|
Activity Number:
|
506
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #304324 |
|
Title:
|
Bayesian Optimal Discovery Procedure for Simultaneous Significance Testing
|
|
Author(s):
|
Jing Cao*+ and Xian-Jin Xie and Song Zhang
|
|
Companies:
|
Southern Methodist University and The University of Texas Southwestern Medical Center at Dallas and The University of Texas Southwestern Medical Center
|
|
Address:
|
5551 Monticello Av., Dallas, TX, 75206,
|
|
Keywords:
|
Bayesian hierarchical model ; high throughput screening ; optimal discovery procedure ; shrinkage
|
|
Abstract:
|
In high throughput screening (HTS), such as genome-wide RNAi screening, the standard p-value approach needs to be modified as often the number of statistical comparisons is far bigger than that of biological replicates. Some current approaches demonstrate that test statistics with shrinking variances have more power over the traditional t statistic. We propose a Bayesian hierarchical model to incorporate the shrinkage concept in the estimation of the gene variance by introducing a mixture structure on the variance. We show that the performance of the Bayesian mixture model is comparable to the commonly used frequentist test statistics. The estimates from the Bayesian model are further implemented in the optimal discovery procedure (ODP) proposed in Storey (2007). We illustrate that the Bayesian model-based ODP produces a more thorough list of hits in HTS.
|
