ccFDR: Conjunctional conditional false discovery rate

In alexploner/condFDR: Conditional and conjunctional false discovery rate

Description

Estimate the conjunctional conditional false discovery rate for two set of p-values, conditional on each other, as measure for pleiotropy.

Usage

ccFDR(data, p1, p2, p_threshold = 0.001, mc.cores = 1)

Arguments

data	data frame or (numeric) matrix with p-values as columns
p1	p-values for which the ccFDR will be estimated; either the name/index of the columns in data, or a vectors of p-values
p2	the p-values on which the ccFDR will be conditioned on; either name or vector of p-values data, or a vectors of p-values
p_threshold	cutoff for pre-filtering the p-values; a vector of either length one (in which case both sets of pvalues have the same threshold) or length two (where the thresholds refer to p1 and p2 in that order).
mc.cores	number of cores to use for parallel calculation; defaults to one (i.e. no parallel calculation), but should absolutely be increased if your system supports it, as this will speed up execution very nicely.

Value

A data frame with columns cFDR1, cFDR2, and ccFDR: if data was specified, the columns are simply added at the end; if only p1 and p2 were specified, a data frame with five columns (the original p-values, plus the cFDRs and the ccFDR).

See Also

mclapply for details on parallel calculations and mc.cores

Examples

data(psynth)
res = ccFDR(psynth, "p1", "p2", p_threshold = 1E-5)
head(res)
head(subset(res, ccFDR < 0.01))

alexploner/condFDR documentation built on Dec. 31, 2020, 7:43 p.m.