Online Program Home
My Program

Abstract Details

Activity Number: 133 - Gene-Set Based Analysis in Genomic Studies
Type: Contributed
Date/Time: Monday, July 30, 2018 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics in Genomics and Genetics
Abstract #330837
Title: Resampling-Based Control of the False Discovery Rate Incorporating Shrinkage Estimation for the Covariance Matrix
Author(s): Josephine Sarpong Akosa* and Melinda McCann
Companies: Oklahoma State University and Oklahoma State University
Keywords: false discovery rate; multiple comparison; simultaneous inference; shrinkage estimators; resampling

High-throughput gene expression experiments such as microarray experiments involve statistically testing thousands of hypotheses simultaneously to identify genes that are differentially expressed. An unguarded use of single-inference procedures for such analyses inflates the overall type I error rates. Correlation between genes and across arrays further complicates this problem. Even when the overall error rate is controlled, correlation can greatly influence the genes deemed differentially expressed. Utilizing an empirical null distribution for simultaneous inference can mitigate the correlation effect. In addition, one needs to choose a suitable test statistic such that even though all measures on a gene are condensed into one number, relevant information is not lost with respect to the test of interest. Motivated by these and the limitations of existing multiple testing procedures, this study proposes a resampling-based procedure that incorporates the estimation of the empirical null distribution and a shrinkage estimation for the variance components. The estimated variance components are later used in the computation of a suitable test statistic.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program