reproducibilityFilter {BindingSiteFinder}R Documentation

Replicate reproducibility filter function

Description

For each replicate the number of binding sites with a certain number of crosslinks is calculated. A quantile based threshold (cutoff) is applied to each replicate. This indicates how many of the merged binding sites are supported by crosslinks from the respective replicate. Next, one can specify how many replicates need to pass the defined threshold for a binding site to be considered reproducible.

Usage

reproducibilityFilter(
  object,
  cutoff = 0.05,
  n.reps = 1,
  min.crosslinks = 1,
  returnType = c("BSFDataSet", "data.frame")
)

Arguments

object

a BSFDataSet object

cutoff

a vector of length = 1, or of length = levels(getMeta(object)$conditions) with a single number (between 0-1) indicating the quantile cutoff

n.reps

a vector of length = 1, or of length = l evels(getMeta(object)$conditions) indicating how many replicates need to meet their threshold for a binding site to be called reproducible.

min.crosslinks

numeric of length = 1, defines the lower boundary for the minimum number of crosslinks a binding site has to be supported by all replicates, regardless of the replicate specific quantile threshold

returnType

one of "BSFDataSet" or "data.frame". "BSFDataSet" is the default and "matrix" can be used for easy plotting.

Details

If cutoff is a single number then the indicated cutoff will be applied to all replicates. If it is a vector then each element in the vector is applied to all replicates of the respective condition. The order is hereby given by the levels of the condition column of the meta data (see BSFDataSet,getMeta). If the condition specific filter is applied, a meta column is added to the GRanges of the BSFDataSet object, indicating the support for each condition.

If n.reps is a single number then this number is used as treshold for all binding sites. If it is a vector then it is applied to the replicates of the respective condition (like in cutoff). This allows the application of different thresholds for experiments of different experimental conditions. If the condition specific filter is applied, a meta column is added to the GRanges of the BSFDataSet object, indicating the support for each condition.

Value

an object of type BSFDataSet

Examples

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

# merge binding sites
bds <- makeBindingSites(object = bds, bsSize = 9, minWidth = 2,
minCrosslinks = 2, minClSites = 1)

# use default return with single threshold
s = reproducibilityFilter(bds, cutoff = c(0.05), n.reps = c(3))

# use default return with condition specific threshold
s = reproducibilityFilter(bds, cutoff = c(0.1, 0.05), n.reps = c(1, 2))

# use data.frame return type for plotting
s = reproducibilityFilter(bds, cutoff = c(0.1, 0.05), n.reps = c(1, 2),
returnType = "data.frame")
library(ComplexHeatmap)
m = make_comb_mat(s)
UpSet(m)


[Package BindingSiteFinder version 1.0.0 Index]