Improving Detection of Rare Biological Events in High-Throughput Screens. (February 2015)

• Record Type: Journal Article
• Authors: Murie, Carl; Barette, Caroline; Button, Jennifer; Lafanechère, Laurence; Nadon, Robert
• Is Part Of: Journal of biomolecular screening. Volume 20:Number 2(2015)
• Page Start: 230
• Page End: 241
• Publication Date: 2015-02
• Subjects: study design -- statistical analysis -- normalization -- randomization; Drugs -- Analysis -- Periodicals; Drugs -- Testing -- Periodicals; Biomolecules -- Analysis -- Periodicals; 572.36
• DOI: 10.1177/1087057114548853
• Languages: English
• ISSNs: 1087-0571

Abstract

The success of high-throughput screening (HTS) strategies depends on the effectiveness of both normalization methods and study design. We report comparisons among normalization methods in two titration series experiments. We also extend the results in a third experiment with two differently designed but otherwise identical screens: compounds in replicate plates were either placed in the same well locations or were randomly assigned to different locations. Best results were obtained when randomization was combined with normalization methods that corrected for within-plate spatial bias. We conclude that potent, reliable, and accurate HTS requires replication, randomization design strategies, and more extensive normalization than is typically done and that formal statistical testing is desirable. The Statistics and dIagnostic Graphs for HTS (SIGHTS) Microsoft Excel Add-In software is available to conduct most analyses reported here.