I wrote a CCP4 script for this:
http://bl831.als.lbl.gov/~jamesh/mlfsom/ano_sfall.com
The script is actually based on a CCP4BB reply from Elanor I found in
the archives. It does Elanor's jigery-hokery, and a few other things.
The script is meant to be the front-end for a simulator so I did a
number of things to make the output "data" more "realistic". You run it
like this:
ano_sfall.com model.pdb sites.pdb P212121 2A fpp=4 fp=-8 sigma=1
or
ano_sfall.com model.pdb 2A wave=0.98
where:
model.pdb is a brookhaven file of the protein structure
optional:
sites.pdb is a second brookhaven file containing position/occupancy/B of
anomalous scatterers
P212121 is the desired space group
2A render to a desired resolution
wave=0.98 calculate the anomalous f' and f'' contribution at 0.98 A
energy=12680 calculate the anomalous f' and f" contrubution at 12680 eV
fpp=4 sets the anomalous f'' contribution to 4 electrons
fp=-8 sets the dispersive f' contribution to -8 electrons
grid=0.3 use a particular grid spacing in sfall
sigma=1 gives a value of 1 to the sigma of each column
nohydrogen do not add hydrogens first
What the script does:
1) add hydrogens with "hadd"
yeah, you can't "see" them at 3A, but i.e. a methylene carbon has 8
electrons, not 6.
2) calculate f' and f" for each atom type in the PDB at the wavelength
provided using "crossec"
unless there is a user-defined f' and f" on the command line
3) calculate a solvent mask with "sfall"
4) smooth the solvent mask with "mapman" (if it is available)
this avoids series-termination errors from the "sharp" solvent edge
5) convert everything into HKL and F using "sfall"
6) apply the f' and f" to the contribution of each atom type using "sftools"
(this is the jigery-hokery)
7) add up all the (phased) partial structure factors using "sftools"
8) output FP PHIP FH PHIH FPH DANO Fplus Fminus Fmean (with "sigmas")
into "ideal.mtz"
I add a fake "sigma" column for each "data" column because so many
programs require one.
FP PHIP is the contribution of the protein (and solvent)
FH PHIH is the vector sum of all the f' and f" contributions
FPH is the vector sum of FP with all the f' contributions
Fplus is FPH + all the f" contributions
Fminus is FPH - all the f" contributions
Fmean is the mean of Fplus and Fminus
DANO is the difference between Fplus and Fminus
Comments and questions are always welcome.
DISCLAIMER:
I do not consider this script ready for "prime time" yet. I'm sure
there are quite a few entries in the PDB that will break it. If it
works for you, great. If it doesn't work for you, sorry. I'm still
working on it. If you use it to publish some fake results, then this is
your fault, not mine. If it does something "wrong" that makes you have
to retract your structure then this is also your fault, not mine. Why
did you trust me anyway?
END DISCLAIMER
-James Holton
MAD Scientist
Santosh Panjikar wrote:
Hi all,
Does anybody have a program which can calculate anomalous differences or
F+ and F- from a refined structure at given wavelength and resolution ?
Thanks
Santosh
Santosh Panjikar, Ph.D. [EMAIL PROTECTED]
Staff Scientist
EMBL Hamburg outstation http://www.embl-hamburg.de/~panjikar/
DESY, Notkestrasse 85 Tel.:+49-40-89902-141
22603, Hamburg, Germany Fax: +49-40-89902-149
