Since I did not see a reaction on the BB a few comments from me.
The first is that it never hurts to grow better crystals. Life and the referees will be much easier.
The good news is that it is often possible to get surprisingly useful data from badly looking diffraction images. However, it does involve a lot of (manual) trial and error. In this case I would use XDS.
The first thing to do is to look at your complete range of diffraction images. Are the spots just split, or do they really diverge after a number of images. If they are just split but otherwise stay together in x,y and phi, you might consider a slightly larger integration box, to catch both. If there are frames with serious problems: no diffraction spots, smeaered diffraction spots etc. you may want to exclude them from processing.
The next step is to adjust the input parameters such that the autoindexing routine will lock into the major diffraction pattern and will not get confused by the minor patterns. In XDS, parameters to adjust are STRONG_PIXEL and MINIMUM_NUMBER_OF_PIXELS_IN_A_SPOT. You want to set them as large as possible to get as little as possible spots from the minor patterns, while still getting enough spots from the stronger pattern for autoindexing. Another parameter to play with is the SPOT_RANGE. Sometimes it works best to use your complete scan, sometimes it is better to use a limited number of frames from the start, the middle or the end of a run. If you know the cell parameters, sometimes it helps to give them, to force autoindxing to look for these cell parameters, sometimes it is better not to give them.
Von: CCP4 bulletin board [mailto:[log in to unmask]] Im Auftrag von RHYS GRINTER
Gesendet: Dienstag, 2. Juli 2013 16:44
An: [log in to unmask]
Betreff: [ccp4bb] Split Crystal Dataprocessing
I collected some data on the weekend on forked crystal, I collected data on this crystal at the base before the crystal split into two.
The crystal didn't stand up well to the radiation damage so I shot a number of places along the crystal and got maybe 45 degrees of good data per position. Auto-processing failed on all but one data set, this dataset processed to 3.99 A, but only with around 80% completeness. However looking at the diffraction images I see spots in the first 45 degrees to at least 3.2 angstroms.
I tried quickly to manually process in mosflm, but I noticed that many of the spots appear to be in fact made up to two very closely located spots. This data was collected at a micro-focus station so it was impossible to tell this without careful analysis of the spots. I guess these spots are an indication that the lattice was splitting even at this point.
As a relative novice at data processing, I'm wondering if this kind of data is processable and if so what is the best strategy (or if I should just get back to the bench and grow some more crystals) and program to use?