Reindex or change spacegroup¶
This task provides a number of utility functions for merged reflection data files, primarily to change space group or reindex the data relative to different axes, otherwise know as a change of basis. It will reindex a set of observations, and optionally a set of Free-R flags, to match reference data as an external file, or from explicit operators. There is also an option just to analyse the symmetry of merged data, to check for under-merging or to check eg the cell conventions in P1
Sometimes when comparing similar structures, the spacegroup or indexing convention may be different even when the crystal packing is similar. In this case the symmetry and the indexing convention changed to match the existing structure.
Similarly during the course of a structure solution it may become obvious that the spacegroup has been incorrectly determined. For example, a symmetry which is present in the data may have been missed leading to the selection of a spacegroup with lower symmetry, or a higher spacegroup may have been selected due to the presence of twinning or pseudo-symmetry. If point group has changed, then the data processing should be repeated to obtain a correct list of unique reflections. If the space group needs to be changed with the same point group (eg P 2 1 2 1 2 1 to P 2 1 2 1 2) then this utility can be used. Illegal changes will fail.

Figure 1: Input pane¶
Input¶
One or two data files are processed, a reflection data object, and optionally a matching Free-R set (1.1)
The pull-down (1.2) specifies what is to be done: options 1-4 need a reference file (1.3) , see below.
Options:
Observed data reference: an observed data object
Calculated data reference: calculated data
2FoFc coefficient reference: map coefficients, eg from Refmac
Coordinate reference: a coordinate file from which intensities will be calculated
User specified spacegroup and reindex, see below
Remove lattice absences, see below
Just analyse data symmetry: possibly useful after an explicit reindex job, to check the result - no output file is produced
Explicit space group and reindex¶

Figure 2: User specified spacegroup and reindex¶
As stated in the figure 2 screenshot, you can give either a new space group (in the same point group), or a reindex operator, or both. Pointless will try to generate the other, but it doesn’t always work in complicated cases. If the reindexing operator leads to non-integral hkl indices, these will be removed, eg [h/2, k, l] will halve the cell in the a direction (a drastic step, be careful that you know what you are doing!): such Reindex operators will change the space group - check that you think it correct, and if necessary run the “Just analyse data symmetry” option on the output
Examples:
P 3 1 to P 3 2
P 2 1 2 1 2 1 to P 2 1 2 1 2
Reindex P 2 2 1 2 1 to P 2 1 2 1 2 using the reindex operator [k,l,h] (a cyclic permutation to preserve the hand)
C 2 to I 2
[k, h, -l] in P 3 etc
[h/2+k/2, -h/2+k/2, l] to remove reflections which would be C-centred: rotates the axes by 45°
Remove lattice absences¶

Figure 3: Removed centred lattice absences¶
Warning
THIS IS NOT RECOMMENDED UNLESS YOU KNOW WHAT YOU ARE DOING.
Occasionally a dataset has been wrongly integrated with a doubled cell, such that eg half the spots are absent, or you would like to see what happens in a half-cell when there is tNCS. Spots corresponding to centred lattices can be removed with this option, eg h+k+l odd for an I lattice. This is easier than working out the equivalent reindexing operator. The space group will change: check that you think the new space group is correct
Note
an equivalent option is available in the Data reduction pipeline (under Additional options), with the same caveats. Maybe you should go back and look closely at your images