fdr_p: Internal functions for p-value and FDR estimation

In randRotation: Random Rotation Methods for High Dimensional Data with Batch Structure

Description

Internal functions for p-value and FDR estimation

Usage

.fdr.qu(
  s0,
  stats,
  beta = 0.05,
  na.rm = FALSE,
  ref.vector = sort(s0, decreasing = TRUE, na.last = TRUE)
)

.fdr.q(
  s0,
  stats,
  beta = 0.05,
  na.rm = FALSE,
  ref.vector = sort(s0, decreasing = TRUE, na.last = TRUE)
)

.pFdr(s0, stats, method, pooled, na.rm, beta)

Arguments

s0	numeric vector of original (non-rotated) statistics.
stats	numeric matrix of rotated statistics.
beta	numeric between 0 and 1. See \insertCiteYekutieli1999randRotation.
na.rm	logical. Should missing values be removed ?
ref.vector	Reference vector defining at which grid points of s0 and (stats) the FDRs are approximated. All other points are approximated by spline interpolation. NAs are removed from ref.vector
method	A p-value or FDR adjustment method, see pFdr.
pooled	logical. TRUE if marginal distributions are exchangeable for all features so that rotated stats can be pooled, see pFdr.

Value

numeric vector of (adjusted) p-value or FDR estimations for s0.

References

\insertAllCited

randRotation documentation built on April 14, 2021, 6:01 p.m.