ContactMatrix-overlaps: Find overlaps between GRanges and a ContactMatrix

In LTLA/InteractionSet: Base Classes for Storing Genomic Interaction Data

Find overlaps between GRanges and a ContactMatrix

Description

Find overlaps between a set of linear intervals in a GRanges object, and the set of regions representing the rows or columns of a ContactMatrix.

Usage

## S4 method for signature 'ContactMatrix,GRanges'
overlapsAny(query, subject, maxgap=0L, minoverlap=1L,
    type=c("any", "start", "end", "within", "equal"),
    ignore.strand=TRUE)

## S4 method for signature 'ContactMatrix,GRangesList'
overlapsAny(query, subject, maxgap=0L, minoverlap=1L,
    type=c("any", "start", "end", "within", "equal"),
    ignore.strand=TRUE, use.region="both")

# The same call is used for the methods taking GRangesList, GInteractions and
# InteractionSet objects as 'subject'. For brevity, these will not be listed.

Arguments

query	A ContactMatrix object.
subject	A GRanges, GRangesList, GInteractions or InteractionSet object.
maxgap, minoverlap, type	See ?findOverlaps in the GenomicRanges package.
ignore.strand	See ?findOverlaps in the GenomicRanges package.
use.region	A string specifying how the interacting regions are to be matched to row/column regions.

Details

When subject is a GRanges, overlaps are identified between the row regions of query and the regions in subject. This is repeated for the column regions of query. A list of two logical vectors is returned, specifying the rows and columns in query that are overlapped by any region in subject. These vectors can be directly used to subset query. Alternatively, they can be used in outer to generate a logical matrix for masking – see Examples.

For all other classes of subject, two-dimensional overlaps are identified. A logical matrix is returned indicating which entries in the ContactMatrix have overlaps with the specified interactions. For any given entry, an overlap is only considered if the regions for the corresponding row and column both overlap anchor regions in the subject. See ?"findOverlaps,GInteractions,GInteractions-method" for more details.

If use.region="both", overlaps between any row/column region and the first/second interacting region of subject are considered. If use.region="same", only overlaps between row regions and the first interacting regions, or between column regions and the second interacting regions are considered. If use.region="reverse", only overlaps between row regions and the second interacting regions, or between row regions and the first interacting regions are considered.

Value

For overlapsAny,ContactMatrix,GRanges-method, a named list of two logical vectors is returned specifying the rows of columns of query overlapped by subject.

For the other overlapsAny methods, a logical matrix is returned indicating which entries in query are overlapped by subject.

Author(s)

Aaron Lun

See Also

ContactMatrix-class, findOverlaps

Examples

example(ContactMatrix, echo=FALSE)

of.interest <- resize(sample(regions(x), 2), width=1, fix="center")
olap <- overlapsAny(x, of.interest)
olap
x[olap$row,]
x[,olap$column]
x[olap$row,olap$column]

keep <- outer(olap$row, olap$column, "|") # OR mask
temp <- as.matrix(x)
temp[!keep] <- NA

keep <- outer(olap$row, olap$column, "&") # AND mask
temp <- as.matrix(x)
temp[!keep] <- NA

# Two dimensional overlaps.
pairing <- GRangesList(first=regions(x), second=regions(x))
olap <- overlapsAny(x, pairing)
olap
olap <- overlapsAny(sort(x), pairing) # A bit prettier
olap

LTLA/InteractionSet documentation built on July 3, 2023, 8:44 a.m.