Monitoring the anomalous scattering signal and noise levels in X-ray diffraction of crystals

A statistical index R_as is proposed in order to monitor the overall signal-to-noise ratio in an anomalous scattering data set. In this approach, symmetry-equivalent reflections are merged and grouped into centric and non-centric subsets. Reflections in the centric subset, which in theory should be equal, are used to estimate the noise level in the data. This approach differs from that used by most data-processing programs, in which the centric reflections are merged and averaged. By preserving the differences in centric reflections during data processing, an internal measure of the noise level can be estimated and used to analyze the quality of the anomalous signal in the data. An index R_as is defined as the ratio of the average Bijvoet difference of merged acentric reflections to merged centric reflections. Test results on a variety of data show that R_as has good correlation with the capability to determine the anomalous scattering substructure from the data. R_as can also be useful in monitoring the quality of the data in terms of the data-collection strategy, instrument settings and data-processing software used. R_as analysis has been implemented in the program 3DSCALE as part of a data-processing program suite that is under development in our laboratory.