Thanks Herman, George and Leo.I will try the methods.Their suggestions are
summarized below.
Dear Hongnan--
On 25 Aug 2008, at 17:47, conancao wrote:Or does any program could refine the
the same ligand with 2 different orientations into a mixed electron density?
That's an easy
one....http://scripts.iucr.org/cgi-bin/paper?ba5073http://cns-online.org/v1.21/
(see the alternate.inp
file)http://www.phenix-online.org/documentation/refinement.htm#anch323 (see the
infos on conformation)
HTH
Kind regards.
-- Leo --
SHELXL has been able to refine such occupancies (e.g. as p and 1-p for the two
ligands, i.e. one extra parameter) for many years, and if I have understood
correctly last week's version of phenix_refine can do so too.
George
Dear Hongnan,What I usually do is to calculate an omitmap and calculate the
electron density value(s) of (an) atom(s) which do not overlap and convert that
to relative occupancies. The sum can of course not be more then one and all
atoms of each molecule have the same occupancy. I use my own program to
calculate the electron density value at atomimic positions, but I am pretty
sure the CCP4 or phenix or so will also have programs to do this. An even
easier alternative might be to scroll in coot the electron density up and down
and note point where the density just disappears for the atom(s) in question.
From my experience, there is no need to determine the occupancies with more
accuracy then 10%, e.g. occupancies like 0.1, 0.2, 0.3 etc. since they are
linked to the B-factors anyways at this resolution. Best regards,Herman
_________________________________________________________________
看MSN史诗巨片，票选人气角色，赢取PSP等诸多好礼！
http://im.msn.cn/
