Estimating the proportion of equivalently expressed genes in microarray data based on transformed test statistics

Shuo Jiao, Shunpu Zhang

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

In microarray data analysis, false discovery rate (FDR) is now widely accepted as the control criterion to account for multiple hypothesis testing. The proportion of equivalently expressed genes (π0) is a key component to be estimated in the estimation of FDR. Some commonly used π0 estimators (BUM, SPLOSH, QVALUE, and LBE ) are all based on p-values, and they are essentially upper bounds of π0. The simulations we carried out show that these four methods significantly overestimate the true π0 when differentially expressed genes and equivalently expressed genes are not well separated. To solve this problem, we first introduce a novel way of transforming the test statistics to make them symmetric about 0. Then we propose a π0 estimator based on the transformed test statistics using the symmetry assumption. Real data application and simulation both show that the π0 estimate from our method is less conservative than BUM, SPLOSH, QVALUE, and LBE in most of the cases. Simulation results also show that our estimator always has the least mean squared error among these five methods.

Original languageEnglish (US)
Pages (from-to)177-187
Number of pages11
JournalJournal of Computational Biology
Volume17
Issue number2
DOIs
StatePublished - Feb 1 2010

Fingerprint

Microarrays
Microarray Data
Test Statistic
Proportion
Genes
Statistics
Gene
Estimator
Multiple Hypothesis Testing
Microarray Data Analysis
Simulation
p-Value
Microarray Analysis
Mean Squared Error
Upper bound
Symmetry
Testing
Estimate
False

Keywords

  • Gene expression analysis
  • Microarray
  • Proportion of null hypothesis (π)
  • Transformed test statistics

ASJC Scopus subject areas

  • Modeling and Simulation
  • Molecular Biology
  • Genetics
  • Computational Mathematics
  • Computational Theory and Mathematics

Cite this

Estimating the proportion of equivalently expressed genes in microarray data based on transformed test statistics. / Jiao, Shuo; Zhang, Shunpu.

In: Journal of Computational Biology, Vol. 17, No. 2, 01.02.2010, p. 177-187.

Research output: Contribution to journalArticle

@article{d885d88a9d694e5d947f8ad8357b790d,
title = "Estimating the proportion of equivalently expressed genes in microarray data based on transformed test statistics",
abstract = "In microarray data analysis, false discovery rate (FDR) is now widely accepted as the control criterion to account for multiple hypothesis testing. The proportion of equivalently expressed genes (π0) is a key component to be estimated in the estimation of FDR. Some commonly used π0 estimators (BUM, SPLOSH, QVALUE, and LBE ) are all based on p-values, and they are essentially upper bounds of π0. The simulations we carried out show that these four methods significantly overestimate the true π0 when differentially expressed genes and equivalently expressed genes are not well separated. To solve this problem, we first introduce a novel way of transforming the test statistics to make them symmetric about 0. Then we propose a π0 estimator based on the transformed test statistics using the symmetry assumption. Real data application and simulation both show that the π0 estimate from our method is less conservative than BUM, SPLOSH, QVALUE, and LBE in most of the cases. Simulation results also show that our estimator always has the least mean squared error among these five methods.",
keywords = "Gene expression analysis, Microarray, Proportion of null hypothesis (π), Transformed test statistics",
author = "Shuo Jiao and Shunpu Zhang",
year = "2010",
month = "2",
day = "1",
doi = "10.1089/cmb.2009.0060",
language = "English (US)",
volume = "17",
pages = "177--187",
journal = "Journal of Computational Biology",
issn = "1066-5277",
publisher = "Mary Ann Liebert Inc.",
number = "2",

}

TY - JOUR

T1 - Estimating the proportion of equivalently expressed genes in microarray data based on transformed test statistics

AU - Jiao, Shuo

AU - Zhang, Shunpu

PY - 2010/2/1

Y1 - 2010/2/1

N2 - In microarray data analysis, false discovery rate (FDR) is now widely accepted as the control criterion to account for multiple hypothesis testing. The proportion of equivalently expressed genes (π0) is a key component to be estimated in the estimation of FDR. Some commonly used π0 estimators (BUM, SPLOSH, QVALUE, and LBE ) are all based on p-values, and they are essentially upper bounds of π0. The simulations we carried out show that these four methods significantly overestimate the true π0 when differentially expressed genes and equivalently expressed genes are not well separated. To solve this problem, we first introduce a novel way of transforming the test statistics to make them symmetric about 0. Then we propose a π0 estimator based on the transformed test statistics using the symmetry assumption. Real data application and simulation both show that the π0 estimate from our method is less conservative than BUM, SPLOSH, QVALUE, and LBE in most of the cases. Simulation results also show that our estimator always has the least mean squared error among these five methods.

AB - In microarray data analysis, false discovery rate (FDR) is now widely accepted as the control criterion to account for multiple hypothesis testing. The proportion of equivalently expressed genes (π0) is a key component to be estimated in the estimation of FDR. Some commonly used π0 estimators (BUM, SPLOSH, QVALUE, and LBE ) are all based on p-values, and they are essentially upper bounds of π0. The simulations we carried out show that these four methods significantly overestimate the true π0 when differentially expressed genes and equivalently expressed genes are not well separated. To solve this problem, we first introduce a novel way of transforming the test statistics to make them symmetric about 0. Then we propose a π0 estimator based on the transformed test statistics using the symmetry assumption. Real data application and simulation both show that the π0 estimate from our method is less conservative than BUM, SPLOSH, QVALUE, and LBE in most of the cases. Simulation results also show that our estimator always has the least mean squared error among these five methods.

KW - Gene expression analysis

KW - Microarray

KW - Proportion of null hypothesis (π)

KW - Transformed test statistics

UR - http://www.scopus.com/inward/record.url?scp=77249128449&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77249128449&partnerID=8YFLogxK

U2 - 10.1089/cmb.2009.0060

DO - 10.1089/cmb.2009.0060

M3 - Article

C2 - 20078228

AN - SCOPUS:77249128449

VL - 17

SP - 177

EP - 187

JO - Journal of Computational Biology

JF - Journal of Computational Biology

SN - 1066-5277

IS - 2

ER -