There is a very useful program written by Phil Evans called othercell
which gives alternate indexing
( The same functionality is in pointless)
In this case it suggests:
Alternative indexing schemes which lead to identical or similar
cells are grouped on continuation lines if they are symmetry-related
Laue group a b c alpha beta gamma Deviation ReindexOperator
I m m m 58.1 76.7 95.8 90.0 90.0 90.2 0.24 [-l,h+l,-k]
Possible spacegroups:
<I 2 2 2> <I 21 21 21>
So if the data is not orthorhombic then you probably have one othogonal
sym op as a twinning operator.
If you use the latest refmac to do twinned refinement it will select
which of the operator is most likely to be a twinning one.
Eleanor
Jeffrey D Brodin wrote:
Dear ccp4bb,
I've recently run into a problem refining a crystal structure that I indexed
and performed a molecular replacement on as C2. The unit cell dimensions are
a=96, b=95.8, c=58.1 and beta=127; Rmrg=.195. The data go down to a resolution
of 2.8 angstroms. The data also index in I222, but scaling statistics are
worse (Rmrg = .598) and Molrep does not find an acceptable solution. Using the
C2 space group Molrep finds a solution with the expected 4 monomers (protein
size is ~12.5kd) in the asymmetric unit, but I have only been able to refine it
to an R/Rfree of 33/37. The asymmetric unit consists of two dimers, and
performing a two-fold rotation around an axis perpendicular to the helices of
the dimers generates the tetramer that we would expect from other experimental
data. Also, the backbone density for two of the monomers is much worse than
normal. Truncate shows possible twinning in that the 4th centric moment of E
has an observed value of 2, which is what would be expected for a perfect twin
(actual expected value is 3) and the first and third centric moments are .87
and 1.45 respectively, with expected values of .8 and 1.6 and .89 and 1.33 for
perfect twins. I am new to resolving twinning issues and was wondering if
anyone has any ideas as to whether I may have the space group wrong or if there
is a twinning operator that I could use to resolve this issue. Thanks in
advance for your help,
Jeff
