Dear Alastair, Since there does not seem to be a compelling reason to use P 32 to describe the contents of your unit cell, I would use P 32 1 2. Refining a structure in a lower symmetry space group often gives slightly lower Rfactors since the refinement program has more parameters to play with . Also selection of the free reflections plays a role. If you have free and working reflections related by the twofold, the free reflections are not really free in P 32, which may cause artificially lower Rfrees.
A different story would be if your missing residues would be close to the twofold and would adopt alternative conformations. However, even in this case, it would be very unlikely that all molecules say in the "upper part" of the asymmetric unit would adopt conformation A and in the "lower part" conformation B. Much more likely would be a random distribution, which brings you back to P 32 1 2. Best, Herman Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Alastair MC EWEN Gesendet: Dienstag, 7. Januar 2014 17:38 An: CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] Space group ambiguity in Zanuda Dear All, I have a question regarding the output of Zanuda. We are working on a structure solved in P 32 1 2 with one molecule in the asymmetric unit. The structure was solved by molecular replacement and refines well to R = 20.6%, Rfree = 25.2% with data to 1.8Å and 90% of the model built. Most of the missing residues are 15 residue long section that is predicted to form a helix and although there is some ambiguous density it is not possible to construct anything. To check the solution I submitted the structure to Zanuda which suggested that the symmetry was too high and the correct space group should be P 32. I have refined the structure in P 32 and although the stats are better (R = 19.9%, Rfree = 23.1%) there is no improvement in the density and the two copies remain largely identical. However when I resubmitted this structure to Zanuda it suggests that the symmetry is too low and the correct space group is P 32 1 2. But looking at the log file I am not sure why this should be so. The program does pick P 32 in step 2 but then selects P 32 1 2 at the end of step 3 even though the R free is better in P 32 (R is slightly worse). I'm not sure what I am missing here. I have tried molecular replacement in P 1 followed by refinement and still get P 32 1 2 picked over P 32 which has the best R and Rfree. I have attached an extract from the log files at the end of this email. Does anyone have any suggestions? Thanks in advance, Alastair McEwen P 32 logfile Step 1. R-factors for the starting model. Transformation into a supergroup. current time: Nov 27 14:19 CET expected end of job (rough estimate): Nov 27 14:30 CET --------------------------------------------------------------------- | Subgroup | Spacegroup | R.m.s.d. | Refinement in tested group | | | | from the |--------------------------------| | Ref | | starting | Rigid | Restrained | | | | model, A |----------|---------------------| | | | | R | R | R-free | |----------|------------|----------|----------|----------|----------| | >> 2 | P 32 | 0.0006 | -- | 0.2716 | 0.3084 | | 6 | P 32 1 2 | 0.1035 | -- | -- | -- | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Step 2. Refinements in subgroups. There are 4 subgroups to test. current time: Nov 27 14:19 CET expected end of job: Nov 27 14:32 CET ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | >> 2 | P 32 | 0.0006 | -- | 0.2716 | 0.3084 | --------------------------------------------------------------------- | 1 | P 1 | 0.1728 | 0.2777 | 0.2521 | 0.3288 | | 2 | P 32 | 0.1738 | 0.2779 | 0.2509 | 0.3279 | | 3 | C 1 2 1 | 0.2080 | 0.2872 | 0.2519 | 0.3315 | | 6 | P 32 1 2 | 0.2040 | 0.2899 | 0.2526 | 0.3434 | --------------------------------------------------------------------- | << 2 | P 32 | 0.1738 | 0.2779 | 0.2509 | 0.3279 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Step 3. Refinement of the best model. Candidate symmetry elements are added one by one. current time: Nov 27 14:25 CET expected end of job: Nov 27 14:32 CET ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | >> 2 | P 32 | 0.1738 | 0.2779 | 0.2509 | 0.3279 | --------------------------------------------------------------------- | 1 | P 1 | 0.1728 | 0.2778 | 0.2521 | 0.3290 | | 2 | P 32 | 0.2231 | -- | 0.2467 | 0.3305 | | 6 | P 32 1 2 | 0.2679 | -- | 0.2452 | 0.3444 | --------------------------------------------------------------------- | << 6 | P 32 1 2 | 0.2679 | -- | 0.2452 | 0.3444 | --------------------------------------------------------------------- R-factor in the original subgroup is NOT the best. The original spacegroup assignment seems to be incorrect P1 logfile Step 1. R-factors for the starting model. Transformation into a supergroup. current time: Dec 11 11:18 CET expected end of job (rough estimate): Dec 11 11:32 CET --------------------------------------------------------------------- | Subgroup | Spacegroup | R.m.s.d. | Refinement in tested group | | | | from the |--------------------------------| | Ref | | starting | Rigid | Restrained | | | | model, A |----------|---------------------| | | | | R | R | R-free | |----------|------------|----------|----------|----------|----------| | >> 1 | P 1 | 0.0000 | -- | 0.2753 | 0.3090 | | 6 | P 32 1 2 | 0.1947 | -- | -- | -- | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Step 2. Refinements in subgroups. There are 6 subgroups to test. current time: Dec 11 11:18 CET expected end of job: Dec 11 11:35 CET ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | >> 1 | P 1 | 0.0000 | -- | 0.2753 | 0.3090 | --------------------------------------------------------------------- | 1 | P 1 | 0.1973 | 0.2780 | 0.2482 | 0.3480 | | 2 | C 1 2 1 | 0.2431 | 0.2837 | 0.2517 | 0.3449 | | 3 | C 1 2 1 | 0.2196 | 0.2863 | 0.2525 | 0.3473 | | 4 | P 32 | 0.2859 | 0.2855 | 0.2503 | 0.3393 | | 5 | C 1 2 1 | 0.2815 | 0.2907 | 0.2530 | 0.3483 | | 6 | P 32 1 2 | 0.2914 | 0.2935 | 0.2574 | 0.3493 | --------------------------------------------------------------------- | << 4 | P 32 | 0.2859 | 0.2855 | 0.2503 | 0.3393 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Step 3. Refinement of the best model. Candidate symmetry elements are added one by one. current time: Dec 11 11:29 CET expected end of job: Dec 11 11:37 CET ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | >> 4 | P 32 | 0.2859 | 0.2855 | 0.2503 | 0.3393 | --------------------------------------------------------------------- | 1 | P 1 | 0.2767 | 0.2863 | 0.2482 | 0.3455 | | 4 | P 32 | 0.3320 | -- | 0.2497 | 0.3409 | | 6 | P 32 1 2 | 0.3698 | -- | 0.2550 | 0.3578 | --------------------------------------------------------------------- | << 6 | P 32 1 2 | 0.3698 | -- | 0.2550 | 0.3578 | --------------------------------------------------------------------- R-factor in the original subgroup is NOT the best. The original spacegroup assignment seems to be incorrect. *************************************************** Alastair McEwen, PhD Structural Biology and Genomics Technology Platform Integrated Structural Biology IGBMC 1 rue Laurent Fries 67404 ILLKIRCH - FRANCE tel: +33 (0)3 88 65 57 92 ***************************************************
