calculateFDRs: Determine false discovery rate
In MethylSeekR: Segmentation of Bis-seq data

Description Usage Arguments Value Author(s) Examples

This function calculates the false discovery rate (FDR) based on randomized Bis-seq data.

1
2
3

calculateFDRs(m, CGIs, PMDs = NA, pdfFilename=NULL, num.cores = 1,
nCpG.smoothing = 3, meth.cutoffs = seq(0.3, 0.7, by=0.1), nCpG.cutoffs =
seq(1, 6, by=1), minCover = 5)

`m`	GRanges object containing the methylation data.
`CGIs`	A GRanges object of CpG island coordinates. All CpGs overlapping CpG islands will be removed for the randomization.
`PMDs`	The GRanges object of the PMDs. Set to either the return value of the function segmentPMDs (see example) or to NA (default) if there are no PMDs.
`pdfFilename`	Name of the pdf file in which the figure is saved. If no name is provided (default), the figure is printed to the screen.
`num.cores`	Number of cores used for the calculations.
`nCpG.smoothing`	The number of consecutive CpGs that the methylation levels are averaged over.
`meth.cutoffs`	A vector containing the cut-offs in methylation for which the FDR should be calculated. Numbers must be between 0 and 1.
`nCpG.cutoffs`	A vector containing the cut-offs on the minimal number of CpGs per region for which the FDR should be calculated.
`minCover`	Only CpGs with a coverage of at least minCover reads will be used.

A list containing a matrix with FDR values and a matrix with the number of inferred segments for each methylation cut-off (rows) and each cut-off on the minimal number of CpGs per region (columns). The function creates a figure showing the relationship between the methylation cut-off, the cut-off on the minimal number of CpGs per region, the number of inferred segments and the FDR. The figure is either printed to the screen (default) or saved as a pdf if a filename is provided.

Lukas Burger lukas.burger@fmi.ch

library(MethylSeekR)

# get chromosome lengths
library("BSgenome.Hsapiens.UCSC.hg18")
sLengths=seqlengths(Hsapiens)

# read methylation data
methFname <- system.file("extdata", "Lister2009_imr90_hg18_chr22.tab",
package="MethylSeekR")
meth.gr <- readMethylome(FileName=methFname, seqLengths=sLengths)

#load CpG islands
library(rtracklayer)
session <- browserSession()
genome(session) <- "hg18"
query <- ucscTableQuery(session, "cpgIslandExt")
CpGislands.gr <- track(query)
genome(CpGislands.gr) <- NA
CpGislands.gr <- resize(CpGislands.gr, 5000, fix="center")

#calculate FDRs, assuming no PMDs
stats <- calculateFDRs(m=meth.gr, CGIs=CpGislands.gr)