cherByThreshold: Function to identify chers based on thresholds

Description

Given a vector of probe positions on the chromosome, a vector of smoothed intensities on these positions, and a threshold for intensities to indicated enrichment, this function identifies Chers (ChIP-enriched regions) on this chromosome. This function is called by the function findChersOnSmoothed.

Usage

cherByThreshold(positions, scores, threshold, distCutOff, minProbesInRow = 3)

Arguments

positions

numeric vector of genomic positions of probes

scores

scores (intensities) of probes on those positions

threshold

threshold for scores to be called a cher

distCutOff

maximal positional distance between two probes to be part of the same cher

minProbesInRow

integer; minimum number of enriched probes required for a cher; see details for further explanation.

Value

A LIST with n components, where the first n components are the cher clusters, each one holding the scores and, as their names, the genomic positions of probes in that cluster.

Details

Specifying a minimum number of probes for a cher (argument minProbesInRow) guarantees that a cher 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.

Examples

Run this code

 ## example with random generated data:
 rpos <- cumsum(round(runif(200)*5))
 rsco <- rnorm(200)+0.2
 plot(rpos, rsco, type="l", col="seagreen3", lwd=2)
 rug(rpos, side=1, lwd=2); abline(h=0, lty=2)              
 rchers <- cherByThreshold(rpos, rsco, threshold=0, distCutOff=2)
 sapply(rchers[-length(rchers)], function(thisClust){
  points(x=as.numeric(names(thisClust)), y=thisClust, type="h", lwd=2,
 col="gold")})