A Bayesian discovery procedure

Michele Guindani, Peter Müller, Song Zhang

Research output: Contribution to journalArticle

29 Citations (Scopus)

Abstract

We discuss a Bayesian discovery procedure for multiple-comparison problems. We show that, under a coherent decision theoretic framework, a loss function combining true positive and false positive counts leads to a decision rule that is based on a threshold of the posterior probability of the alternative. Under a semiparametric model for the data, we show that the Bayes rule can be approximated by the optimal discovery procedure, which was recently introduced by Storey. Improving the approximation leads us to a Bayesian discovery procedure, which exploits the multiple shrinkage in clusters that are implied by the assumed non-parametric model. We compare the Bayesian discovery procedure and the optimal discovery procedure estimates in a simple simulation study and in an assessment of differential gene expression based on microarray data from tumour samples. We extend the setting of the optimal discovery procedure by discussing modifications of the loss function that lead to different single-thresholding statistics. Finally, we provide an application of the previous arguments to dependent (spatial) data.

Original languageEnglish (US)
Pages (from-to)905-925
Number of pages21
JournalJournal of the Royal Statistical Society. Series B: Statistical Methodology
Volume71
Issue number5
DOIs
StatePublished - Nov 2009

Fingerprint

Loss Function
Bayes Rule
Multiple Comparisons
Differential Expression
Semiparametric Model
Nonparametric Model
Dependent Data
Posterior Probability
Spatial Data
Decision Rules
Thresholding
Shrinkage
Microarray Data
False Positive
Gene Expression
Tumor
Count
Simulation Study
Statistics
Alternatives

Keywords

  • Bayes optimal rule
  • False discovery rate
  • Loss function
  • Multiple comparison

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

A Bayesian discovery procedure. / Guindani, Michele; Müller, Peter; Zhang, Song.

In: Journal of the Royal Statistical Society. Series B: Statistical Methodology, Vol. 71, No. 5, 11.2009, p. 905-925.

Research output: Contribution to journalArticle

Guindani, Michele ; Müller, Peter ; Zhang, Song. / A Bayesian discovery procedure. In: Journal of the Royal Statistical Society. Series B: Statistical Methodology. 2009 ; Vol. 71, No. 5. pp. 905-925.
@article{6db1a8716fdb42cea45e426d2e747478,
title = "A Bayesian discovery procedure",
abstract = "We discuss a Bayesian discovery procedure for multiple-comparison problems. We show that, under a coherent decision theoretic framework, a loss function combining true positive and false positive counts leads to a decision rule that is based on a threshold of the posterior probability of the alternative. Under a semiparametric model for the data, we show that the Bayes rule can be approximated by the optimal discovery procedure, which was recently introduced by Storey. Improving the approximation leads us to a Bayesian discovery procedure, which exploits the multiple shrinkage in clusters that are implied by the assumed non-parametric model. We compare the Bayesian discovery procedure and the optimal discovery procedure estimates in a simple simulation study and in an assessment of differential gene expression based on microarray data from tumour samples. We extend the setting of the optimal discovery procedure by discussing modifications of the loss function that lead to different single-thresholding statistics. Finally, we provide an application of the previous arguments to dependent (spatial) data.",
keywords = "Bayes optimal rule, False discovery rate, Loss function, Multiple comparison",
author = "Michele Guindani and Peter M{\"u}ller and Song Zhang",
year = "2009",
month = "11",
doi = "10.1111/j.1467-9868.2009.00714.x",
language = "English (US)",
volume = "71",
pages = "905--925",
journal = "Journal of the Royal Statistical Society. Series B: Statistical Methodology",
issn = "1369-7412",
publisher = "Wiley-Blackwell",
number = "5",

}

TY - JOUR

T1 - A Bayesian discovery procedure

AU - Guindani, Michele

AU - Müller, Peter

AU - Zhang, Song

PY - 2009/11

Y1 - 2009/11

N2 - We discuss a Bayesian discovery procedure for multiple-comparison problems. We show that, under a coherent decision theoretic framework, a loss function combining true positive and false positive counts leads to a decision rule that is based on a threshold of the posterior probability of the alternative. Under a semiparametric model for the data, we show that the Bayes rule can be approximated by the optimal discovery procedure, which was recently introduced by Storey. Improving the approximation leads us to a Bayesian discovery procedure, which exploits the multiple shrinkage in clusters that are implied by the assumed non-parametric model. We compare the Bayesian discovery procedure and the optimal discovery procedure estimates in a simple simulation study and in an assessment of differential gene expression based on microarray data from tumour samples. We extend the setting of the optimal discovery procedure by discussing modifications of the loss function that lead to different single-thresholding statistics. Finally, we provide an application of the previous arguments to dependent (spatial) data.

AB - We discuss a Bayesian discovery procedure for multiple-comparison problems. We show that, under a coherent decision theoretic framework, a loss function combining true positive and false positive counts leads to a decision rule that is based on a threshold of the posterior probability of the alternative. Under a semiparametric model for the data, we show that the Bayes rule can be approximated by the optimal discovery procedure, which was recently introduced by Storey. Improving the approximation leads us to a Bayesian discovery procedure, which exploits the multiple shrinkage in clusters that are implied by the assumed non-parametric model. We compare the Bayesian discovery procedure and the optimal discovery procedure estimates in a simple simulation study and in an assessment of differential gene expression based on microarray data from tumour samples. We extend the setting of the optimal discovery procedure by discussing modifications of the loss function that lead to different single-thresholding statistics. Finally, we provide an application of the previous arguments to dependent (spatial) data.

KW - Bayes optimal rule

KW - False discovery rate

KW - Loss function

KW - Multiple comparison

UR - http://www.scopus.com/inward/record.url?scp=70350100452&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350100452&partnerID=8YFLogxK

U2 - 10.1111/j.1467-9868.2009.00714.x

DO - 10.1111/j.1467-9868.2009.00714.x

M3 - Article

VL - 71

SP - 905

EP - 925

JO - Journal of the Royal Statistical Society. Series B: Statistical Methodology

JF - Journal of the Royal Statistical Society. Series B: Statistical Methodology

SN - 1369-7412

IS - 5

ER -