Oscillation camera data processing. 4. Results and recommendations for the processing of synchrotron radiation data in macromolecular crystallography

The accuracy of oscillation camera data in macromolecular crystallography is dependent upon a variety of factors including crystal quality, experimental conditions and the processing of the data itself. In cases where the best possible data have been collected on film or by use of an area detector, an accurate knowledge of the data collection conditions and a careful choice of the processing technique generally provide structure amplitudes far superior to those obtained by less exhaustive procedures. Particular areas in the processing of oscillation camera data which require careful attention are the choice of a rocking curve, the prediction and treatment of partial reflections if they are present, the method of integration of the full reflections and post-refinement procedures for the verification and improvement of the crystal setting matrix and experimental conditions. A detailed description of these procedures is given for the processing of film data from a synchrotron source. In particular, it is shown that profile-fitting techniques are far superior to simple summation, while a theoretically justified profile fitting of partial reflections is shown to improve these measurements greatly, particularly where contiguous film summation is used. These modifications are shown to give a considerable increase in the number of significant intensities towards the resolution limit. Partial bias is discussed in terms of the integration method employed, fraction recorded and resolution.