Re: using colvars for bending modulus?

From: Jérôme Hénin (
Date: Wed Jul 27 2011 - 03:12:33 CDT

Measuring the curvature as one angle does not mean that the shape will
be biased towards a triangle. All it means is, you won't be able to
resolve a triangle from other bent shapes - but if the chain has a
propensity to bend as a semi-circle, then that's what you'll get.
Unless the chain is much longer than the persistence length, I think
that would be fine.

On 27 July 2011 08:40, JC Gumbart <> wrote:
> Curvature is defined locally, but simple derivations for the bending modulus
> presume a constant curvature throughout the chain.  I'm not terribly worried
> about twisting motions, which ideally would be averaged over anyway.  The
> problem with defining just one angle is that the final structure won't be a
> semi-circle but rather a triangle of sorts.
> -----Original Message-----
> From: [] On Behalf Of Jérôme Hénin
> Sent: Wednesday, July 27, 2011 1:30 AM
> To: JC Gumbart
> Cc: namd-l list
> Subject: Re: namd-l: using colvars for bending modulus?
> Hi JC,
> I'm not very familiar with this, but isn't curvature a local property?
> Here you want a kind of mesoscopic measurement on the whole chain, so
> if the chain twists, it gets complicated. My best guess for a simple
> yet relevant quantity is the angle between three groups, two ends and
> the center of the chain.
> Cheers,
> Jerome
> On 26 July 2011 22:43, JC Gumbart <> wrote:
>> Has anyone considered how to use a collective variable to measure a PMF as
> a function of radius of curvature for a polymer?  I thought about something
> like a combination of angle terms between successive links, but it wouldn't
> necessarily stay planar.
>> Thanks!
>> JC

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:20:37 CST