Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing

Sang Cheol Kim, Seul Ji Lee, Won Jun Lee, Young Na Yum, Joo Hwan Kim, Soojung Sohn, Jeong Hill Park, Jeongmi Lee, Johan Lim, Sung Won Kwon

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

In microarray data analysis, we are often required to combine several dependent partial test results. To overcome this, many suggestions have been made in previous literature; Tippett's test and Fisher's omnibus test are most popular. Both tests have known null distributions when the partial tests are independent. However, for dependent tests, their (even, asymptotic) null distributions are unknown and additional numerical procedures are required. In this paper, we revisited Stouffer's test based on z-scores and showed its advantage over the two aforementioned methods in the analysis of large-scale microarray data. The combined statistic in Stouffer's test has a normal distribution with mean 0 from the normality of the z-scores. Its variance can be estimated from the scores of genes in the experiment without an additional numerical procedure. We numerically compared the errors of Stouffer's test and the two p-value based methods, Tippett's test and Fisher's omnibus test. We also analyzed our microarray data to find differentially expressed genes by non-genotoxic and genotoxic carcinogen compounds. Both numerical study and the real application showed that Stouffer's test performed better than Tippett's method and Fisher's omnibus method with additional permutation steps.

Original languageEnglish (US)
Article numbere63290
JournalPLoS One
Volume8
Issue number5
DOIs
StatePublished - May 14 2013

Fingerprint

Microarrays
Testing
Genes
testing
Normal distribution
Carcinogens
Statistics
Normal Distribution
Microarray Analysis
Experiments
carcinogens
methodology
data analysis
genes
statistics

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Medicine(all)

Cite this

Kim, S. C., Lee, S. J., Lee, W. J., Yum, Y. N., Kim, J. H., Sohn, S., ... Kwon, S. W. (2013). Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing. PLoS One, 8(5), [e63290]. https://doi.org/10.1371/journal.pone.0063290

Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing. / Kim, Sang Cheol; Lee, Seul Ji; Lee, Won Jun; Yum, Young Na; Kim, Joo Hwan; Sohn, Soojung; Park, Jeong Hill; Lee, Jeongmi; Lim, Johan; Kwon, Sung Won.

In: PLoS One, Vol. 8, No. 5, e63290, 14.05.2013.

Research output: Contribution to journalArticle

Kim, SC, Lee, SJ, Lee, WJ, Yum, YN, Kim, JH, Sohn, S, Park, JH, Lee, J, Lim, J & Kwon, SW 2013, 'Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing', PLoS One, vol. 8, no. 5, e63290. https://doi.org/10.1371/journal.pone.0063290
Kim SC, Lee SJ, Lee WJ, Yum YN, Kim JH, Sohn S et al. Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing. PLoS One. 2013 May 14;8(5). e63290. https://doi.org/10.1371/journal.pone.0063290
Kim, Sang Cheol ; Lee, Seul Ji ; Lee, Won Jun ; Yum, Young Na ; Kim, Joo Hwan ; Sohn, Soojung ; Park, Jeong Hill ; Lee, Jeongmi ; Lim, Johan ; Kwon, Sung Won. / Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing. In: PLoS One. 2013 ; Vol. 8, No. 5.
@article{8bdb248fb6f64959b0f7d90b9e6c1220,
title = "Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing",
abstract = "In microarray data analysis, we are often required to combine several dependent partial test results. To overcome this, many suggestions have been made in previous literature; Tippett's test and Fisher's omnibus test are most popular. Both tests have known null distributions when the partial tests are independent. However, for dependent tests, their (even, asymptotic) null distributions are unknown and additional numerical procedures are required. In this paper, we revisited Stouffer's test based on z-scores and showed its advantage over the two aforementioned methods in the analysis of large-scale microarray data. The combined statistic in Stouffer's test has a normal distribution with mean 0 from the normality of the z-scores. Its variance can be estimated from the scores of genes in the experiment without an additional numerical procedure. We numerically compared the errors of Stouffer's test and the two p-value based methods, Tippett's test and Fisher's omnibus test. We also analyzed our microarray data to find differentially expressed genes by non-genotoxic and genotoxic carcinogen compounds. Both numerical study and the real application showed that Stouffer's test performed better than Tippett's method and Fisher's omnibus method with additional permutation steps.",
author = "Kim, {Sang Cheol} and Lee, {Seul Ji} and Lee, {Won Jun} and Yum, {Young Na} and Kim, {Joo Hwan} and Soojung Sohn and Park, {Jeong Hill} and Jeongmi Lee and Johan Lim and Kwon, {Sung Won}",
year = "2013",
month = "5",
day = "14",
doi = "10.1371/journal.pone.0063290",
language = "English (US)",
volume = "8",
journal = "PLoS One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "5",

}

TY - JOUR

T1 - Stouffer's Test in a Large Scale Simultaneous Hypothesis Testing

AU - Kim, Sang Cheol

AU - Lee, Seul Ji

AU - Lee, Won Jun

AU - Yum, Young Na

AU - Kim, Joo Hwan

AU - Sohn, Soojung

AU - Park, Jeong Hill

AU - Lee, Jeongmi

AU - Lim, Johan

AU - Kwon, Sung Won

PY - 2013/5/14

Y1 - 2013/5/14

N2 - In microarray data analysis, we are often required to combine several dependent partial test results. To overcome this, many suggestions have been made in previous literature; Tippett's test and Fisher's omnibus test are most popular. Both tests have known null distributions when the partial tests are independent. However, for dependent tests, their (even, asymptotic) null distributions are unknown and additional numerical procedures are required. In this paper, we revisited Stouffer's test based on z-scores and showed its advantage over the two aforementioned methods in the analysis of large-scale microarray data. The combined statistic in Stouffer's test has a normal distribution with mean 0 from the normality of the z-scores. Its variance can be estimated from the scores of genes in the experiment without an additional numerical procedure. We numerically compared the errors of Stouffer's test and the two p-value based methods, Tippett's test and Fisher's omnibus test. We also analyzed our microarray data to find differentially expressed genes by non-genotoxic and genotoxic carcinogen compounds. Both numerical study and the real application showed that Stouffer's test performed better than Tippett's method and Fisher's omnibus method with additional permutation steps.

AB - In microarray data analysis, we are often required to combine several dependent partial test results. To overcome this, many suggestions have been made in previous literature; Tippett's test and Fisher's omnibus test are most popular. Both tests have known null distributions when the partial tests are independent. However, for dependent tests, their (even, asymptotic) null distributions are unknown and additional numerical procedures are required. In this paper, we revisited Stouffer's test based on z-scores and showed its advantage over the two aforementioned methods in the analysis of large-scale microarray data. The combined statistic in Stouffer's test has a normal distribution with mean 0 from the normality of the z-scores. Its variance can be estimated from the scores of genes in the experiment without an additional numerical procedure. We numerically compared the errors of Stouffer's test and the two p-value based methods, Tippett's test and Fisher's omnibus test. We also analyzed our microarray data to find differentially expressed genes by non-genotoxic and genotoxic carcinogen compounds. Both numerical study and the real application showed that Stouffer's test performed better than Tippett's method and Fisher's omnibus method with additional permutation steps.

UR - http://www.scopus.com/inward/record.url?scp=84877782583&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84877782583&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0063290

DO - 10.1371/journal.pone.0063290

M3 - Article

C2 - 23691011

AN - SCOPUS:84877782583

VL - 8

JO - PLoS One

JF - PLoS One

SN - 1932-6203

IS - 5

M1 - e63290

ER -