In the present case, the analysis of the filtering procedure is
straightforward: a single peak is produced, which corresponds
to the orientation of one of the molecules. The second molecule
is not found in this particular run, since its orientation was
not among the first 240 peaks of the rotation function. This shows
a problem with the **PC**-refinement strategy: if no solution is found,
one cannot be sure that there is no solution. A ``direct"
rotation search
(see Section 17.7) may help in these cases.
Another aspect of the
result of the filtering procedure is that sometimes several **PC**-refinements
converge on the same solution. The Rotman statement provides a
facility to measure the ``distance" or metric between two rotation
matrices by removing crystallographic redundancies. Suppose
one wants to know the distance between
two orientations of the search model
(=80.477, =85.000, =24.806) and
(=261.392, =90.000, = 336.243)
taking into account the crystallographic
symmetry. Using the Rotman statement

xrefin a=44.144 b=164.69 c=70.17 alpha=90. beta=108.50 gamma=90. symmetry ( x, y, z ) symmetry ( -x, y+1/2, -z ) end rotman euler=(80.477 85.000 24.806) swap euler=( 261.392 90.000 336.243) dist endone finds that the difference between these two orientations is about 5, due to the crystallographic symmetry.

Sat Mar 11 09:37:37 PST 1995