tests/testthat/test-getGlobalMeans.R

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

nrows <- 100
ncols <- 6
mat.counts <- matrix(runif(nrows * ncols, 1, 10), nrows)
gr <- GRanges(
  Rle(c("chr1", "chr2", "chr3", "chr4"), c(25, 25, 25, 25)),
  IRanges(1:100, width = 1:1)
)
se.rna <- SummarizedExperiment(
  assays = SimpleList(counts = mat.counts),
  rowRanges = gr
)
mat.rownames <- as.character(gr)

test_that("getGlobalMeans", {
  mat.beta <- matrix(runif(nrows * ncols, 0, 1), nrows)
  se.rna <- SummarizedExperiment(
    assays = SimpleList(counts = mat.counts),
    rowRanges = gr
  )
  expected.rna <- matrix(rowMeans(mat.counts), ncol = 1)
  colnames(expected.rna) <- "globalMean"
  rownames(expected.rna) <- mat.rownames

  expected.rna.two_col <- matrix(rowMeans(mat.counts[, 1:2]))
  colnames(expected.rna.two_col) <- "globalMean"
  rownames(expected.rna.two_col) <- mat.rownames

  se.array <- SummarizedExperiment(
    assays = SimpleList(Beta = mat.beta),
    rowRanges = gr
  )
  expected.array <- matrix(rowMeans(flogit(mat.beta)), ncol = 1)
  colnames(expected.array) <- "globalMean"
  rownames(expected.array) <- mat.rownames

  expect_equal(
    getGlobalMeans(se.rna, assay = "rna"),
    expected.rna
  )
  expect_equal(
    getGlobalMeans(se.rna, assay = "rna"),
    expected.rna
  )
  expect_equal(
    getGlobalMeans(se.rna, assay = "atac"),
    expected.rna
  )
  expect_equal(
    getGlobalMeans(se.array, assay = "array"),
    expected.array
  )
  expect_equal(
    getGlobalMeans(se.rna, assay = "rna", targets = c(1, 2)),
    expected.rna.two_col
  )

  se.invalid <- SummarizedExperiment(
    assays = SimpleList(badname = mat.counts),
    rowRanges = gr
  )
  expect_error(
    getGlobalMeans(se.invalid, assay = "rna")
  )
})

test_that("precomputeBootstrapMeans", {
  expected.rownames <- as.character(gr)
  boot.mean <- precomputeBootstrapMeans(se.rna, num.bootstraps = 2)
  expect_equal(rownames(boot.mean), expected.rownames)

  lapply(1:50, function(boot_count) {
    boot.mean <- precomputeBootstrapMeans(se.rna, num.bootstraps = boot_count)
    expect_equal(ncol(boot.mean), boot_count)
  })
  expect_error(
    precomputeBootstrapMeans(se.rna, num.bootstraps = 2, targets = c(1:4)),
    "Need 5 or more samples for targeted bootstrapping to work."
  )
  expect_no_error(
    precomputeBootstrapMeans(se.rna, num.bootstraps = 2, targets = c(1:5))
  )
})