Guidelines for de novo phasing using multiple small-wedge data collection

Intense micro-focus X-ray beamlines available at synchrotron facilities have achieved high-quality data collection even from the microcrystals of membrane proteins. The automatic data collection system developed at SPring-8, named ZOO, has contributed to many structure determinations of membrane proteins using small-wedge synchrotron crystallography (SWSX) datasets. The ‘small-wedge’ (5–20°) datasets are collected from multiple crystals and then merged to obtain the final structure factors. To our knowledge, no systematic investigation on the dose dependence of data accuracy has so far been reported for SWSX, which is between ‘serial crystallography’ and ‘rotation crystallography’. Thus, herein, we investigated the optimal dose conditions for experimental phasing with SWSX. Phase determination using anomalous scattering signals was found to be more difficult at higher doses. Furthermore, merging more homogeneous datasets grouped by hierarchical clustering with controlled doses mildly reduced the negative factors in data collection, such as ‘lack of signal’ and ‘radiation damage’. In turn, as more datasets were merged, more probable phases could be obtained across a wider range of doses. Therefore, our findings show that it is essential to choose a lower dose than 10 MGy for de novo structure determination by SWSX. In particular, data collection using a dose of 5 MGy proved to be optimal in balancing the amount of signal available while reducing the amount of damage as much as possible.