RnBeads (version 1.4.0)

rnb.execute.computeDiffMeth: rnb.execute.computeDiffMeth

Description

computes differential methylation

Usage

rnb.execute.computeDiffMeth(x, pheno.cols, region.types = rnb.region.types.for.analysis(x), covg.thres = rnb.getOption("filtering.coverage.threshold"), pheno.cols.all.pairwise = rnb.getOption("differential.comparison.columns.all.pairwise"), columns.pairs = rnb.getOption("columns.pairing"), columns.adj = rnb.getOption("covariate.adjustment.columns"), adjust.sva = rnb.getOption("differential.adjustment.sva"), pheno.cols.adjust.sva = rnb.getOption("inference.targets.sva"), adjust.celltype = rnb.getOption("differential.adjustment.celltype"), skip.sites = !rnb.getOption("analyze.sites"), disk.dump = rnb.getOption("disk.dump.big.matrices"), disk.dump.dir = tempfile(pattern = "diffMethTables_"), ...)

Arguments

x

RnBSet object

pheno.cols

column names of the pheno slot in x on which the dataset should be partitioned. Those columns are required to be factors or logical. In case of factors, each group in turn will be compared to all other groups

region.types

which region types should be processed for differential methylation

covg.thres

coverage threshold for computing the summary statistics. See computeDiffTab.extended.site for details.

pheno.cols.all.pairwise

integer or character vector specifying the colomns of pheno(x) on which all pairwise comparisons should be conducted. A value of NULL (default) indicates no columns.

columns.pairs

argument passed on to rnb.sample.groups. See its documentation for details.

columns.adj

Column names or indices in the table of phenotypic information to be used for confounder adjustment in the differential methylation analysis.

adjust.sva

flag indicating whether the adjustment table should also contain surrogate variables (SVs) for the given target variable.

pheno.cols.adjust.sva

Column names or indices in the table of phenotypic information to be used for SVA adjustment in the differential methylation analysis.

adjust.celltype

flag indicating whether the resulting table should also contain estimated celltype contributions. See rnb.execute.ct.estimation for details.

skip.sites

flag indicating whether differential methylation in regions should be computed directly and not from sites. This leads to skipping of site-specific differential methylation

disk.dump

Flag indicating whether the resulting differential methylation object should be file backed, ie.e the matrices dumped to disk

disk.dump.dir

disk location for file backing of the resulting differential methylation object. Only meaningful if disk.dump=TRUE. must be a character specifying an NON-EXISTING valid directory.

...

arguments passed on to binary differential methylation calling. See computeDiffTab.extended.site for details.

Value

an RnBDiffMeth object. See class description for details.

Examples

library(RnBeads.hg19)
data(small.example.object)
logger.start(fname=NA)
dm <- rnb.execute.computeDiffMeth(rnb.set.example,pheno.cols=c("Sample_Group","Treatment"))
get.comparisons(dm)