makeTable: Create a Results Table
In MatthiasLienhard/qsea: IP-seq data analysis and vizualization

makeTable

R Documentation

Create a Results Table

Description

This function creates a table from the qsea objects qseaSet and qseaTvN_glm

Usage

makeTable(qs,glm,norm_methods="counts",samples,groupMeans, keep, ROIs,
    annotation, minPvalSummarize, CNV=FALSE, verbose=TRUE, minEnrichment=3,
    chunksize=1e5)

Arguments

`qs`	a qseaSet object (mandatory)
`glm`	a list of one or more qseaGLM objects (optional)
`norm_methods`	ether a character vector of pre-defined normalization combinations, or a list defining normalization combinations. This affects both individual and mean values.
`samples`	The indices of the samples for which individual values are to be written out in the specified order
`groupMeans`	a named list of indices vectors, defining groups for which mean values are to be written out
`keep`	a vector of indices of the windows that are considered (as created by isSignificant)
`ROIs`	A GRanges object, containing regions of interest (ROIs). Only windows overlapping ROIs are considered.
`annotation`	a named list of GRange objects, containing annotations (e.g. genes, CpG islands, ...) that are added to the table.
`minPvalSummarize`	If ROIs are given, you can specify a QseaTvN_glm object. For each ROI the window with the most significant differential coverage is written out
`CNV`	If set TRUE, the CNV logFC for the samples specified by samples are written out.
`verbose`	verbosity level
`minEnrichment`	for transformation to absolute methylation level, you can specify the minimal number of expected reads for a fully methylated window. This avoids inaccurate estimates, due to low enrichment.
`chunksize`	For efficient memory usage, the table is built up in chunks. With this parameter, the maximum number of windows processed in one chunk is specified.

Details

Note that, if overlapping ROIs are specified, windows might emerge in the table several times.

Value

A result table containing the specified normalized values for the selected windows and samples/groups

Author(s)

Mathias Lienhard

Examples

#create example set
qs=getExampleQseaSet()
design=model.matrix(~group, getSampleTable(qs))
TvN_glm=fitNBglm(qs, design, norm_method="beta")
TvN_glm=addContrast(qs,TvN_glm, coef=2, name="TvN")
sig=isSignificant(TvN_glm, fdr_th=0.01)

##Table containing all significant windows
tab1=makeTable(qs=qs, glm=TvN_glm, 
    keep=sig, samples=getSampleNames(qs))
##additional CNV logFC for the selected samples
tab2=makeTable(qs=qs, glm=TvN_glm, 
    keep=sig, samples=getSampleNames(qs), CNV=TRUE)
##explicit selection of normalization: 
##counts (i.e. no normalization, only counts)
tab3=makeTable(qs=qs, glm=TvN_glm, keep=sig,
    samples=getSampleNames(qs), norm_method="counts")

##counts AND %methylation values for individual samples and group means
tab4=makeTable(qs=qs, glm=TvN_glm, keep=sig,
    samples=getSampleNames(qs), groupMeans=getSampleGroups(qs),
    norm_method=c("counts", "beta"))

MatthiasLienhard/qsea documentation built on March 22, 2023, 1:15 p.m.