Robust Analysis of High-Throughput Screening (HTS) Assay Data

• Activity Number: 153
• Type: Invited
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Changwon Lim*+ and Pranab K Sen and Shyamal D. Peddada
• Companies: Loyola University and The University of North Carolina at Chapel Hill and NIEHS, NIH
• Keywords: Dose-response study ; False discovery rate (FDR) ; Heteroscedasticity ; Hill model ; M-estimation procedure ; Toxicology

Abstract

Quantitative high throughput screening (qHTS) assays use cells or tissues to screen thousands of compounds in a short period of time. Data generated from qHTS assays are evaluated using nonlinear regression models such as the Hill model and decisions regarding toxicity are made using the estimates of the parameters of the model. For any given compound, the variability in the observed response may either be constant across dose groups (homoscedasticity) or vary with dose (heteroscedasticity). Secondly, since thousands of chemicals are evaluated simultaneously, it is not uncommon to find outliers and influential observations in the data. It is well-known that the variance structure and outliers play an important role in the analysis of nonlinear regression models. Since thousands of chemicals are processed in a given run of the assay, it is necessary to develop and use methodology that is robust to variance structure and outliers so that a researcher does not have to perform model diagnostics for each individual model. In this talk we describe such a robust methodology and illustrate it using a data set obtained from the U.S. National Toxicology Program.