Estimating a positive false discovery rate for variable selection in pharmacogenetic studies

Lang Li, Siu Hui, Gene Pennello, Zeruesenay Desta, Todd Skaar, Anne Nguyen, David Flockhart

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Selecting predictors to optimize the outcome prediction is an important statistical method. However, it usually ignores the false positives in the selected predictors. In this paper, we develop a positive false discovery rate (pFDR) estimate for a conventional step-wise forward variable selection procedure. We propose two views of a variable selection process, an overall and an individual test. An interesting feature of the overall test is that its power of selecting non-null predictors increases with the proportion of non-null predictors among all candidate predictors. Data analysis is illustrated with a pharmacogenetics example.

Original languageEnglish
Pages (from-to)883-902
Number of pages20
JournalJournal of Biopharmaceutical Statistics
Volume17
Issue number5
DOIs
StatePublished - Sep 2007

Fingerprint

Variable Selection
False Positive
Predictors
Pharmacogenetics
Selection Procedures
Statistical method
Data analysis
Proportion
Optimise
Pharmacogenomic Testing
Prediction
Estimate

Keywords

  • Cross-validation
  • False discovery rate
  • Multiple-comparisons
  • Pharmacogenetics
  • Variable selection

ASJC Scopus subject areas

  • Pharmacology (medical)
  • Pharmacology, Toxicology and Pharmaceutics(all)

Cite this

Estimating a positive false discovery rate for variable selection in pharmacogenetic studies. / Li, Lang; Hui, Siu; Pennello, Gene; Desta, Zeruesenay; Skaar, Todd; Nguyen, Anne; Flockhart, David.

In: Journal of Biopharmaceutical Statistics, Vol. 17, No. 5, 09.2007, p. 883-902.

Research output: Contribution to journalArticle

@article{a042b29392d9438db3d0e58778a66cab,
title = "Estimating a positive false discovery rate for variable selection in pharmacogenetic studies",
abstract = "Selecting predictors to optimize the outcome prediction is an important statistical method. However, it usually ignores the false positives in the selected predictors. In this paper, we develop a positive false discovery rate (pFDR) estimate for a conventional step-wise forward variable selection procedure. We propose two views of a variable selection process, an overall and an individual test. An interesting feature of the overall test is that its power of selecting non-null predictors increases with the proportion of non-null predictors among all candidate predictors. Data analysis is illustrated with a pharmacogenetics example.",
keywords = "Cross-validation, False discovery rate, Multiple-comparisons, Pharmacogenetics, Variable selection",
author = "Lang Li and Siu Hui and Gene Pennello and Zeruesenay Desta and Todd Skaar and Anne Nguyen and David Flockhart",
year = "2007",
month = "9",
doi = "10.1080/10543400701514056",
language = "English",
volume = "17",
pages = "883--902",
journal = "Journal of Biopharmaceutical Statistics",
issn = "1054-3406",
publisher = "Taylor and Francis Ltd.",
number = "5",

}

TY - JOUR

T1 - Estimating a positive false discovery rate for variable selection in pharmacogenetic studies

AU - Li, Lang

AU - Hui, Siu

AU - Pennello, Gene

AU - Desta, Zeruesenay

AU - Skaar, Todd

AU - Nguyen, Anne

AU - Flockhart, David

PY - 2007/9

Y1 - 2007/9

N2 - Selecting predictors to optimize the outcome prediction is an important statistical method. However, it usually ignores the false positives in the selected predictors. In this paper, we develop a positive false discovery rate (pFDR) estimate for a conventional step-wise forward variable selection procedure. We propose two views of a variable selection process, an overall and an individual test. An interesting feature of the overall test is that its power of selecting non-null predictors increases with the proportion of non-null predictors among all candidate predictors. Data analysis is illustrated with a pharmacogenetics example.

AB - Selecting predictors to optimize the outcome prediction is an important statistical method. However, it usually ignores the false positives in the selected predictors. In this paper, we develop a positive false discovery rate (pFDR) estimate for a conventional step-wise forward variable selection procedure. We propose two views of a variable selection process, an overall and an individual test. An interesting feature of the overall test is that its power of selecting non-null predictors increases with the proportion of non-null predictors among all candidate predictors. Data analysis is illustrated with a pharmacogenetics example.

KW - Cross-validation

KW - False discovery rate

KW - Multiple-comparisons

KW - Pharmacogenetics

KW - Variable selection

UR - http://www.scopus.com/inward/record.url?scp=34548833067&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34548833067&partnerID=8YFLogxK

U2 - 10.1080/10543400701514056

DO - 10.1080/10543400701514056

M3 - Article

C2 - 17885872

AN - SCOPUS:34548833067

VL - 17

SP - 883

EP - 902

JO - Journal of Biopharmaceutical Statistics

JF - Journal of Biopharmaceutical Statistics

SN - 1054-3406

IS - 5

ER -