When working with low-resolution datasets (especially those with very few reflections), some non-default options might be necessary to process the data with autoPROC. It is also often necessary to run some additional steps by hand, ie. taking the initial, autoPROC-generated files as a starting point for more fine-tuning.

Here we will collect some recipes for dealing with such datasets.

Start with defaults

The first assumption is, that a dataset can be processed with all defaults, ie. in the same way a high-resolution dataset from the same beamline/instrument and shift would be processed. Of course, this might include some parameter settings defining beam centre convention etc, but nothing specific to a low-resolution dataset.

Even if that run terminates without error, some checks should be done:

If an autoPROC run terminates with an error or some of the above criteria are not met, the following sections will show some possible actions one could take.

Some useful macros and settings

A good set of options is to run with

% process -M LowResOrTricky ...

since this will set a few parameters typical for low-res (or any other more challenging) dataset. See output of process -M show what parameters this macro sets.

If the cell and/or spacegroup is known, running with

% process cell="40 50 60 90 90 90" symm=P21212 ...

can help XDS to find the correct indexing solution.

It is also a good idea to restrict the processing to the resolution where there actually are some reflections, eg by giving a resolution limit

% process -R 100.0 3.2 ...

or by specifying the detector edge via

% process autoPROC_XdsKeyword_TRUSTED_REGION="0.0 1.05" ...