Wrapper to denoise a correlation matrix using a Random Matrix Theory approach

Usage

getDenoisedCorMatrix(
  obj,
  res = 1000000,
  chr = "chr14",
  genome = c("hg19", "hg38", "mm9", "mm10"),
  iter = 2,
  targets = NULL,
  prior.means = NULL,
  assay = c("rna", "atac", "array")
)

Arguments

obj: SummarizedExperiment object with rowRanges for each feature and colnames
res: The resolution desired (default is a megabase 1e6)
chr: Which chromosome to perform the denoising
genome: Which genome (default is hg19)
iter: How many iterations to perform denoising
targets: Samples/cells to shrink towards
prior.means: The means of the bin-level prior distribution (default will compute them for you)
assay: What assay type this is ("rna", "atac")

Value

A denoised correlation matrix object for plotting with plotCorMatrix

Examples

data("k562_scrna_chr14", package = "compartmap")
denoised_cor_mat <- getDenoisedCorMatrix(k562_scrna_chr14, genome = "hg19", assay = "rna")
#> Shrinking bins with the JSE.
#> Number of means fewer than 4. Using Bayes instead of JSE.
#> 108 bins created...
#> Calculating correlations...
#> Done...
#> Denoising the correlation matrix using RMT.
#> Iterative denoising. Iteration: 2