Identification of the point of diminishing returns in high-multiplicity data collection for sulfur SAD phasing

High-quality high-multiplicity X-ray diffraction data were collected on five different crystals of thaumatin using a homogeneous-profile X-ray beam at E = 8 keV to investigate the counteracting effects of increased multiplicity and increased radiation damage on the quality of anomalous diffraction data collected on macromolecular crystals. By comparing sulfur substructures obtained from subsets of the data selected as a function of absorbed X-ray dose with sulfur positions in the respective refined reference structures, the doses at which the highest quality of anomalous differences could be obtained were identified for the five crystals. A statistic σ{ΔF}_D, calculated as the width σ of the normalized distribution of a set {ΔF} of anomalous differences collected at a dose D, is suggested as a measure of anomalous data quality as a function of dose. An empirical rule is proposed to identify the dose at which the gains in data quality due to increased multiplicity are outbalanced by the losses due to decreases in signal-to-noise as a consequence of radiation damage. Identifying this point of diminishing returns allows the optimization of the choice of data collection parameters and the selection of data to be used in subsequent crystal structure determination steps.