W0213

High Throughput Crystallization Screening For Structural Genomics and Data Mining. Brent Segelke, Macromolecular Crystallography Group at Lawrence Livermore National Laboratory, Livermore, CA 94551

With the coming wave of structural genomics comes the technological challenge of high throughput crystallization. Crystallization is known to be a stumbling block in structural studies and can be reasonably anticipated as a major stumbling block in structural genomics. Crystallization as it is commonly pursued, is a simple combination of small numbers of components from a large finite set of parameters known to influence protein solubility. As such, experimental design for crystallization is amenable to software automation. We have developed, the program CRYSTOOL for random crystallization screening and have demonstrated the inherent efficiency of random screening. We are currently working toward full automation for high throughput CRYSTOOL screening. There are several dispensing robots commercially available, which are either designed for crystallization setup or reasonably easily adapted to crystallization setup with CRYSTOOL. Having automated the design and setup of crystallization trials, automated analysis will bring great benefit for high throughput crystallization screening. There is at least one commercially available system (Diversified Scientific, Inc. CRYSTALSCORE) that addresses automated crystal detection. Our plan is to combine several of these available components, CRYSTOOL, dispensing robot, and CRYSTALSCORE, into an all-in-one, fully automated, crystallization lab. From our experience to date it is apparent that dispensing will be the rate-limiting step (total throughput for one dispensing robot may be as high as 1000 experiments/day).

A lab equipped with such an automated system for CRYSTOOL crystallization trials will not only be able to generate and analyze a large number of crystallization experiments but will also be generating a wealth of data. This data will provide new insights into crystallization generally and will lead to efficiency gains, improved estimates of throughput, higher throughput, and reduced costs.

Lawrence Livermore National Laboratory is operated by the Univ. of California for the US DOE under contract W-7405-ENG-48.