# RE: NAMD-L: question about harmonic constraints

From: JC Gumbart (gumbart_at_ks.uiuc.edu)
Date: Tue Apr 11 2006 - 23:43:40 CDT

That's a good question.

If you look at the formula (see
http://www.ks.uiuc.edu/Research/namd/2.6b1/ug/node32.html), with v = 0, you
get the potential to be U = 1/2k[(R-R0) dot n]^2. Ignoring R0 (assume it's
zero), this becomes U = 1/2k[Rx*nx+Ry*ny+Rz*nz]^2. So, if your constrained
atoms drift along a direction such that the sum is zero, you will have no
applied force! This is avoided though by the vector n only having one
component.

I actually figured this out after running into my own problem with using SMD
to constrain something. I was initially shocked to see atoms I thought were
constrained moving very freely, but now I understand why, and how to get
around it.

-----Original Message-----
From: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] On Behalf
Of Margaret Shun Cheung
Sent: Tuesday, April 11, 2006 9:07 PM
To: JC Gumbart
Cc: namd-l_at_ks.uiuc.edu
Subject: Re: namd-l: NAMD-L: question about harmonic constraints

Dear JC and NAMD users,

Thank you very much for sharing this clever script. I guess to comply with
the pulling axis, an rotation of the long axis of the aspherical molecule
to align with the pulling axis may be the first step to do. (I sent VMD-L
an email and asked if there are tcl scripts for rotations).

But I am intrigued by the message in which the pulling axis on the
SMDgroup to be only one of
x-, y-, or z- (e.g. (0,0,\$flz) ):
it looks like if \$fx and \$fy can be treated
like \$f1z, then the directions of pulling forces for SMDgroup can be
(\$f1x,\$f1y,\$flz).

Or is it because of some other reasons that directions of forces coupled
with SMD module should be one of the x-,y-, or z- axis?

Thank you very much!
Sincerely,
Margaret

On Tue, 11 Apr 2006, JC Gumbart wrote:

> The solution to your problem is to use SMD for one part and tcl forces for

> the other.
>
> For example, you can use SMD to restrain one part of the protein with zero

> velocity along the axis of pulling (please remember, it can only be one
axis,
> x, y or z) and then use tcl forces to pull the other part at constant
> velocity.
>
> I've attached a script I used to pull a small deca-alanine helix in the -z

> direction at constant velocity. I pulled the center of mass of the 10
> C-alpha atoms. This along with the NAMD user guide section on tclforces
> should give you a good start in using it for your purposes.
>
>

Margaret S. Cheung,
Ph. D. Postdoctoral Fellow,

------------------------------------------------------------
Room 2117 Tel: (301) 405-4892
Inst. for Phys. Sci. & Tech. Fax: (301) 314-9404
University of Maryland http://glue.umd.edu/~mscheung
College Park, MD 20742-2431 email: mscheung_at_ipst.umd.edu
------------------------------------------------------------

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:41:54 CST