Implementing multilevel dynamic scheduling for a highly flexible 5-rail high throughput screening system.

Abstract

As automation solves the bottleneck involved in drug screening, new bottlenecks present themselves. Some of these bottlenecks include sample management, hit picking and confirmation, and reagents lost as a result of incomplete runs. To keep up with the demands of a large HTS department, scientists spend a disproportionate amount of time simply feeding these systems with samples and reagents. Automating the sample management functions directly on the screening systems would solve this problem. With the use of online data analysis, an integrated sample store permits automated hit picking and confirmation. In addition to these issues, other bottlenecks are often caused by instrument malfunctions. A single lost run can now mean a loss of hundreds of plates and the reagents associated with their testing. A system was designed to include four assaying systems that are fed by an automated online sample repository system. Redundancy between the assaying systems allows for an extra level of error handling in case of a malfunction. The control of such a system requires a sophisticated scheduler/controller software package capable of coordinating the interaction between multiple systems and reacting to changes in the robotic environment in realtime. This paper discusses the design of the system as well as the requirements and selection of an appropriate scheduler/controller package.