extractPatch: Extract a patch of the interaction space

In LTLA/diffHic: Differential Analysis of Hi-C Data

Extract a patch of the interaction space

Description

Extract and count read pairs into bin pairs for a subset of the interaction space.

Usage

extractPatch(file, param, first.region, second.region=first.region, 
    width=10000, restrict.regions=FALSE)

Arguments

file	character string specifying the path to an index file produced by preparePairs
param	a pairParam object containing read extraction parameters
first.region	a GRanges object of length 1 specifying the first region
second.region	a GRanges object of length 1 specifying the second region
width	an integer scalar specifying the width of each bin in base pairs
restrict.regions	A logical scalar indicating whether the output regions should be limited to entries in param$restrict.

Details

This function behaves much like squareCounts, but only for the “path” of the interaction space defined by first.region and second.region. Read pairs are only counted into bin pairs where one end overlaps first.region and the other end overlaps second.region. This allows for rapid extraction of particular regions of interest without having to count across the entire interaction space.

Note that the first anchor region (i.e., bin) in each bin pair is not necessarily the bin that overlaps first.region. In each pair, the bins are sorted so that the first bin has a higher genomic coordinate than the second bin. The flipped flag in the metadata of the output object indicates whether this order is flipped. If TRUE, the first bin in each pair corresponds to second.region, and vice versa.

If restrict.regions=TRUE, only bins on the chromosomes in first.region and second.region will be reported in the regions slot of the output object. This avoids the overhead of constructing many bins when only a small subset of them are used. By default, restrict.regions=FALSE to ensure that the anchor IDs of the output object are directly comparable between different calls to extractPatch.

Value

An InteractionSet object containing the number of read pairs for each bin pair in the specified patch.

Author(s)

Aaron Lun

See Also

squareCounts

Examples

hic.file <- system.file("exdata", "hic_sort.bam", package="diffHic")
cuts <- readRDS(system.file("exdata", "cuts.rds", package="diffHic"))
param <- pairParam(fragments=cuts)

# Setting up the parameters
fout <- tempfile(fileext=".h5")
invisible(preparePairs(hic.file, param, file=fout))

stuff <- extractPatch(fout, param, GRanges("chrA:1-100"))
interactions(stuff)

stuff <- extractPatch(fout, param, GRanges("chrA:1-100"), GRanges("chrB:1-20"))
interactions(stuff)

LTLA/diffHic documentation built on Dec. 20, 2024, 7:06 p.m.