Category: Data Analysis and Informatics
The lead optimization stage of drug discovery is where researchers attempt to improve upon lead compounds and select a small number to advance as preclinical candidates. Numerous types of assays may be run at this stage, including primary assays against a therapeutic target, selectivity assays against related targets, and a host of toxicity and ADME (Absorption Distribution Metabolism and Excretion) assays. For any given drug discovery program, many of these assays will be run regularly for the duration of the program which could last several years, continuing to produce back-up compounds as early leads go into human testing. The long-running nature of these assays lends well to a monitoring program to track variability of the data produced. Such a monitoring program can help investigators to identify problems while enabling decision makers to take into account normal variability. Although it is common practice to use reference compounds to monitor assay performance over time, analysis tools for this type of data are lacking. Here we present Analyze™, an integrated system that seamlessly incorporates production monitoring and state-of-the-art data analysis in one tool. Following recommendations outlined in the NIH Assay Guidance Manual, we use the MSR (Minimum Significant Ratio) as the primary statistic to characterise overall assay variability. For each assay run, the system automatically updates the MSR and associated plots. Both the overall and within-in run MSR can be calculated, provided that a second reference compound is included in the analysis. We demonstrate the utility of this approach using a large set of simulated lead optimization data. We also show how decisions based on a basic understanding and knowledge of assay variability can help the lead optimization process.
Stephen Day– Application Scientist, Scigilian Software, Inc, Montreal, QC, Canada