Software for Automatic Data Handling in High-Throughput Crystallography
Dr. Paul Adams, Berkeley Center for Structural Genomics, speaking Protein Crystallography in Drug Discovery 2005.
Date Posted: Friday, January 13, 2006
To purchase a DVD containing all the presentations featured at Protein Crystallography in Drug Discovery 2005, please go to the Select Biosciences website.
Launch Presentation
Abstract
The process of drug discovery and target validation increasingly relies on high-resolution structural information from X-ray crystallography. Analysis of large numbers of macromolecular crystals in complex with drug candidates presents unique challenges for synchrotron beamlines and home-lab X-ray facilities. The development of robotic hardware to manipulate cryo-cooled samples makes it possible to generate large numbers of datasets, which then need to be indexed, classified, and reduced. The new LABELIT package from Berkeley National Lab provides software tools to dramatically reduce the human effort required. Diffraction patterns are indexed with more robust procedures which 1) tolerate inaccuracy in the given position of the incident beam of up to several millimeters; 2) check that the deduced unit cell is not an integer multiple of the true unit cell; and 3) correctly accommodate experimental errors in the identification of Bravais symmetry.
As a result, images can be indexed without the need for visual interaction. Complete analysis of oscillation photographs taken at two angular settings can be accomplished in less than 1 minute, thus keeping pace with robotic hardware performing a rapid, high-throughput screen of many crystals. Crystal quality information obtained from this analysis can then be used to rank the set of crystals so that the best samples can be selected for further data collection. While the full dataset is being collected, LABELIT can determine the Patterson symmetry and space group from partial data. This information can be used to finalize or revise the data collection strategy prior to completion.
LABELIT has both a simple command-line interface, and a web-browser interface for displaying autoindexing results. At synchrotron beamline facilities or home sources, data collection systems can be configured to automatically call LABELIT, assuring that data will be efficiently indexed in real time without the need for human intervention. Users can experiment with LABELIT and obtain more information at: http://cci.lbl.gov/labelit.
Work was funded in part by the US Department of Energy under Contract No. DE-AC03-76SF00098, and by NIH/NIGMS under grant number 1P50GM62412.About the speaker
Paul Adams obtained a B.Sc. in Biological Sciences in 1988, and his PhD in 1992 at the University of Edinburgh (Scotland) in the laboratory of Lindsay Sawyer. He then went to Yale University to undertake postdoctoral research with Professor Axel Brunger, which led to the development of the Crystallography & NMR System package used by many structural biologists for structure determination and refinement. In 1999 he joined Lawrence Berkeley National Laboratory as a Staff Scientist. He leads the development of the PHENIX software for automated crystallographic structure solution. This is an international collaboration involving the groups of Tom Terwilliger (Los Alamos National Laboratory), Randy Read (University of Cambridge, U.K.), and Tom Ioerger and Jim Sacchettini (Texas A&M University).
At the Berkeley Structural Genomics Center, where he is Deputy Principal Investigator, his group is developing new software tools for automated data collection and analysis at synchrotron beamlines. He is currently Acting Head of the Berkeley Center for Structural Biology which oversees the running of 5 beamlines at the Advanced Light Source at Lawrence Berkeley National Laboratory.
Launch Presentation