Re: Using ABF to explore the conformational space of a spin label attached to a membrane protein

From: Ajasja Ljubetič (
Date: Tue Jan 18 2011 - 08:40:56 CST

Dear Jrme,

Two questions on this:
> 1) in what frame of reference do you need these to be expressed? I
> assume it would be a local frame, based on nearby atoms? You'll
> probably need to enable rotateReference, and related options.

1)Currently I'm still cheating a bit. I aligned the helix along the z-axis
and the center of the cone along the y. I also restrained the backbone
atoms. So I use the Cartesian axis as a reference for now.

> 2) if polar coordinates are okay, wouldn't Cartesian coordinates work
> just as well? It would be a lot simpler...

2) Actually I have a feeling that the length of the vector would not be a
good colvar, as the length is defined in a narrow range. Ideally I would
just like to sample along theta and phi.
(The third colvar would be perhaps the rotation arund the S-S bond, which
has a very high energy barrier. Or I could just run two simulations with the
S-S dihedral locked in one or the other minima).

If I sampled along two projections (say z and x), do you think that it would
be possible to obtain the avarage length of the vector in that bin without
introducing another colvar? The average length in each (x,z) bin could
probably be obtained from the trajectory in post processing, If the DCD is
written frequently enough. Or would it be possible do do something like
#define the x z and r colvars
abf {
colvars x z
fullsamples ...

histogram {
colvars x z r

In this way I would still only have a 2D surface, but from the 3D histogram
I could compute the average r as a function of x and z projections.

Or perhaps just
analysis on
#define the r
colvar {
name r
runAve on
runAveLength $fullsamples
runAveStride $fullsamples
runAveOutputFile r
abf {
colvars x z
fullsamples ...

For the angle, it is just a matter of doing some math and finding the
> Jacobian term in a usable form. I never needed it badly enough to do
> it. If you feel like trying and send me your result, I will gladly
> implement it.
> Hmm here a little bit of my biochemical background shows through. Is this
the Jacobian for the 3N cartesian --> 3N generalized coordinates (like
or just for the colvar? And what are the q_1 to q_(3N-1)? So I'll have to do
some reading before I'll be able to derive the Jacobian.

Best regards,

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:56:32 CST