Dear all, I have a small three helix bundle domain (~14 kDa) which after data analysis I believed crystallized in either P41212 or P43212 (a=b=121.8 c=118.0) with between 4 and 8 molecules in the ASU and I have data to 3.2A. However after looking at a native Patterson there is a non-origin peak which is also detected in phenix.xtriage and in MolRep. Due to the presence of this pseudo-translation vector I could have any of the P422 space groups. Then to be thorough (and because on some of my images the spots show splitting) I tested for twinning. These tests do not detect obvious twinning (although <I^2/<I>^2 is rather high - see below), however, if I process in P4 assuming merohedral twinning or P222 (a=118.3, b=122.0, c=122.8) assuming pseudo-merohedral twinning, the H-test and Britton-test suggest a twin fraction of ~0.42. All data have similar Rsym values (P422=0.094, P4=0.085, P222=0.082). So my first 2 questions are -
1) Are the twinning tests not revealing twinning due to the presence of the pseudo-translation vector? 2) When I process in P4 or P222 are the twinning fractions from the H-test and Britton-test towards those of a perfect twin because the real point group is actually P422? ----------------------------------------------------------------------------------------------------------------------- 'P422 - xtriage' Wilson ratio and moments Acentric reflections <I^2>/<I>^2 :2.466 (untwinned: 2.000; perfect twin 1.500) <F>^2/<F^2> :0.745 (untwinned: 0.785; perfect twin 0.885) <|E^2 - 1|> :0.815 (untwinned: 0.736; perfect twin 0.541) Centric reflections <I^2>/<I>^2 :3.809 (untwinned: 3.000; perfect twin 2.000) <F>^2/<F^2> :0.615 (untwinned: 0.637; perfect twin 0.785) <|E^2 - 1|> :1.087 (untwinned: 0.968; perfect twin 0.736) NZ test (0<=z<1) to detect twinning and possible translational NCS ----------------------------------------------- | Z | Nac_obs | Nac_theo | Nc_obs | Nc_theo | ----------------------------------------------- | 0.0 | 0.000 | 0.000 | 0.000 | 0.000 | | 0.1 | 0.107 | 0.095 | 0.217 | 0.248 | | 0.2 | 0.211 | 0.181 | 0.340 | 0.345 | | 0.3 | 0.304 | 0.259 | 0.432 | 0.419 | | 0.4 | 0.386 | 0.330 | 0.488 | 0.474 | | 0.5 | 0.451 | 0.394 | 0.538 | 0.520 | | 0.6 | 0.508 | 0.451 | 0.578 | 0.561 | | 0.7 | 0.558 | 0.503 | 0.616 | 0.597 | | 0.8 | 0.602 | 0.551 | 0.652 | 0.629 | | 0.9 | 0.642 | 0.593 | 0.677 | 0.657 | | 1.0 | 0.679 | 0.632 | 0.705 | 0.683 | ----------------------------------------------- | Maximum deviation acentric : 0.058 | | Maximum deviation centric : 0.031 | | | | <NZ(obs)-NZ(twinned)>_acentric : +0.042 | | <NZ(obs)-NZ(twinned)>_centric : +0.010 | ----------------------------------------------- L test for acentric data using difference vectors (dh,dk,dl) of the form: (2hp,2kp,2lp) where hp, kp, and lp are random signed integers such that 2 <= |dh| + |dk| + |dl| <= 8 Mean |L| :0.485 (untwinned: 0.500; perfect twin: 0.375) Mean L^2 :0.317 (untwinned: 0.333; perfect twin: 0.200) The distribution of |L| values indicates a twin fraction of 0.01. Note that this estimate is not as reliable as obtained via a Britton plot or H-test if twin laws are available. ------------------------------------------------------------------------------------------------------------ I also have an NMR structure as a molecular replacement model and a 4 wavelength Ta6Br12 MAD data set and a few derivative IR data sets. Molecular replacement does not work in any of the P422 space groups using MolRep and this may be because there are too many molecules in the ASU or because it is an NMR structure. The highly isomorphous Ta6Br12 data set has been processed to 4A although there is no anomalous signal between 8-4A. The heavy atoms in the 4A isomorphous replacement data seem to have been incorporated although some of the cells are no longer isomorphous. My final question is - 3) Although the lack of anomalous signal above within the MAD data may be due to other issues, is there any way twinning can cause this? I apologise for the long email but before I collect more data I would like to get a few things resolved and hopefully limit the possible space groups to make like much easier (as I feel a little like Alice going deeper into the rabbit hole at the moment). Any responses to these questions and any other suggestions will be very much appreciated. Kind regards James Dr James Garnett Division of Molecular Biosciences, Imperial College London Level 5, Biochemistry Building, South Kensington, LONDON, SW7 2AZ, UK. Tel: +44 (0) 207 594 5464 Fax: +44 (0) 207 594 3057
