How to calculate sigmaa weighting map?

[liuxz at PUBMS.PKU.EDU.CN]

Hi, 

I looked up in the old archive discussing sigmaa weighting map and 
noticed that Mr. Marcelo C. Sousa told that there were some errors in the
tutorial file $XPLOR/tutorial/xtal_maps/nfo-mfc_phicalc_map.inp of
Xplor3.851. the relative lines are:
  
      do (diff = combine($nn * fom * ampl(fobs) - $mm * dd * ampl(fcalc),
phase(fcalc)) )          ( acentric and sel=1 ) 
      do (diff = combine(fom * ampl(fobs), phase(fcalc)) ) ( centric and
sel=1 ) 

Mr. Marcelo C. Sousa suggested they should be corrected as:

do (deri1=combine(2 $nn fom ampl(fobs) - 2 $mm dd ampl(fcalc+fpart),
phase(fcalc+fpart))) ($low_res >= d >= $high_res and acentric and
ampl(fobs) > 0 ) 
do (deri1=combine($nn fom ampl(fobs) - $mm dd ampl(fcalc+fpart),
phase(fcalc+fpart))) ($low_res >= d >= $high_res and centric and
ampl(fobs) > 0 ) 

I also saw Mr. Eckhard Hofmann's article which analized this problem
and raised a question to the solution. But I did not see any answer to
his question. I would like to know which weighting scheme is correct.
Could any experts help me please?
  
When I use Xplor3.8's tutorial file to calculate Fo-Fc map, I got very
noisy map, even at 3 sigma level. But this file can give fairly good 
2Fo-Fc map. So I think its weighting scheme for centric reflections maybe 
wrong. 

Do you have any suggestion? think you very much.

-------------------------------------------------------
  Xiaozhou Liu
  Biophysics and Physiology Department
  College of Life Science
  Peking University
  Beijing 100871, P.R.China
                                             
  E-mail: liuxz at iris.lsc.pku.edu.cn
          liuxz at pubms.pku.edu.cn
  Telephone: +86 10 62751864
-------------------------------------------------------