mergeCalls: Automatic merging of overlapped nucleosome calls

Description

This function joints close nucleosome calls into one larger, fuzzy nucleosome.

Usage

mergeCalls(calls, min.overlap=50, discard.low=0.2, mc.cores=1, verbose=TRUE)

Arguments

calls

RangedData with scored and ranged nucleosome calls from peakScoring or peakDetection(..., score=TRUE).

min.overlap

Minimum overlap between two reads for merge them

discard.low

Discard low covered calls (i.e. calls with score_h < discard.low will be discarded)

mc.cores

Number of cores available to parallel data processing.

verbose

Show progress info?

Value

RangedData with merged calls and the additional data column nmerge, with the count of how many original ranges are merged in the resulting range.

Details

This functions looks for overlapped calls and join those with more than min.overlap bases overlapped. More than two reads can be joined in one single call if all of them are overlapped at least that distance with almost another read in the range. Joining is performed in chain, so if nucleosome call A is close to B and B is close to C, the final call will comprise the range A-B-C. The resulting scores (mixed, width, height) of the final joined call will be the average value of the individual scores.

The parameter discard.low allows to ignore the small peaks that could be merged with larger ones, originating large calls. In the case that all of the overlapped reads in a given position have score_h less than discard.low, all of them will be selected instead of deleting that call.

Examples

Run this code

	#Generate a synthetic coverage map (assuming reads of 40bp and fragments of 130)
	map = syntheticNucMap(wp.num=20, fuz.num=20,  nuc.len=40, lin.len=130, rnd.seed=1)
	cover = filterFFT(coverage(map$syn.reads))

	#Find peaks over FFT filtered coverage
	calls = peakDetection(filterFFT(cover, pcKeepComp=0.02), width=130, score=TRUE)

	#Merge overlapped calls
	merged_calls = mergeCalls(calls)

	plotPeaks(merged_calls, cover)