|
Activity Number:
|
536
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308940 |
|
Title:
|
Covariate-Modulated False Discovery Rates
|
|
Author(s):
|
Egil Ferkingstad*+ and Arnoldo Frigessi and Gudmar Thorleifsson and Augustine Kong
|
|
Companies:
|
University of Oslo and University of Oslo and deCODE genetics and deCODE genetics
|
|
Address:
|
Department of Biostatistics, Oslo, N-0317, Norway
|
|
Keywords:
|
multiple testing ; bioinformatics ; data integration ; genomics ; false discovery rate ; empirical Bayes
|
|
Abstract:
|
In an empirical Bayes setting, the local false discovery rate (local FDR) is defined as the posterior probability that the null hypothesis is true given data. We extend this methodology further and introduce the covariate-modulated false discovery rate (cmFDR), useful when an additional covariate is available that influences the probability of each null hypothesis being true. cmFDR measures the posterior significance of each test conditionally on the covariate and the data, leading to greater power. The cmFDR uses covariate-based prior information to produce a list of significant hypotheses which differs in length and order from the list obtained by the local FDR. We estimate the cmFDR with MCMC for an approximate model on p-values. The new method is applied to expression quantitative trait loci (eQTL) data, and to gene expressions modulated by copy number alterations in breast cancer.
|
