Hi
Here is another way of scaling sigmas after
anisotropic bfactor scaling, which avoids having to
manually enter the Bii's. I get the scale factors
by comparing FOBS before and after scaling, using the following
csh script.
I would like to hear of people's experience with
anisotropic bfactor scaling. I wonder what one really
achieves, beyond what would be done during data processing.
Peter Metcalf
fix_boverall_sigmas:
#===================
set hkl_before_boverall = xxx.cv
set hkl_after_boverall = xxx_boverall.cv
set hklout = xxx_fix_boverall.cv
/bin/rm -f $hklout
set tempfile_1 = $$_1
set tempfile_2 = $$_2
paste $hkl_after_boverall $hkl_before_boverall | sed's/\.0* /\.0 /g' | sed 's/ */ /g' > $tempfile_1
grep IND $tempfile_1 | awk '{ print$1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11, $17 }' > $tempfile_2
# file tempfle_2 looks like this - i.e. it is hkl_after_boverall with the unscaled fobs appended
#INDE 16 4 0 FOBS= 242.035 0.0 SIGMA= 4.0 TEST= 0 228.200
#now replace sigma by sigma * bover_fobs/fobs i.e in above case 4.0 * 242.035/228.200 = 4.24
perl -ane 'print "@F[0..7] ",sprintf("%-8.4f", @F[8]*@F[5]/@F[11]) , " @F[9..10]", "\n";' $tempfile_2 > $hklout
/bin/rm -f $tempfile_1 $tempfile_2
exit
