu<mailto:wpornpat_at_mail.ucsd.edu>>
Subject: namd-l: Colvars wall boundaries for CoordNum

Hi, I wonder if the actual force constant for CoordNumb s wall boundary is obtained from the b ForceConstantb in the colvars configuration file divided by the square of the width defined in colvars module for unbiasing usage?

- -Wasut

Sent from Mail<https://go.microsoft.com/fwlink/?LinkId=550986e7c080%7C1%7C0%7C636326379651084654&sdata=qC3B9fBOmQiFJzsfjn41IXXytgXYdsOvpWb8NEYbbAE%3D&reserved=0>> for Windows 10

------------------------------

Date: Thu, 8 Jun 2017 20:42:46 -0400
From: Giacomo Fiorin <
giacomo.fiorin_at_gmail.com<mailto:giacomo.fiorin_at_gmail.com>>
Subject: Re: namd-l: Colvars wall boundaries for CoordNum

Hi Wasut, the wall constant for a wall boundary is not rescaled (check the
description of the physical unit):
http://www.ks.uiuc.edu/Research/namd/2.12/ug/node55.html#SEC8ba7%7Ca0f29d7e28cd4f5484427885aee7c080%7C1%7C0%7C636326379651084654&sdata=qdIQiJX5mo7IAgUupB%2FBYZLjwbs5YzMOEhxr3H0A%2F74%3D&reserved=0>
TION000132200000000000000

The rescaling by width^2 is only needed because a harmonic bias can be
defined on multiple variables, with very different units.

Note, however, than in the latest versions you have the ability to define
walls as a separate bias (and also change its force constant umbrella
sampling-style). The current syntax that you are referring to will remain
for backward compatibility.

Giacomo

On Thu, Jun 8, 2017 at 8:27 PM, Wasut Pornpatcharapong <
wpornpat_at_mail.ucsd.edu<mailto:wpornpat_at_mail.ucsd.edu>> wrote:

> Hi, I wonder if the actual force constant for CoordNumb s wall boundary is
> obtained from the b ForceConstantb in the colvars configuration file
divided
> by the square of the width defined in colvars module for unbiasing usage?
>
>
>
> -Wasut
>
>
>
> Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986e7c080%7C1%7C0%7C636326379651084654&sdata=qC3B9fBOmQiFJzsfjn41IXXytgXYdsOvpWb8NEYbbAE%3D&reserved=0>> for
> Windows 10
>
>
>

- --
Giacomo Fiorin
Associate Professor of Research, Temple University, Philadelphia, PA
Contractor, National Institutes of Health, Bethesda, MD
http://goo.gl/Q3TBQUWLGlDuNv8q%2FUM3ETIzJB7LK93OloeI%3D&reserved=0>
https://github.com/giacomofiorin4&sdata=FGKYkXJn%2FlmG9d3LhiqVMHspAwPdNNrZncYLAxnxGwg%3D&reserved=0>

------------------------------

End of namd-l digest V1 #2196
*****************************

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:20:21 CST