Nick, there was a discussion of this three weeks ago. Check this thread
http://www.mail-archive.com/ccp4bb@jiscmail.ac.uk/msg14133.html I still maintain the view that appropriately tight restraints are the way to go and not the grouped B-factor refinement (at least not the way it is currently implemented). As for weight optimization, you should probably stick with traditional "minimize-the-R-free" approach. Or take some high resolution structures refined with refmac and check in the "ISOTROPIC THERMAL FACTOR RESTRAINTS" section to see what RMS is expected and then choose weights to match it. Cheers, Ed. On Wed, 2010-02-17 at 09:34 +0000, Nicholas Keep wrote: > We have been solving a protein complex at 3.1 A by molecular replacement. One > copy of two chains in the asymmetric unit. > The two chains have fairly different average B factors. As would be expected > isotropic B factors in refmac overfits the > data and leaves a large gap between R and Rfree. TLS refinement with overall > B factor gave a more reasonable separation > of R and Rfree and more importantly more interpretable difference maps. > Phenix. refine with TLS and group B factors > bought the Rfree down further. Refmac does not seem to have a group B factor > option but I presume that by setting tight > limits on the B factor constraints (I tried 0.1 on all of them (at most 1/10 > of the defaults)- HAS ANYONE DONE ANY > SYSTEMATIC WORK ON WHAT THIS SHOULD BE?) I think you effectively achieve two > B factors per residue- main and side. Is > this correct? > > Does anyone else have comments on Bfactor use at low resolution in refmac? > Best wishes > Nick > -- Edwin Pozharski, PhD, Assistant Professor University of Maryland, Baltimore ---------------------------------------------- When the Way is forgotten duty and justice appear; Then knowledge and wisdom are born along with hypocrisy. When harmonious relationships dissolve then respect and devotion arise; When a nation falls to chaos then loyalty and patriotism are born. ------------------------------ / Lao Tse /
