The use in Biology of current high-throughput genetic and post-genomic data leads to the simultaneous evaluation of a huge number of statistical hypothesis and at the same time, to the multiple-testing problem. As an alternative to the too conservative Family-Wise Error-Rate (FWER), the False Discovery Rate (FDR) has appeared for the last ten years as the more appropriate criterion to handle the problem. One drawback is that the FDR is associated to a given rejection region for the statistic considered, without distinguishing those that are close to the boundary and those that are not.

As a result, the local FDR has been recently proposed to quantifies the specific probability, given the p-value, for being under the null hypothesis. In this context we present a semi-parametric approach based on kernel estimators.

• kerfdr: source, package, RDoc, pseudo-code
• kerfdr on CRAN

Publications

• kerfdr: A semi-parametric kernel-based approach to local FDR estimations. Guedj, Célisse, Robin and Nuel. BMC Bioinformatics. 2009. 10:84. link

• A semi-parametric approach for mixture models: application to local false discovery rate estimation. Robin, Bar-Hen et al. Comput. Statist. and Data Analysis. 2007. 51: 5483-5493. abstract

Talk

• kerfdr: A semi-parametric kernel-based approach to local FDR estimation. Guedj et al. Statistical Methods for Post-Genomics Data 2008, Rennes. slides

Poster

• kerfdr: A semi-parametric kernel-based approach for local FDR estimations. Nuel, Guedj et al. IGES 2007, York (UK). poster

This work is the result of the SSB group.