clipCoverage: Coverage function for BSFDataSet objects

In ZarnackGroup/BindingSiteFinder: Binding site defintion based on iCLIP data

Coverage function for BSFDataSet objects

Description

Function that computes a crosslink coverage with all samples over all ranges given in the BSFDataSet. The coverage can be summarized over all combinations of the three dimension (samples, ranges, positions).

Usage

clipCoverage(
  object,
  ranges.merge = FALSE,
  ranges.merge.method = c("sum", "mean"),
  positions.merge = FALSE,
  positions.merge.method = c("sum", "mean"),
  samples.merge = TRUE,
  samples.group = c("all", "condition"),
  samples.merge.method = c("sum", "mean"),
  out.format = c("granges", "data.frame"),
  out.format.overwrite = FALSE,
  match.rangeID = "bsID",
  quiet = FALSE
)

Arguments

object	a BSFDataSet object
ranges.merge	logical; whether to merge ranges
ranges.merge.method	character; how to combine ranges ('sum' or 'mean')
positions.merge	logical; whether to merge positions
positions.merge.method	character; how to combine positions ('sum' or 'mean')
samples.merge	logical: whether to merge samples
samples.group	character; how samples should be grouped when combining ('all', 'condition')
samples.merge.method	charater; how to combine positions ('sum' or 'mean')
out.format	character; how the coverage should be returned ('data.frame' or 'granges'). Note that option 'granges' only exists if the output coverage is of the same rows as the input ranges.
out.format.overwrite	logical; if out.format='granges', then decide wheter the meta columns should be extended by the coverage information or be overwritten
match.rangeID	character; unique internal identifier. Name of the meta column of the input ranges that should be used as identifier to match the coverage back to the input ranges. Is 'bsID' as default, since that ID exists for all binding sites after makeBindingSites was called.
quiet	logical; whether to print messages

Details

When summarizing the crosslink coverage over samples (samples.merge=TRUE) one can decide whether to summarize all samples or whether to keep conditions separate (samples.group). This either reduces the samples dimension to a single matrix, or a list. For a binding site set with 100 binding sites of width=7 and 4 replicates with 2 conditions, the following options are possible. With merging enabled and samples.group='all' the coverage of all samples is combined. With samples.group='condition' only samples of the same condition are grouped.

When summarizing the crosslink coverage over ranges, all ranges are combined which reduces the ranges dimension to a single vector. This turns eg. a binding site set of 100 binding sites with width=7 into a vector of length 100 with exactly one column. Depending on how the samples were summarized, the result can be a single such vector, or a list.

When summarizing the crosslink coverage over positions, all positions are combined which reduces the positions dimension to a single vector. This turns eq. a binding site set of 100 binding sites with width=7 into a vector of length 1 with 7 columns. Depending on how the samples were summarized, the result can be a single such vector, or a list.

For all summarizing operations options sum and mean exists. This allows for normalization by the eg. the number of binding sites, size of the range, number of sample, etc..

If the resulting object does have a dimension that fits to the number of input ranges the result can be directly attached to them. Basically extending the GRanges object (out.format).

Value

an object of class specified in out.format

See Also

BSFDataSet, BSFind

Examples

# load data
files <- system.file("extdata", package="BindingSiteFinder")
load(list.files(files, pattern = ".rda$", full.names = TRUE))

bds = makeBindingSites(bds, bsSize = 7)
# sum of each replicate over each binding site position
c1 = clipCoverage(bds, out.format = "data.frame", positions.merge = TRUE,
 ranges.merge = FALSE, samples.merge = FALSE)
# total signal per binding site from all samples
c2 = clipCoverage(bds, out.format = "granges", positions.merge = FALSE,
 ranges.merge = TRUE, samples.merge = TRUE, samples.group = "all")
# total signal per binding site from all samples - split by condition
c3 = clipCoverage(bds, out.format = "granges", positions.merge = FALSE,
 ranges.merge = TRUE, samples.merge = TRUE, samples.group = "condition")

ZarnackGroup/BindingSiteFinder documentation built on May 31, 2024, 3:29 a.m.