I would keep the hexagonal lattice, assuming your integration was OK - no missed spots and then index the data as P1. If the anom Pattersons should then show the correct symmetry - if the Patterson shows P6/mmm symmetry you should be safe in either assigning P6122 or P6522 to the structure SG. Eleanor But actually it should be enough to index and merge as P3 then let the Patterson symmetry suggest the SG. Cant you run pointless on the unmerged data? That gives a useful indication of what symmetry operators are present. Eleanor
On 14 Mar 2013, at 20:31, wtempel wrote: > Dear colleagues, > we have collected 200 degrees worth of oscillation data on a selenomethionine > derivative (1 "non-terminal" methionyl per 120 residues) to 2A resolution. > XDS provides the following output: > <COLLECT.LP> > LATTICE- BRAVAIS- QUALITY UNIT CELL CONSTANTS (ANGSTROEM & DEGREES) > CHARACTER LATTICE OF FIT a b c alpha beta gamma > > * 44 aP 0.0 47.9 47.9 171.2 90.1 90.0 119.9 > * 31 aP 0.4 47.9 47.9 171.2 89.9 90.0 60.1 > * 39 mC 0.8 83.0 47.9 171.2 90.0 90.1 89.9 > * 14 mC 2.2 47.9 82.9 171.2 90.0 90.1 90.0 > * 34 mP 2.3 47.9 171.2 47.9 90.1 119.9 90.0 > * 29 mC 2.6 47.9 82.9 171.2 90.0 90.1 90.0 > * 38 oC 2.8 47.9 83.0 171.2 89.9 90.0 90.1 > * 10 mC 2.8 83.0 47.9 171.2 90.0 90.1 89.9 > * 13 oC 2.9 47.9 82.9 171.2 90.0 90.1 90.0 > * 12 hP 3.4 47.9 47.9 171.2 90.1 90.0 119.9 > 35 mP 249.9 47.9 47.9 171.2 90.0 90.1 119.9 > 33 mP 250.4 47.9 47.9 171.2 90.0 90.1 119.9 > 32 oP 251.8 47.9 47.9 171.2 90.1 90.0 119.9 > . > . > . > SPACE-GROUP UNIT CELL CONSTANTS UNIQUE Rmeas COMPARED > LATTICE- > NUMBER a b c alpha beta gamma > CHARACTER > > 5 82.9 48.0 171.3 90.0 90.0 90.0 13033 15.3 36401 > 10 mC > 143 47.9 47.9 171.3 90.0 90.0 120.0 8371 17.4 41063 > 12 hP > 149 47.9 47.9 171.3 90.0 90.0 120.0 4496 17.0 44938 > 12 hP > 150 47.9 47.9 171.3 90.0 90.0 120.0 4720 16.8 44714 > 12 hP > 168 47.9 47.9 171.3 90.0 90.0 120.0 4255 16.9 45179 > 12 hP > * 177 47.9 47.9 171.3 90.0 90.0 120.0 2549 16.7 46885 > 12 hP > 21 48.0 82.9 171.3 90.0 90.0 90.0 6758 15.1 42676 > 13 oC > 5 48.0 82.9 171.3 90.0 90.1 90.0 12751 15.8 36683 > 14 mC > 5 48.0 82.9 171.3 90.0 90.1 90.0 12751 15.8 36683 > 29 mC > 1 47.9 47.9 171.3 89.9 90.0 60.1 24755 9.4 24679 > 31 aP > 21 47.9 83.0 171.3 90.0 90.0 90.0 6765 9.2 42669 > 38 oC > 3 47.9 171.3 47.9 90.0 119.9 90.0 12671 9.6 36763 > 34 mP > 5 83.0 47.9 171.3 90.0 90.1 90.0 13029 8.9 36405 > 39 mC > 1 47.9 47.9 171.3 90.1 90.0 119.9 24755 9.4 24679 > 44 aP > </COLLECT.LP> > We scaled reflections, with reasonable residuals, assuming space group P622 > and have a reasonable solution in P6122. Refinement does not take us past an > Rfree of aprrox. 0.4. The apparent quality of the model-phased map varies > across the asymmetric unit. > The L-test does not indicate merohedral twinning. Translational > pseudosymmetry was not detected. <Merohedral> twinning has been tentatively > ruled out. > > Before considering Order-Disorder-type crystal defects (as reviewed by Andrey > Lebedev here: http://www.ysbl.york.ac.uk/mxstat/andrey/ecm2010.pdf), we would > like to rule out that we might have erroneously applied too many symmetry > constraints. To that end, we have scaled data in C2. We observed anomalous > difference fourier peaks corresponding to positions of 6 selenium atoms in > the current model, but refinement still does not proceed, and maps look > similar to P6122 treatment. To "cover all bases" we would like to extend the > asymmetric unit to P1. And here is the question: > > Does it suffice to pursue any one of the P1-associated lattices in the XDS > table or do both need to be tested? > > Many thanks as always, > Wolfram Tempel
