Zero-dose extrapolation as part of macromolecular synchrotron data reduction

Abstract
Radiation damage to macromolecular crystals at third-generation synchrotron sites constitutes a major source of systematic error in X-ray data collection. Here, a computational method to partially correct the observed intensities during data reduction is described and investigated. The method consists of a redundancy-based zero-dose extrapolation of a decay function that is fitted to the intensities of all observations of a unique reflection as a function of dose. It is shown in a test case with weak anomalous signal that this conceptually simple correction, when applied to each unique reflection, can significantly improve the accuracy of averaged intensities and single-wavelength anomalous dispersion phases and leads to enhanced experimental electron-density maps. Limitations of and possible improvements to the method are discussed.