robust.fdr: "robust FDR" estimation

Description

Implements robust method of FDR estimation (Pounds and Cheng 2006, Bioinformatics)

robust.fdr(p, sides = 1, p2 = 1 - p, discrete = F, use8 = T)

vector of p-values from the analysis.

sides

indicate whether p-values are 1-sided (set sides=1) or 2-sided (set sides=2), default=1.

for one-sided testing, p-values from testing the "other alternative", default=1-p.

discrete

logical. Indicates whether p-values are discrete

use8

indicates whether the constant 8 should be used if p-values are discrete, see Pounds and Cheng (2006) for more details.

p: the vector of p-values provided by the user.
fdr: the vector of smoothed FDR estimates.
q: the vector of q-values based on the smoothed FDR estimates.
cdf: the vector with p-value empirical distribution function at corresponding entry of p.
loc.fdr: the local (unsmoothed) FDR estimates.
fp: the estimated number of false positives at p-value cutoff in p.
fn: the estimated number of false negatives at p-value cutoff in p.
te: the total of fp and fn.
pi: the null proportion estimate.
ord: a vector of indices to order the vectors above by ascending p-value.

This function uses the code from Stan Pounds available at http://www.stjuderesearch.org/depts/biostats/documents/robust-fdr.R and is included in prot2D package for convenience and comparison purpose.

Pounds, S. & Cheng, C. (2006) "Robust estimation of the false discovery rate" Bioinformatics, vol. 22 (16): 1979-1987.