SIMULATION ANALYSIS OF EXPERIMENTAL-DESIGN STRATEGIES FOR SCREENING RANDOM COMPOUNDS AS POTENTIAL NEW DRUGS AND AGROCHEMICALS

Computer simulations have been performed to investigate the effectiveness of experiment design techniques when applied to pharmaceutical and agrochemical random screening. Two design algorithms have been investigated, a maximum dissimilarity technique (stepwise elimination) and an approach based on sampling clusters. Results show that the former technique tends to reduce the chances of finding active molecules quickly. In contrast, the cluster sampling approach can afford small improvements in the rate at which active molecules are identified. However, use of maximum dissimilarity methods may still be beneficial if a premium is placed on finding active molecules with unusual structures.