Hi all, The LPE (Local Pooled Error) package has 2 methods to compute the FDR correction for multiple testing : the Benjamini-Hochberg (BH) method, and the resampling based FDR adjustment. Can somebody advise me on which one to choose ? Is one always the best, or does it depend on circumstances, like the sample size or something else ? In a test run of two 2 groups and 3 samples/group, the resampling method gives 2x more differentially expressed genes than the BH method, at FDR=1%. Jean-Louis ==================================================== Jean-Louis Ruelle, PhD GlaxoSmithKline Biologicals, R&D Rue de l'Institut, 89 1330 Rixensart Belgium Phone : + 32-2-6568451 Fax : +32-2-6568436 email : jean-louis.ruelle@gskbio.com The information contained in this message is confidential an...{{dropped}}

Hi, In a recent paper submitted to BMC bioinformatics, we show that resampling based FDR adjustment method is better than BH method for small sample size. For small sample sizes, BH is quite conservative as shown in our simulation study. Will put the copy of the submitted paper at: http://www.healthsystem.virginia.edu/internet/hes/biostat/bioinformati cs/Sof tware/LPE/index.cfm Best, -Nitin ______________________ Nitin Jain, PhD Non Clinical Statistics Pfizer, Inc. (Groton, CT) Bldg: 260, # 1451 Ph: (860) 686-2526 (Office) Fax: (860) 686-6170 -----Original Message----- From: bioconductor-bounces@stat.math.ethz.ch [mailto:bioconductor-bounces@stat.math.ethz.ch]On Behalf Of jean-louis.ruelle@gskbio.com Sent: Monday, May 23, 2005 8:52 AM To: bioconductor@stat.math.ethz.ch Subject: [BioC] FDR adjustment in LPE Hi all, The LPE (Local Pooled Error) package has 2 methods to compute the FDR correction for multiple testing : the Benjamini-Hochberg (BH) method, and the resampling based FDR adjustment. Can somebody advise me on which one to choose ? Is one always the best, or does it depend on circumstances, like the sample size or something else ? In a test run of two 2 groups and 3 samples/group, the resampling method gives 2x more differentially expressed genes than the BH method, at FDR=1%. Jean-Louis ==================================================== Jean-Louis Ruelle, PhD GlaxoSmithKline Biologicals, R&D Rue de l'Institut, 89 1330 Rixensart Belgium Phone : + 32-2-6568451 Fax : +32-2-6568436 email : jean-louis.ruelle@gskbio.com The information contained in this message is confidential an...{{dropped}} _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/bioconductor LEGAL NOTICE\ Unless expressly stated otherwise, this messag...{{dropped}}