A Model-Based Approach for Assessing Reproducibility and Outlier Detection in High-Throughput Biological Data

• Activity Number: 349
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 11:15 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Karthik Devarajan*+ and Xiaowei Chen and Jeffrey Peterson
• Companies: Fox Chase Cancer Center and Fox Chase Cancer Center and Fox Chase Cancer Center
• Keywords: outlier ; technical reproducibility ; coefficient of variation ; asymmetric Laplace model ; lognormal model ; maximum likelihood

Abstract

A problem that arises frequently in large-scale biological studies is the assessment of technical reproducibility of data obtained under homogeneous experimental conditions. This is an important problem considering the tremendous growth in the number of high-throughput technologies that have become available in molecular biology in the past decade. Although some visualization tools for determining the quality of microarray data have been in existence for many years, these methods have various shortcomings and do not necessarily translate to data obtained from other technologies. There is an inherent need for quantitative evaluation of the reproducibility of technical replicates obtained using novel approaches such as high-throughput screening, next-generation sequencing and SNP arrays. In this paper, we develop an empirical model-based approach that accounts for technical variability and potential asymmetry that arise naturally in replicate data, and aids in the identification of outliers. We illustrate our methods using experimental data obtained from the application of several high-throughput technologies in studies involving cancer.