mpreprocess: methylation data QC and preprocessing pipeline for Illuminal...

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

methylation data QC and preprocessing pipeline for Illuminal BeadChips

Description

The pipeline performs background correction, dye bias correction, inter-array normalization and probe type bias correction for HumanMethylation 450 and MethylationEPIC BeadChip data. It removes or mitigates background noise and systematic experimental bias,It also perform quality controls, identifing and excluding low quality samples and probes, removing low quality and outlier values, and performing imputation.

Usage

    mpreprocess(rgSet,nCores=2,bgParaEst="oob",dyeCorr="RELIC",
                qc=TRUE,qnorm=TRUE,qmethod="quantile1",
                fqcfilter=FALSE,rmcr=FALSE,impute=FALSE)

Arguments

rgSet	An object of class rgDataSet,methDataSet, RGChannelSetExtended, RGChannelSet or MethylSet.
nCores	Number of cores will be used for computation
bgParaEst	Method to estimate background normal distribution parameters. Possible options: "oob","est", or "neg".
dyeCorr	Dye bias correction, "mean": correction based on averaged red/green ratio; or "RELIC": correction with RELIC method; or "none": no dye bias correction. The default is RELIC
qc	If TRUE, QC will be performed. Low quality samples and CpGs will be excluded before background correction.
qnorm	If TRUE, inter-array quantile normalization will be performed.
qmethod	Quantile normalization method. This should be one of the following strings: "quantile1", "quantile2", or "quantile3". See details in function norm.quantile.
fqcfilter	If TRUE, outlier and low quality values will be filtered out.
rmcr	TRUE: excluded rows and columns with more than 5% of missing values. FALSE is in default
impute	Whether to impute missing values. If TRUE, k-nearest neighbor's methods will be used for imputation. FALSE is in default.

Details

Fuction mpreprocess is a pipeline that perform methylaiton data preprocessing and quality controls using functions: preprocessENmix, norm.quantile, rcp, QCinfo and qcfilter. More customized preprocessing steps can be achieved using the individual functions, see user's guide.

Value

A methylation beta value matrix with rows for CpGs and columns for samples.

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.

Zongli Xu, Sabine A. S. Langie, Patrick De Boever, Jack A. Taylor1 and Liang Niu, RELIC: a novel dye-bias correction method for Illumina Methylation BeadChip, BMC Genomics, 2017

Liang Niu, Zongli Xu and Jack A. Taylor: RCP: a novel probe design bias correction method for Illumina Methylation BeadChip, Bioinformatics 2016

See Also

Package minfi for classes RGChannelSet and MethylSet

Examples

if (require(minfiData)) {
#rgDataSet as input
path <- file.path(find.package("minfiData"),"extdata")
rgSet <- readidat(path = path,recursive = TRUE)
beta=mpreprocess(rgSet,nCores=6,qc=TRUE,fqcfilter=TRUE,rmcr=TRUE,impute=TRUE)

#methDataSet as input
mdat=getmeth(rgSet)
beta=mpreprocess(mdat,nCores=6)

#RGChannelSet as input
beta=mpreprocess(RGsetEx,nCores=6)

#RGChannelSetExtended as input
sheet <- read.metharray.sheet(file.path(find.package("minfiData"),"extdata"),
 pattern = "csv$")
rgSet <- read.metharray.exp(targets = sheet,extended = TRUE)
beta=mpreprocess(rgSet,nCores=6,qc=TRUE,fqcfilter=TRUE,rmcr=TRUE,impute=TRUE)

#MethylSet as input
mdat=preprocessRaw(rgSet)
beta=mpreprocess(mdat,nCores=6)

}

xuz1/ENmix documentation built on Nov. 24, 2024, 4:31 a.m.