Statistical Design of Pools Using Optimal Coverage and Minimal Collision

• Activity Number: 425
• Type: Contributed
• Date/Time: Wednesday, August 6, 2003 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Katja Remlinger*+ and Jacqueline M. Hughes-Oliver and S. Stanley Young and Raymond L. H. Lam
• Companies: North Carolina State University and North Carolina State University and Glaxo Wellcome Inc. and Glaxo Wellcome Inc. Canada
• Keywords: synergism ; uniform cell coverage designs ; blocking ; drug discovery ; high throughput screening ; pooling experiments

Abstract

Discovery of a new drug involves identification of compounds that are potent with respect to a given biological assay. Chemical libraries, collections of compounds, are very large but only a small number of compounds available for screening are actually potent. One approach to improving screening efficiency is to test compounds in pools, e.g., sets of 10 compounds. The questions that arise are: how should the pools be designed, and can chemical structure help to create pools? We consider two criteria: optimal coverage of the chemical space and minimal similarity between compounds within a pool, i.e., minimal collisions between compounds. This work compares four pooling designs that fulfill one or both of these criteria. The computationally intensive methods are applied to a public dataset and we evaluate each design method by determining how well the design criteria are met. One design, called MC, uniformly dominates all other designs, but all designed pools outperform randomly created pools. Furthermore, we discuss blocking and synergism between compounds as other effects that must be investigated when performing pooling experiments.