An Improved Variance-Smoothing Method for Testing Differential Expression in Small-Sample-Size Affymetrix Oligonucleotide Microarrays

• Activity Number: 70
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Parul Gulati*+ and David Jarjoura and Soledad Fernandez and Lianbo Yu and Michael Pennell
• Companies: The Ohio State University and The Ohio State University and The Ohio State University and The Ohio State University and The Ohio State University
• Keywords: Oligonucleotide ; microarrays ; hierarchical ; empirical Bayes ; coefficient of variation ; t-statistic

Abstract

Oligonucleotide microarray experiments with few replications lead to great variability in estimates of residual variance across genes. Several hierarchical Bayesian methods have been used to reduce this variability and to increase power. A recently developed method incorporates the relationship between the gene expression and the variance of gene expression into an empirical Bayes approach. This generates a modified t-statistic, which outperforms methods that do not smooth variances. However this method assumes a constant coefficient of variation (CV) for the residual variances. We provide evidence and arguments against this assumption, and extend the method by allowing the CV to vary with gene expression. The superior performance of the extended method is demonstrated using simulated and spike-in data.