Realizing New Horizons in X-ray Crystallography Data Processing

University Of Calif-Lawrenc Berkeley Lab, Berkeley CA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): This project responds to NIH/NIGMS Program Announcement PAR-10-073, Technology Development for High-Throughput Structural Biology. The atomic structure of macromolecules as determined by X-ray crystallography has led to key biological insights in the last 60 years, and remains a prime focus of both academic and industrial research today. X-ray diffraction experiments are largely performed at synchrotron radiation facilities in the U.S. and abroad. The success of the experiment is dependent on the collection of large amounts of diffraction data, often from numerous crystalline samples. The best outcome is achieved if the diffraction datasets can be processed immediately, allowing real-time experimental adjustments and/or the examination of alternate crystal specimens. Aging software used to process the diffraction images is rapidly becoming obsolete for two reasons. First, the diversity of diffraction results is not well-modeled, and secondly, a new generation of X-ray detectors using pixel-array technology will far outpace the ability to process the data. These challenges will be addressed in this proposal by a team of scientists working at two synchrotron facilities (the Advanced Light Source at Lawrence Berkeley National Laboratory and the Stanford Synchrotron Radiation Laboratory). A new software system, New Horizons, will be designed for high-throughput operation at synchrotron beamlines, where it can be incorporated into existing software environments that control the experiment. Rapid measurements of crystal quality will be reported to the experimenter, so that protocol adjustments can be made. The new system will be benchmarked against existing data processing programs such as MOSFLM, to assure adherence to standard practice. Orders-of-magnitude advances in speed will be obtained by implementing the software on general purpose graphics processing units (GPGPUs), while at the same time adhering to standards such as OpenCL to assure portability. GPU acceleration will permit the implementation of superior models based on small-molecule work, extending both the data processing accuracy and range of applicability; in particular it will be possible to analyze a wider variety of crystals containing lattice defects. An open-source model will be used to assure that New Horizons can be utilized and adapted by synchrotron facilities worldwide.