The Crystal Screening Interface at ALS

1
The Crystal Screening Interface at ALS
Crystallography Beam Line Automation Work
Smarter Not Harder Stanford Synchrotron Radiation
Laboratory30th Annual Users Meeting October 8,
2003

• Nicholas Sauter, Lawrence Berkeley National Lab

Paul Adams, Computational Crystallography
Initiative
Thomas Earnest, Berkeley Center for Structural
Biology
2
Robohutch Sector 5 Beamlines
3
(No Transcript)
4
Initial Data Entry
5
The Screening Interface
6
How many images should I collect?
Symmetry results based on
Two images 90 apart
One Image
True Beam Center
Conventional Autoindexing
Orthorhombic (correct)
Monoclinic (not quite correct)
0.5 mm
Wrong Indexing
No Indexing
7
Fix Misindexing With LABELIT

• Common experience the predicted pattern looks
  almost right, but something is funny
• Systematic absence of h k odd
• Easily detected and corrected, once the problem
  is recognized

8
Data Processing Sequence
Final Score Success Resolution cutoff Mosaic
spread RMS residual Good strategy Few
overlaps No ice rings Well-shaped spots Minimal
diffuse scatter Short exposure time
Autoindex Determine Bravais lattice
Collect 2 oscillation frames 90 apart
LABELIT 15 seconds

• Heuristic score within each group
• score 1 (.7e 4/resolution)
  (1.5rmsResidual) (.02mosaicity)
• Representative results protein Syrrx-004

9
Software Goals
10
Acknowledgements

• Computational Crystallography Initiative (LBNL)
• Paul Adams
• Ralf Grosse-Kunstleve
• Nigel Moriarty
• Berkeley Center for Structural Biology (LBNL)
• Thomas Earnest
• Robert Nordmeyer
• Carl Cork
• Earl Cornell
• John Taylor
• Stanford Synchrotron Radiation Laboratory
• Ashley Deacon
• Zepu Zhang
• Syrrx, Inc.
• Gyorgy Snell
• MRC Laboratory of Molecular Biology
• Andrew Leslie
• Harry Powell
• Daresbury Laboratory