computeExpectations

computeExpectations

computeExpectations(object, ...)

# S4 method for MatrixOrmatrix
computeExpectations(object, norm = FALSE,
  group = NULL)

# S4 method for SummarizedExperiment
computeExpectations(object, norm = FALSE,
  group = NULL)

Arguments

object	SummarizedExperiment
...	additional arguments
norm	weight all samples equally?
group	an group vector, optional

Value

vector with expected fraction of reads per peak.

Details

By default, this function will compute the expected fraction of reads per peak as the the total fragments per peak across all samples divided by total reads in peaks in all samples. Optionally, norm can be set to TRUE and then the expectation will be the average fraction of reads in a peak across the cells. This is not recommended for single cell applications as cells with very few reads will have a large impact. Another option is to give a vector of groups, in which case the expectation will be the average fraction of reads per peak within each group. If group vector is provided and norm is set to TRUE then within each group the fraction of reads per peak is the average fraction of reads per peak in each sample. Otherwise, the within group fraction of reads per peak is based on the reads per peak within the sample divided by the total reads within each sample. The group can also be given by a length 1 character vector representing the name of a column in the colData of the input object if the input is a SummarizedExperiment

Methods (by class)

• MatrixOrmatrix: method for Matrix or matrix

• SummarizedExperiment: method for SummarizedExperiment with counts slot

Examples

# NOT RUN {
# First get some data
data(mini_counts, package = "chromVAR")

# Compute expectations
expectations <- computeExpectations(mini_counts)

# }