findPeaksOnSmoothed {Ringo}R Documentation

Find peaks on smoothed ExpressionSet

Description

Given an ExpressionSet of smoothed probe intensities, an environment with the mapping of probes to chromosomes, and a vector of thresholds for calling genomic sites enriched, this function finds the peaks consisting of enriched genomic positions, with probes mapped to them. 'Adjacent' enriched positions are condensed into a single peak.

Usage

findPeaksOnSmoothed(smoothedX, probeAnno, thresholds, allChr = c(1:19,
"X", "Y"), distCutOff = 600, minProbesInRow = 3, cellType=NULL,
verbose = TRUE)

Arguments

smoothedX Object of class ExpressionSet holding the smoothed probe intensities, e.g. the result of function computeRunningMedians.
probeAnno environment containing the probe to genome mapping
thresholds numeric vector of threshold above which smoothed probe intensities are considered to correspond to enriched probes. The vector has to be of length equal the number of samples in smoothedX, with a single threshold for each sample.
allChr character vector of all chromosomes, on which peaks are to be found. Every chromosome here has to have probes mapped to it in the probeAnno environment.
distCutOff integer; maximum amount of base pairs at which enriched probes are condensed into one peak.
minProbesInRow integer; minimum number of enriched probes required for a peak; see details for further explanation.
cellType character; name of cell type the data stems from, maybe could be taken from pData(smoothedX)
verbose logical; extended output to STDOUT?

Details

Specifying a minimum number of probes for a peak (argument minProbesInRow) guarantees that a peak is supported by a reasonable number of measurements in probe-sparse regions. For example, if there's only one enriched probe within a certain genomic 1kb region and no other probes can been mapped to that region, this single probe does arguably not provide enough evidence for calling this genomic region enriched.

Value

A list of class peakList, holding objects of class peak that were found on the supplied data.

Author(s)

Joern Toedling toedling@ebi.ac.uk

See Also

peakByThreshold,computeRunningMedians, relatePeaks

Examples

  exDir <- system.file("exData",package="Ringo")
  load(file.path(exDir,"exampleProbeAnno.rda"))
  load(file.path(exDir,"exampleX.rda"))
  smoothX <- computeRunningMedians(exampleX, probeAnno=exProbeAnno,
     modColumn = "Cy5", allChr = c("8"), winHalfSize = 400, combineReplicates=TRUE)
  peaksX <- findPeaksOnSmoothed(smoothX, probeAnno=exProbeAnno,
thresholds=1.5, allChr = c("8"),distCutOff=600, cellType="HL1")
  if (interactive())
    plot(peaksX[[2]], smoothX, probeAnno=exProbeAnno, gff=exGFF)
  peaksX <- relatePeaks(peaksX, exGFF)
  as.data.frame.peakList(peaksX)

[Package Ringo version 1.0.0 Index]