Hi All:
I have a structure with two chains in P21212 with cell parameters 85, 123,
51A. Co-crystallization with a ligand results in a new crystal form, and
we collected a data set to 2.3A with cell parameters of 85, 123, and 203A.
There (h,k,4n) spots are all strong, but the remainder are not all weak.
The structure solves, naturally, with pseudo-translation along c with
dimers separated by 51A. Refinement is stuck at 40% with 8 chains. The
B-factors of chains A and B are near 2A^2, while the other 6 chains have
normal B-factors.
I assumed that there was non-merohedral twinning, with a combination of
the small 51A cell and larger 203A cell, in roughly a 2:1 ratio. If you
use full occupancies for atoms in chains A and B, and 0.5 for atoms in
chains C-H, all of the B-factors are more normal, since the overall scale
factor is more correct. If true, then the intensities of the (h,k,4n)
reflections should be more intense with the contribution from the smaller
51A cell, and, in principle, I tried to scale them back to arrive at an
"untwinned" 203A data set. There isn't a good way to estimate the twin
fraction if this is the correct model.
The electron density is quite good, and reveals features of the ligands
and co-factors without including them in the refinement model.
Any other suggestions on how to correct the intensities, or refine the
structure?
Thanks for your help in advance.
-Shahila
Shahila Mehboob, PhD
Research Assistant Professor
Center for Pharmaceutical Biotechnology
University of Illinois at Chicago
900 South Ashland Ave.
Chicago, IL 60607
Phone: 312-413-9304 (lab)
Email: [log in to unmask]
|