getCorMatrix: Calculate Pearson correlations of a binned matrix

In biobenkj/compartmap: Higher-order chromatin domain inference in single samples from methylation arrays and single cells from scRNA-seq and scATAC-seq

Calculate Pearson correlations of a binned matrix

Description

This function is used to generate a list object to be passed to getABSignal

Usage

getCorMatrix(binmat, squeeze = FALSE)

Arguments

binmat	A binned matrix list object from getBinMatrix
squeeze	Whether to squeeze the matrix for Fisher's Z transformation

Value

A list object to pass to getABSignal

Examples

library(GenomicRanges)

# Generate random genomic intervals of 1-1000 bp on chr1-22
# Modified from https://www.biostars.org/p/225520/
random_genomic_int <- data.frame(chr = rep("chr14", 100))
random_genomic_int$start <- apply(random_genomic_int, 1, function(x) { 
  round(runif(1, 0, getSeqLengths(getGenome("hg19"), chr = x)[[1]]), 0)
})
random_genomic_int$end <- random_genomic_int$start + runif(1, 1, 1000)
random_genomic_int$strand <- "*"

# Generate random counts
counts <- rnbinom(1000, 1.2, 0.4)

# Build random counts for 10 samples
count.mat <- matrix(sample(counts, nrow(random_genomic_int) * 10, replace = FALSE), ncol = 10)
colnames(count.mat) <- paste0("sample_", seq(1:10))

# Bin counts
bin.counts <- getBinMatrix(
  count.mat,
  makeGRangesFromDataFrame(random_genomic_int),
  chr = "chr14",
  genome = "hg19"
)

#Calculate correlations
bin.cor.counts <- getCorMatrix(bin.counts)

biobenkj/compartmap documentation built on April 5, 2025, 9:17 a.m.