B-factor refinement

Jack Tanner chemjjt at SHOWME.MISSOURI.EDU
Thu Aug 21 10:16:26 EST 1997

Try including the low resolution data in the overall bfactor refinement and
use isotropic overall bfactor refinement.  Since the final B-shift is added
to the atomic B-factors make sure that none of your atomic bfactors end up
< 0.

In your case I might do isotropic overall bfactor refinement at 20-2.3 A
followed by isotropic atomic bfactor refinement at 20-2.3 A, picking
waters, and finally bulk solvent mask correction at 20-2.3 A with the
waters included in the protein mask.  You could then do another cycle of
water picking and solvent mask correction.

Also, you might try optimizing the value of the probe radius in bulk
solvent mask using rfree.  It might optimize to less than 1.4 A (Jiang &
Brunger, JMB 1994, 243, p.100).

>Dear X-PLOR users,
>	I am in the late stages of refinement of a 2.3 angstrom structure
>and I have been experimenting with overall anisotropic b refinement as
>well as bulk solvent correction.  The structure is essentially done and
>I am picking waters.  After running baoverall.inp from 6.0 to 2.3
>angstroms, the Rfree drops by 2% to 34%.  Then bulk solvent correction
>(20 to 2.3 angstroms) further drops the Rfree to 33%.   If I follow this
>with individual B factor refinement, the Rfree drops to 31.5%, but the B
>factors for the model all uniformly increase.  This new model also produces
>poor maps, making it more difficult to pick waters.
>If anyone has any thoughts on this I would appreciate it.
>Christelle Sabatier
>University of California, Berkeley
>Department of Chemistry

Dr. John J. Tanner
University of Missouri - Columbia
Chemistry Dept.
123 Chemistry Bldg.
Columbia, MO 65211

phone:  573-884-1280
fax:    573-882-2754
chemjjt at showme.missouri.edu

More information about the X-plor mailing list