In the associated paper we discuss the issues involved in improving the sensitivity of both approaches to clustering, using, as an example, 999 5 degree wedges from lysozyme in four iso-forms: NAG: native with N-acetylglucosamine (NAG) soaked in, benzamadine: native with benzamadine soaked in, benzamadine plus NAG: native with both NAG and benzamadine soaked in. Native: no ligands Although the cell parameters are changed sufficiently to allow recognition of the NAG soak, it is difficult to filter the benzamadine soak simply on the basis of cell parameter changes, suggesting the desirability of switching from cell-based clustering to reflection-based clustering as early in the process as possible. The data is this dataset consists of a bzipped HKL structure factor tarball and a bzipped cluster data files from the four forms tarball.

Available as a preprint at https://www.biorxiv.org/content/10.1101/141770. Submited to Acta F. Work supported in part by US Dept.~of Energy, Office of Science, DE-AC02-98CH10886 and E-SC0012704} US NIH National Institute of General Medical Sciences, P41RR012408, P41GM103473, and P41GM111244 HJB supported in part by Dectris, Ltd.