Resampling-based multiple testing examples and methods for p-value adjustment pdf

The univariate p value was adjusted for multiple testing by a resampling method westfall and young 1993. Michael wolf is assistant professor, department of economics and business, university of pompeau fabra, e08005, barcelona, spain. This document describes the algorithm employed by wyoung. Pointwise testing with functional data using the westfall. Examples and methods for p value adjustment 1st edition. Exact and approximate stepdown methods for multiple. The stepdown minp has been shown to perform well when compared to other resampling procedures 26. However, the power and stability properties of these popular resamplingbased multiple testing procedures have not been extensively evaluated. Software from sas institute is available to execute many of the methods and programming is straightforward for. Westfall,peterhandsstanleyyoung,resamplingbased multiple testing. Application to investigation of antiretroviral drug susceptibility yang yang program of biostatistics and biomathematics, division of public health sciences, fred hutchinson cancer research center, seattle, washington 98109, u.

Resamplingbased multiple testing for microarray data analysis yongchao ge1, sandrine dudoit2, and terence p. Based multiple testing methods with covariate adjustment. Resampling is a general term that encompasses bootstrap, permutation, and parametric simulationbased analyses. Software from sas institute is available to execute many of the methods and programming is straightforward for other applications. Resamplingbased multiple comparison procedure with. In contrast to the proposals of lehmann and romano ann. Find all the books, read about the author, and more. These could be nominal pvalues, for example, from ttables, or permutation pvalues as. Multiple testing is the subject of the present paper. This function computes adjusted pvalues for simple multiple testing procedures from a vector of. The more accurate m eff can be estimated, the more precise will be the p value adjustment.

We have also made additions to the output, added an option to produce a graph, and included support for the predict command. Although this is by no means a new subject in the statistical. Examples and methods for p value adjustment, published by wiley. Examples and methods for p value adjustment wiley series in probability and statistics by westfall, young isbn. Using resampling techniques in proc multitest to evaluate surgeon specif ic results following coronary artery bypass graft cabg. Resamplingbased confidence regions and multiple tests for. The following table defines the possible outcomes when testing multiple null hypotheses. The major revisions involve improvements to the estimation methods and the addition of an option to use a permutation test to estimate pvalues, including an adjustment for multiple testing. Department of statistics, university of california, berkeley 2. The author demonstrates how easy it is to carry out resampling methods using the currently available software from the sasr institute proc multtest, as well as many other types of. Exact and approximate stepdown methods for multiple hypothesis testing joseph p romano joseph p. Division of biostatistics, university of california, berkeley 3. This cited by count includes citations to the following articles in scholar. Resamplingbased multiple testing procedures are widely used in genomic studies to identify differentially expressed genes and to conduct genomewide association studies.

The bonferroni correction sets the signi cance cuto at n. Examples and methods for p value adjustment new york. Adjusted pvalues for simple multiple testing procedures. Millers classic text, simultaneous statistical inference 1981, beautifully describes the development of multiple testing before the era of largescale data sets. Examples and methods for p value adjustment, wiley.

Furthermore, all methods for calculating pvalues or for performing multiple testing correction assume a valid statistical modeleither analytic or empiricalthat captures dependencies in. Romano is professor, department of statistics, stanford university, stanford, ca 94305. Wolfinger 2011 multiple comparisons and multiple testing using sas, 2nd edn, sas institute. Hommels method is more powerful than hochbergs, but the difference is usually small and the hochberg pvalues are faster to compute. These functions compute permutation adjusted pvalues for the stepdown maxt and minp multiple testing procedures. A range of methods to adjust for multiple testing are available see for an overview. The method of westfall and young 1993 is used as a standard for comparison with the proposed methods. Examples and methods for p value adjustment article in technometrics 354. Stepdown multiple testing procedures are generally based on a set of null resampling test statistics t. Methods for dealing with multiple testing frequently call for adjusting in some way, so. Resamplingbased multiple testing for microarray data analysis. The closure principle isincorporated along withthe new.

Suppose we have a number m of null hypotheses, denoted by. The burgeoning field of genomics has revived interest in multiple testing procedures by raising new methodological and computational challenges. The original papers only describe how to carry out multiple testing at a fixed significance level. Methods to adjust for multiple comparisons in the analysis. Examples and methods for pvalue adjustment, published by wiley. Examples and methods for p value adjustment, wiley, new york, 1993. Although, the resampling based methods have previously been recommended for clinical trials with multiple outcomes they are not widely used in practice. Using a statistical test, we reject the null hypothesis if the test is declared significant. Resamplingbased multiple testing for microarray data. Examples and methods for p value adjustment, wiley b. Getting started in gene expression microarray analysis. Everyday low prices and free delivery on eligible orders. Stanley young are the authors of resamplingbased multiple testing. Choice of a null distribution in resamplingbased multiple.

Depending on context, the resampling can be carried out by a bootstrap method, a permutation method, or a randomization method. Efficient computation of adjusted pvalues for resampling. Calculating exact pvalues when permutations are randomly drawn, statistical applications in genetics and molecular biology vol 9 iss. Hochbergs and hommels methods are valid when the hypothesis tests are independent or when they are nonnegatively associated sarkar, 1998. We consider hypothesis testing with smooth functional data by performing pointwise tests and applying a multiple comparisons procedure. Resamplingbased multiple testing for microarray data analysis yongchao ge1, sandrine dudoit2. Gene expression is a prototype application, but the applications are much broader. Permutation tests are considered the gold standard for multiple testing correction in gwas. One important alternative approach is to estimate the effective number of independent tests m eff cheverud, 2001 for a set of m markers, and use this number for an improved bonferroni correction. This condition is the formal analogue of the subset pivotality condition westfall and young, resamplingbased multiple testing. A gallery of examples of implausible correlations sourced by. The biological question of differential expression can be restated as a problem in multiple hypothesis testing. P value adjustments for multiple tests in multivariate binomial models. For exampl westfall and young 1993 propose resamplingbasedpvalue adjustment procedures which are highly relevant to microarray experiments.

Analysing the effect of candidate genes on complex traits. For example, in the example above, with 20 tests and 0. Smyth 2010, permutation pvalues should never be zero. Miller rg 1981 simultaneous statistical inference, 2 nd. Combines recent developments in resampling technology including the bootstrap with new methods for multiple testing that are easy to use, convenient to report and widely applicable. We demonstrate the use of our proposed ordinary bootstrap null distribution with a singlestep multiple testing method which is equivalent to.