I believe you have C2 with pseudo translation and that screws up the
merohedral twinning test. As you also write you know the translation
vector already. Most likely your crystal truly is in a higher symmetry,
but for example during freezing (and the high solvent content) your
multimers (two tetramers is what I expect) slightly move far enough
apart to not fulfill the requirements for the higher symmetry space group.
What I would try to do is the following:
- take your initial model from resolve, figure out the NCS operators in
C2 for both tetramers (if my assumption is correct) with two masks ! One
mask per tetramer, each mask covering only one subunit
- run DM defining the two sets of NCS operators with the correct mask,
use SCHEME RES starting from low resolution e.g. 4-5 A, also use the
AVER REFINE EVERY 3 or so cycles and for a total of perhaps 200 cycles.
For the first run don't let DM update your mask.
- check if your electron density map improved and if so rebuild the
model by hand/resolve
- revisit with updated mask models DM and perhaps also use automatic
update mask option again starting from 4-5 A
Something else you might want to look at is a plot where you calculate a
native patterson map at various resolutions e.g. 20 - 3 A in 1 A steps
check the peak height of your self patterson peak and draw a nice
diagram. On the y-axis the percentage peak height and on the x-axis the
resolution. What I expect to see in your case is a slope where the peak
height at low resolution is e.g. 70% dropping down to e.g. 30% at 3 A.
If you should get a resolution independent more or less constant height
of the peak, then the two tetramers are rotated close to 45 degree to
P.S. I would really be curious to see how far off I was :-)
University of Washington
Dept. of Biochemistry, K-426
1705 NE Pacific Street
Seattle, WA 98195