Characterizing Irreproducibility in Drug Sensitivity Data from a Large Pharmacogenomic Study

• Activity Number: 339 - SPEED: Biopharmaceutical and General Health Studies: Statistical Methods and Applications, Part 1
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Zoe Rehnberg* and Johann A Gagnon-Bartsch
• Companies: University of Michigan and University of Michigan

Abstract

After the first large-scale pharmacogenomic studies were published in 2012, the consistency of high-throughput drug sensitivity data across experiments has been widely discussed. While gene expression data seems to be well replicated, only varying levels of moderate to poor concordance has been found for drug sensitivity measures (half-maximal inhibitory concentration [IC50] and area under the dose-response curve [AUC]) in multiple large databases. We take advantage of the detailed raw data available from the Genomics of Drug Sensitivity in Cancer (GDSC) project to identify six factors, ranging from data collection to data analysis, contributing to a lack of reproducibility of IC50 estimates within GDSC. Further, we create a set of quality control measures to quantify these sources of irreproducibility and aid in the effective use of the GDSC database.