e HTPX crystallization shipping and future Chris Mayo

e. HTPX crystallization, shipping and future Chris Mayo, Ian Berry, Jon Diprose & Robert Esnouf Mayo et al. (2005), Structure 13, 175– 182.

OPPF crystallization facility n n High-throughput nanolitre-scale sitting-drop crystallization ¨ Hydra 96 and 2 Cartesian Micro. Sys 400 ¨ 12 minutes per 96 -well plate, 50 plates/day Plate storage system ¨ TAP Homebase, capacity 10, 000 microtitre plates ¨ Automated hand-off to imaging system Imaging system ¨ Veeco Oasis 1700 ¨ ~50, 000 images/day, >31, 000 images so far Custom IT infrastructure ¨ Control software, Linux farm ¨ Web-based user interface

A Pi. MS crystallization module n Module to cover management of crystallization trials ¨ Shared development with Bio. XHIT allowing external resources to contribute to Pi. MS goals ¨ A self contained and well understood module ¨ One of most demanding in terms of database use and user interface performance ¨ Of immediate use to several Pi. MS sites n Four-way grouping: OPPF, NKI, EMBL Grenoble, Paris-Sud

Oxford e. HTPX trip 28/30 Nov 2005 n n n n 3 rd e. HTPX test trip To collect some new/live data! For testing and debugging beamline automation To get more user experience with the sample changer To test automated submission of frozen samples and develop interface further To demonstrate ISPy. B including remote access For Ian and Ludovic to work on crystal/data link

OPPF 2857 n Day 1 screened 7 crystals, nasty spots and multiple lattices but strong diffraction. Collected slightly weaker but better looking: ¨ Spacegroup P 6? n ¨ a=91. 3Å, b=91. 3Å, c=246. 5Å Data to 2. 1Å: n ~80% complete n R = >30%

OPPF 2857 n Day 2 screened 4 crystals refining stepwise glycerol treatment: ¨ Spacegroup P 21 n a=74. 7Å, b=73. 3Å, c=76. 7Å, β=119. 2 All data to 1. 8Å: n 91. 0% complete n R = 5. 5% n I/σ(I)=28. 1 ¨ Outer shell: n 71. 3% complete n (98. 8% at 1. 94Å) n R = 53. 1% n I/σ(I)=1. 3 ¨

Feedback from 3 rd e. HTPX trial n n Sample changer is revolutionary ¨ Enables better science ¨ Less agonising over marginal (3– 4Å) cases ¨ Allows a more rational approach to cryo-protection ¨ Allows planning of data collection runs to reduce stress ¨ Barcoding needs to be more robust ¨ Checking could still be better and interface easier Limitation ¨ Unlikely to trust auto-centering for best samples ¨ Remounting accurately would be good

Feedback from 3 rd e. HTPX trial n n For Oxford the e. HTPX hub was of marginal benefit ¨ Can enter samples direct into ISPy. B ¨ Requires porting data from OPPF crystallization databases to its tray manager ¨ Requires laborious data entry in a fixed order ¨ Offers no mechanism for recording history of synchrotron use ¨ Not enough flexibility for recording all forms of cryo-treatment ¨ Pfizer group spent 6 man hours entering data already in an Excel spreadsheet! Suggested way forward ¨ Splitting tray manager section and crystal shipping section ¨ More flexible and tolerant model for crystal tracking ¨ Handling storage Dewars at home laboratories ¨ Importing. XLS (and. XML) files directly

Developing a generalized container model for sample tracking / shipping • Have container description as one entry • Have instance with ‘where’, ‘valid from’ and ‘valid to’ dates • These are updated never deleted • Gives full history of all use • Could archive old records Crystal on P 7 beam line? O+P 4 O+P 6 • Location • Sample changer? • Storage Dewar • Shipping Dewar • Pin (extra level for pucks) • Crystal P 1 O 1 P 0 O+P 5 P 2 OPPF Pfizer O+P 3 In transit O 2 O 7 BM 14

Container model requires hierarchy