MULITSCALE manual; anisotropic scaling + solvent

Yoram Puius puius at aecom.yu.edu
Thu Feb 27 11:07:38 EST 1997


Hi all,

Does anyone out there (perhaps someone from Dr. Brunger's group?)
have the manual for the MULTISCALE keyword in X-PLOR 3.851?

Perhaps there's already a script out there to do what I'm interested
in, which is this:

I have data that is fairly anisotropic (I brought down my Rfree by
4% by applying overall anisotropic B-factors!) which seems to
refine best with a solvent model.  I am quite aware that one of
my options is do drop the solvent model, but it seems to help quite
a bit -- when I apply both, I get very reasonable R-factors.

The problem is, each run I get different refinements for the
anisotropic Bs and the solvent, which can result in huge isotropic B
all around (even through the R-factors stay low).  I've been trying to
use multiscale to simultaneously fit an ISOTROPIC solvent model and
and ANISOTROPIC Fobs scale;  this is apparently impossible.  And I
don't want to fit an anisotropic solvent model!

So, my questions are:
1)  DOes anyone have a manual for the MULTISCALE keyword?
2)  Does anyone have a tested routine that fits a solvent model
and anisotropic B-factors?
3)  Am I nuts in refining the Bij factors each time?  Should I
only do it once?  (I have maybe 5-10% more atoms to add to the model,
and was assuming that it will affect the numbers)

Now, I realize that whatever isotropic B-factors I refine (this is
a 1.9A structure, so it's justified) would be on top of the
anisotropic ones. So if my B11,B22,B33 are, say 5, 10, and 20, I
can absorb some of that into my isotropic factors to get something
close to 0, 5, and 15 (I _know_ it doesn't work that way, but you
get the idea).

Which brings me to my last question:

4)  Can I tweak my overall B-factors so that my smallest Bij is
close to zero, and then keep those Bij coefficients?

Thanks in advance to any replies to this wandering and wordy
request.

	- Yoram
-- 
_______________________________________________________________________
Yoram A. Puius                Albert Einstein College of Medicine
mailto:puius at aecom.yu.edu     Department of Biochemistry
5th year M.D.-Ph.D.           1300 Morris Park Avenue, Bronx, NY  10461
_______________________________________________________________________



More information about the X-plor mailing list