getBindLen: Estimate length of binding and support region

In ChIPseqR: Identifying Protein Binding Sites in High-Throughput Sequencing Data

Description

The cross-correlation between forward and reverse strand read counts is used to estimate the distance between peaks on both strands. This is then used to derive suitable values for the length of binding and support regions.

Usage

getBindLen(data, bind, support, summary = median, verbose = FALSE, plot = TRUE, ...)

Arguments

data	An object of class ReadCounts.
bind	Either known length of binding region or minimum and maximum of binding region length to consider.
support	Either known length of support region or minimum and maximum of support region length to consider.
summary	Function to use to summarise cross-correlation across chromosomes.
verbose	Logical indicating whether progress messages should be printed.
plot	Logical indicating whether cross-correlation should be plotted.
...	Further arguments to getBindCor.

Details

This assumes that the first peak in cross-correlation corresponds to the length of the binding site. Note that this is not accurate. The peak is closer to bind + 2*m where m is the median of the read distribution in the support region ('read distribution in the support region' means the read density as a function of distance to binding site start/end). Consequently this method will overestimate the length of the binding site. If either bind or support are of length 1 this is assumed to be the known value and a more accurate estimate for the remaining parameter is used.

Value

A numeric vector giving the estimated lengths of binding and support regions.

Author(s)

Peter Humburg

ChIPseqR documentation built on Nov. 8, 2020, 6:49 p.m.