tests/testthat/test_lm.R

In NKI-CCB/flexgsea-r: Flexible Gene Set Enrichment Analysis

requireNamespace('dplyr')

context("lm")

d = tibble::tibble(
    x1 = c(-0.91036224, -0.23633198,  0.07675882, -0.14270400),
    x2 = c(-0.0005694878, -0.9734813268, -2.0230448058,  0.3552175208),
    x2.big = 10*c(-0.0005694878, -0.9734813268, -2.0230448058,  0.3552175208),
    x3 = c(0.1979443, -1.0453220, -0.1528467,  0.0897873),
    x4 = -c(0.1979443, -1.0453220, -0.1528467,  0.0897873),
    y1 = c(0.19801108, -1.05384953, -0.15942901,  0.09702427),
    y2 = c(1.793303,  1.589571, -1.412860,  2.192915))
x = data.matrix(dplyr::select(d, dplyr::starts_with('x')))
y = data.matrix(dplyr::select(d, dplyr::starts_with('y')))

scores <- flexgsea_lm(x, y)
scores.abs <- flexgsea_lm(x, y, abs=T)
scores.single <- flexgsea_lm(x, y[,1])
test_that("Output is matrix", {
    expect_true(is.matrix(scores))
    expect_true(is.matrix(scores.abs))
    expect_true(is.matrix(scores.single))
})
test_that("Number of rows in output is number of genes", {
    expect_equal(nrow(scores), ncol(x))
    expect_equal(nrow(scores.abs), ncol(x))
    expect_equal(nrow(scores.single), ncol(x))
})
test_that("Number of columns in output is number of response variables", {
    expect_equal(ncol(scores), ncol(y))
    expect_equal(ncol(scores.abs), ncol(y))
    expect_equal(ncol(scores.single), 1)
})
test_that("Names of columns in output is names of response variables", {
    expect_equal(colnames(scores), colnames(y))
    expect_equal(colnames(scores.single), "Response")
    expect_equal(colnames(scores.abs), colnames(y))
})
test_that("Names of rows in output is names of genes", {
    expect_equal(rownames(scores), colnames(x))
    expect_equal(rownames(scores.single), colnames(x))
    expect_equal(rownames(scores.abs), colnames(x))
})
test_that("Output is finite", {
    expect_true(all(is.finite(scores)))
    expect_true(all(is.matrix(scores.abs)))
    expect_true(all(is.matrix(scores.single)))
})
test_that("Correlated response and variable get ranked at top", {
    expect_equal(names(which.max(scores[, 'y1'])), 'x3')
    expect_true(names(which.max(scores.abs[, 'y1'])) %in% c('x3', 'x4'))
    expect_equal(names(which.max(scores.single[,1])), 'x3')
})
test_that("Anti-correlated response and variable get ranked at bottom", {
    expect_equal(names(which.min(scores[, 'y1'])), 'x4')
    expect_equal(names(which.min(scores.single[,1])), 'x4')
})
test_that("Inflated variables get ranked similary to normal one", {
    expect_equal(scores['x2', ], scores['x2.big', ])
    expect_equal(scores.abs['x2', ], scores.abs['x2.big', ])
    expect_equal(scores.single['x2', ], scores.single['x2.big', ])
})