fdrtool: Estimation of (Local) False Discovery Rates and Higher Criticism

This package allows to estimate both tail area-based false discovery rates (Fdr) as well as local false discovery rates (fdr) for a variety of null models (p-values, z-scores, correlation coefficients, t-scores). The proportion of null values and the parameters of the null distribution are adaptively estimated from the data. In addition, the package contains functions for non-parametric density estimation (Grenander estimator), for monotone regression (isotonic regression and antitonic regression with weights), for computing the greatest convex minorant (GCM) and the least concave majorant (LCM), for the half-normal and correlation distributions, and for computing empirical higher criticism (HC) scores and the corresponding decision threshold.

Version: 1.2.13
Depends: R (≥ 2.15.1)
Published: 2014-11-14
Author: Bernd Klaus and Korbinian Strimmer.
Maintainer: Korbinian Strimmer <strimmerlab at gmail.com>
License: GPL (≥ 3)
URL: http://strimmerlab.org/software/fdrtool/
NeedsCompilation: yes
Materials: NEWS
CRAN checks: fdrtool results

Downloads:

Reference manual: fdrtool.pdf
Package source: fdrtool_1.2.13.tar.gz
Windows binaries: r-devel: fdrtool_1.2.13.zip, r-release: fdrtool_1.2.13.zip, r-oldrel: fdrtool_1.2.13.zip
OS X Snow Leopard binaries: r-release: fdrtool_1.2.13.tgz, r-oldrel: fdrtool_1.2.13.tgz
OS X Mavericks binaries: r-release: fdrtool_1.2.13.tgz
Old sources: fdrtool archive

Reverse dependencies:

Reverse depends: DBGSA, ERP, GeneCycle, GeneNet, sda, st, TBEST
Reverse imports: cape, SimSeq
Reverse suggests: frontier, mutoss, qgraph, sand