Current version does not expand data or model. C222 twinned in P2 will
not be handled if data are in C222.
Overmerged data are a bit tricky to deal with. And model in higher
space group would not fit into the asymmetric unit in general.
In short: in the current version space group should be actual space
group and model should correspond to it. We are working with on a
procedure to deal with overmerged data. Meanwhile ZANUDA from YSBL
webserver
www.ysbl.york.ac.uk/YSBLPrograms/index.jsp
can be used to sort out space group if you have accidentally
overmerged and managed to solve structure at higher space group
regards
Garib
On 12 Mar 2009, at 20:49, Frank von Delft wrote:
Hi Pavol, Garib
Actually, that bit of doc doesn't answer a few practical questions
we keep running into:
1) What spacegroup should you provide the data in? The usual
scenario, in practice, is that one merges in too high symmetry, and
wants to check whether it is actually lower symmetry with twinning.
2) What spacegroup should the *model* be? The highest symmetry?
3) What exactly is refmac doing: is it expanding the model to the
lower symmetries starting from your model?
4) Will it also test for the "funny" ones (e.g. C222 which is
actually P2 twinned).
Cheers
phx.
Pavol wrote:
Dear Kim,
yes, Refmac handles pseudo-merohedral twinning. The procedure is
fully
automatic and there is not much to change but the documention can
be found
here: http://www.ccp4.ac.uk/html/refmac5/keywords/keywords_5_5.html#Twin
Pavol
