noControlFDR: FDR Estimation without Control

In WaveSeqR: A Wavelet-based Method for ChIP-Seq peak calling

Description

Estimates FDR for peaks in a ChIP sample using a randomized algorithm. noControlFDR consists of three steps:
(1) Randomized background estimation using randomizedBackground().
(2) Adjust p-values for multiple testing using p.adjust.
(3) Print peaks and p-values to file.

Usage

noControlFDR(peaks, chip, outfile, winsize = 200, 
			minreads = 8, samplesize = 1e+06, adj.method = "fdr")

Arguments

peaks	File containing peak locations. Must be in bedGraph format.
chip	Padded bedGraph file for the ChIP data being analyzed.
outfile	Output file.
winsize	Window size for obtaining peaks (default = 200).
minreads	Minimum reads within a peak (default = 8). This filtering is necessary for a more accurate estimation of FDR.
samplesize	Number of peaks to sample for building ecdf with the randomized algorithm.
adj.method	Method for multiple testing adjustment. Can use any methods accepted by p.adjust (default = "fdr").

Value

This function writes peaks to a file after FDR estimation in the absence of control. This file has the following fields:

chromosome	Chromosome name
start	Start position of peak (1-based indexing)
end	End position of peak
ChIP.reads	Number of ChIP reads in the peak
p.value	P-value for the peak
adj.p.value	P-values adjusted for multiple comparisons using adj.method

Author(s)

Apratim Mitra

See Also

sample, ecdf, p.adjust

WaveSeqR documentation built on May 2, 2019, 5:19 p.m.