getThresholdDistribution: Wavelet Threshold Estimation

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

Description

Obtains wavelet coefficient distributions for the specified input file using the continuous wavelet transform (CWT). Monte Carlo sampling is performed with N samples of size 2^maxscale and coefficients are obtained at a range of quantiles defined by the range [1-p.min,1-p.max]. Statified sampling is performed chromosome-by-chromosome weighted by chromosome length.

Usage

getThresholdDistribution(infile, outdir, exptname, mother = "morlet", N = 5000, maxscale = 12, p.min = 0.001, p.max = 0.3)

Arguments

infile	Input padded graph file (required)
outdir	Output directory (required)
exptname	Unique string to identify name of experiment (required)
mother	Wavelet mother function used for the CWT. Tested choices are "morlet","haar","gaussian1", "gaussian2" (default="morlet")
N	Number of samples (default=5000)
maxscale	Maximum scale of wavelet decomposition. Also defines size of sample (2^maxscale) (default=12)
p.min	Minimum quantile (1-p.min) to be output for wavelet coefficient distribution (default=0.3)
p.max	Maximum quantile (1-p.max) as above (default=0.001)

Details

This function outputs wavelet coefficient thresholds in increments of p.max for each chromosome to the thresdist folder of outdir.

Warning

For large values of N and small values of winsize this step can be time-consuming.

Note

Function messages are piped to a log file which can be found in the log folder of outdir.

Author(s)

Apratim Mitra

See Also

wavCWT

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