Clipper already uses quaternions internally - I suspect one is already
used as an intermediate step when converting between Euler and Polar
conventions. So it *should* be easy to output this (but it is not
assured - the code in question might be part of ssm, which cannot see
clipper).
So I can look into this. Is it useful to print out the quaternion, or is
this information only useful if it can be displayed graphically?
I guess we should be able to put together a clever graphic which shows
the rotation axis, the angle of rotation, and the screw translation all
at once.
Kevin
Ethan A Merritt wrote:
SSM superposition is a great tool, but I wonder if Coot could not
report a few additional useful quantities that result from it.
In particular, I would like it to report both before + after RMSD
values for the superposition. Also, it would be nice if it would
report the rotation as a quaternion in addition to the Euler and
omega/phi/kappa conventions.
For example, suppose that I want to compare two dimeric structures.
Let's say one of them has component monomers A and A'.
The other one has component monomers B and B'.
For simplicity, let's say that really these are the same dimeric
protein but A:A' is the apo form and B:B' is a ligand-bound form.
I can easily use SSM to superimpose B onto A, and B' gets
carried along for the ride. So then I look at this and notice
that B' is not perfectly superimposed onto A'. I think
to myself, "Aha, the intermonomer angle between A:A' is obviously
different than the angle between B:B'. I wonder by how much?"
I would like to be able to report the difference between these two
structures by saying something like:
"Upon ligand binding the second monomer (A') of the dimer rotates by
XXX degrees relative to the first monomer (A). This conformational
change corresponds to an RMSD displacement of YYY for the
CA atoms in monomer A'"
Both of these quantities XXX and YYY should fall out directly from a
second run of SSM that starts from the B-onto-A model and asks
for a superposition of B' onto A'. But unfortunately Coot doesn't
report them (or at least not that I can find), so I am reduced to
exporting my various superpositions and Euler angles into other
tools to derive 2 simple numbers.
For bonus points, Coot could actually display the axis of the
quaternion on the screen. In a favorable case one might immediately
learn that the observed rotation was about a particular hinge point
or other structural element. (I've switched here to an example
use that involved inter-domain rotation rather than inter-monomer
rotation).
cheers,
Ethan
