nmode: Estimating number of mode for each row of data

In xuz1/ENmix: Quality control and analysis tools for Illumina DNA methylation BeadChip

Estimating number of mode for each row of data

Description

Due to SNPs in CpG probe region or other unknow factors, methylation beta values for some CpGs have multimodal distribution. This function is to identify this type of probes (so called gap probes) with obovious multimoal distribution.

Usage

nmode(x, minN = 3, modedist=0.2, nCores = 1)

Arguments

x	A methylation beta value matrix with row for probes and column for samples.
minN	Minimum number of data points at each cluster
modedist	Minimum distance between adjacent modes
nCores	Number of cores used for computation

Details

This function uses an empirical approach to estimate number of modes in methylation beta value for each CpG probe. By default, the function requires the distance between modes have to be greater than 0.2 in methylation beta value, and each mode clusters should has at least 3 data points or 5% of data points whichever is greater.

Value

A vector of integers indicating number of modes. Gap probes are probes with number of mode greater than 1.

Author(s)

Zongli Xu

References

Zongli Xu, Liang Niu, Leping Li and Jack A. Taylor, ENmix: a novel background correction method for Illumina HumanMethylation450 BeadChip. Nucleic Acids Research 2015

Examples

if (require(minfiData)) {
mdat <- preprocessRaw(RGsetEx)
beta=getBeta(mdat, "Illumina")
nmode=nmode(beta, minN = 3,modedist=0.2, nCores = 5)

path <- file.path(find.package("minfiData"),"extdata")
rgSet <- readidat(path = path,recursive = TRUE)
mdat <- getmeth(rgSet)
beta=getB(mdat)
nmode=nmode(beta, minN = 3,modedist=0.2, nCores = 5)
}

xuz1/ENmix documentation built on Aug. 5, 2023, 7:11 a.m.