Hi, I see this case as an example of Order-Disorder structures. This theory was developed by Dornburger-Schiff back in 1950s and very well documented. According to this theory, it is structurally possible to have P1 and C2 (in this case) and there should be a structural transformation between the two SPGs (hit: unit cell constants are fairly close as Artem already showed). Presence of NCS translations are a typical hint for this kind of structures. Actualy, these strcutures are called polysynthetic twins in the strcutural chemistry. Also, it may be possible to have other space groups as well depending on the arrangement of modular layers of protein molecules, but mostly, the lowest energetic arrangements are ruled out. I think in this case P1 and C2 structural arrangements are the lowest energetically favourable arrangements (they are called MDOs, Maximum Degree of Order strcutures). One way to check this is to look for diffuse scattering in raw data frames. If you don't see DS, that means that these structures are MDOs, otherwise non-MDOs. Now the question be to answered is, what you want from this. If you are interested in just the protein structure, forget about packing (because, the modular layer in both cases should be the same). If you want to establish a relation between the two SPGs, then bit more study would be needed.
Best, Rangana On Mon, Oct 31, 2016 at 7:20 AM, Sam Tang <[email protected]> wrote: > Dear all > > Thanks a lot for the numerous input which is highly appreciated. > > I should provide some updates as to what I have done on these two crystals > / datasets. > > I processed both datasets with HKL2000 as well as imosflm. Both softwares > give (essentially) the same unit cells as in the first email. The C2 cell > doesn't look like a transformation of the P1 cell. > > However Phaser (and in fact Xtriage) identified that there is > translational pseudosymmetry in the C2 dataset but not the P1 dataset. We > are lucky enough to obtain MR solutions for the dataset but we remain a bit > cautious about the C2 we had. Currently we are looking at Zanuda to see if > this is what it should be. > > Thanks again! > > Kind regards > > Sam > > > > > On 30 October 2016 at 14:54, Kevin Jude <[email protected]> wrote: > >> 3ICE had two space groups in the same /crystal/ - P6 at one end, P1 with >> pseudo-6-fold NCS at the other. In the P6 case, the RNA ligand (in the >> center of a hexameric protein) was rotationally averaged, but in the P1 >> case it could be resolved. >> >> Best wishes >> Kevin >> >> On Fri, Oct 28, 2016 at 6:13 AM, Sam Tang <[email protected]> wrote: >> >>> Dear all >>> >>> Sorry for going a bit off-topic in this thread. >>> May I seek your advice as on whether you have experienced that crystals >>> being obtained from the same droplet, looking alike under microscope (rod >>> shape) and in fact growing possibly from a same nuclei, give two space >>> groups after indexing? >>> >>> I recently obtain crystals for a protein (co-crystallized with a nucleic >>> acid ligand) and collected two datasets from synchrotron. Although these >>> two crystals are from the same drop, the SG and unit cell dimensions are >>> very different: >>> >>> Xtal1: C121 (156 60 105 90 111 90) (L-test, Pointless shows that there >>> is no twinning), ~2.5 Angstrom >>> Xtal2: P1 (53 60 79 106 105 98), ~3 Angstorm >>> >>> Would it be possible that the ligand changes the SG of the crystal so >>> that only one of the forms contains the ligand? >>> >>> Any advice is appreciated and thanks a lot in advance for your input. >>> >>> Regards >>> >>> Sam Tang >>> Biochemistry Programme, School of Life Sciences, CUHK >>> >>> >> >
