Faster data-collection strategies for structure determination using anomalous dispersion

Many macromolecular structures are being determined using anomalous dispersion phasing methods. Different data-collection strategies at one, two, three or more wavelengths can be used for these experiments. The choice of strategy can determine the success or failure of the experiment and should be based on a clear understanding of the advantages and disadvantages of each approach given the experimental constraints and goals. In this paper, several sets of three-wavelength MAD experiment data were reanalyzed using one, two and three wavelengths and systematically removing reflections from the data sets to determine the minimum amount of data required to yield an automatically traceable map as a function of the number of wavelengths used in phasing. In the cases studied here, two-wavelength MAD consistently required fewer data than three-wavelength MAD, as long as the unique data completeness was high at each wavelength. It was also found in some instances that using one wavelength for phasing required as much or more data as using two wavelengths. These results can help with the design of adequate data-collection strategies which maximize the phasing power from the minimal data collected. This is particularly important for minimizing the effects of radiation damage on phasing while taking sample characteristics, beamline properties and experimental goals into account.