When you describe the thermal motion of an atom as a "ball" it moves in
three dimensions. So you can define three axis at right angles to each
other and relate the motion to three directions. If it moves about the
centre point equally in all directions then the motion is isotropic and
this describes a sphere. Any section through this sphere is the same
circle so you can equally use an area from any circle cut through the
middle of that sphere. Hence a square value describing a three
dimensional object.
However if the atom moves more in any direction the overall shape is now
an ovoid and the motion is anisotropic. You then need to know all three
directions to describe the thermal motion. In early crystallography or
with low resolution data we use isotropic temperature factors as an
approximation to the real situation. Clearly when you are dealing with a
string of atoms they must by definition make bonds to each other and so
be anisotropic in their thermal motion. This becomes important when you
wish to refine structures to improve them and also to understand atomic
motions that may be critical to the biological function.
I hope this helps clear up the point and I will post to the general list
for information.
Best regards,
Jim.
Bernard M. Debono wrote:
> Dear Jim,
>
> I am currently going through a protein X-ray structure analysis
> article.
> It refers to 'isotropic B-values' which have units angstrom^2. I
> reckon
> this is an area of sorts...perhaps a surface area, as it has been
> applied to the context of side chains.
>
> Could you provide me with more information on this particular topic,
> especially the underlying physical principles involved in the
> measurement, and its relevance in describing a protein? Branden &
> Tooze,
> Creighton, and Schulz & Schirmer were not helpful sources at all.
>
> Regards,
>
> Bernard Debono
> PPS'97