tests/testthat/test-filter_blank.R

In metamorphr: Tidy and Streamlined Metabolomics Data Workflows

test_that("works for min_frac = 5", {
  filtered_features <- test_filters %>%
    # only look at "samples"
    dplyr::filter(!(Sample %in% c("q1", "q2", "q3"))) %>%
    filter_blank(min_frac = 5, blank_samples = "b1") %>%
    dplyr::select(2) %>%
    dplyr::pull() %>%
    unique()

  expect_equal(filtered_features, c("f1", "f10"))
})

test_that("works for min_frac = 3", {
  filtered_features <- test_filters %>%
    # only look at "samples"
    dplyr::filter(!(Sample %in% c("q1", "q2", "q3"))) %>%
    filter_blank(min_frac = 3, blank_samples = "b1") %>%
    dplyr::select(2) %>%
    dplyr::pull() %>%
    unique()

  expect_equal(filtered_features, c("f1", "f2", "f10"))
})

test_that("works for min_frac = 2", {
  filtered_features <- test_filters %>%
    # only look at "samples"
    dplyr::filter(!(Sample %in% c("q1", "q2", "q3"))) %>%
    filter_blank(min_frac = 2, blank_samples = "b1") %>%
    dplyr::select(2) %>%
    dplyr::pull() %>%
    unique()

  expect_equal(filtered_features, c("f1", "f2", "f3", "f4", "f10", "f11"))
})

test_that("works for min_frac = 1.5", {
  filtered_features <- test_filters %>%
    # only look at "samples"
    dplyr::filter(!(Sample %in% c("q1", "q2", "q3"))) %>%
    filter_blank(min_frac = 1.5, blank_samples = "b1") %>%
    dplyr::select(2) %>%
    dplyr::pull() %>%
    unique()

  expect_equal(filtered_features, c("f1", "f2", "f3", "f4", "f5", "f10", "f11"))
})

test_that("does not change the structure of the input in an unexpected way", {
  filtered_features <- test_filters %>%
    # only look at "samples"
    filter_blank(min_frac = 0, blank_samples = "b1")

  expect_equal(filtered_features, test_filters)
})

test_that("does not change the structure of the input in an unexpected way", {
  test_tibble <- tibble::tibble(
    UID = c(1, 1),
    Feature = c("f0", "f0"),
    Sample = c("s", "b"),
    Intensity = c(0, 0)
  )

  filtered_features <- test_tibble %>%
    # only look at "samples"
    filter_blank(min_frac = 0, blank_samples = "b")

  expect_equal(nrow(filtered_features), 0)
})


test_that("equal results for blank_as_group = TRUE and FALSE", {
  test_filters_group <- test_filters %>%
    dplyr::select("Sample") %>%
    dplyr::mutate(Group = stringr::str_remove_all(.data$Sample, "[0-9]{1,}")) %>%
    dplyr::distinct()

  filtered_features <- test_filters %>%
    join_metadata(test_filters_group) %>%
    # only look at "samples"
    dplyr::filter(!(Sample %in% c("q1", "q2", "q3")))

  filtered_features_group_5 <- filtered_features %>%
    filter_blank(min_frac = 5, blank_samples = "b", blank_as_group = TRUE, group_column = Group)

  filtered_features_ft_5 <- filtered_features %>%
    filter_blank(min_frac = 5, blank_samples = "b1", blank_as_group = FALSE)

  filtered_features_group_3 <- filtered_features %>%
    filter_blank(min_frac = 3, blank_samples = "b", blank_as_group = TRUE, group_column = Group)

  filtered_features_ft_3 <- filtered_features %>%
    filter_blank(min_frac = 3, blank_samples = "b1", blank_as_group = FALSE)

  filtered_features_group_2 <- filtered_features %>%
    filter_blank(min_frac = 2, blank_samples = "b", blank_as_group = TRUE, group_column = Group)

  filtered_features_ft_2 <- filtered_features %>%
    filter_blank(min_frac = 2, blank_samples = "b1", blank_as_group = FALSE)

  filtered_features_group_1_5 <- filtered_features %>%
    filter_blank(min_frac = 1.5, blank_samples = "b", blank_as_group = TRUE, group_column = Group)

  filtered_features_ft_1_5 <- filtered_features %>%
    filter_blank(min_frac = 1.5, blank_samples = "b1", blank_as_group = FALSE)

  expect_equal(filtered_features_group_5, filtered_features_ft_5)
  expect_equal(filtered_features_group_3, filtered_features_ft_3)
  expect_equal(filtered_features_group_2, filtered_features_ft_2)
  expect_equal(filtered_features_group_1_5, filtered_features_ft_1_5)
})

test_that("throws error if blank_as_group = TRUE and group_column = NULL", {
  test_filters_group <- test_filters %>%
    dplyr::select("Sample") %>%
    dplyr::mutate(Group = stringr::str_remove_all(.data$Sample, "[0-9]{1,}")) %>%
    dplyr::distinct()

  filtered_features <- test_filters %>%
    join_metadata(test_filters_group)

  expect_error(filter_blank(filtered_features, min_frac = 0, blank_samples = "b", blank_as_group = TRUE, group_column = NULL))
})

test_that("throws error if group_column is provided as character", {
  test_filters_group <- test_filters %>%
    dplyr::select("Sample") %>%
    dplyr::mutate(Group = stringr::str_remove_all(.data$Sample, "[0-9]{1,}")) %>%
    dplyr::distinct()

  filtered_features <- test_filters %>%
    join_metadata(test_filters_group)

  expect_error(filter_blank(filtered_features, min_frac = 0, blank_samples = "b", blank_as_group = TRUE, group_column = "Group"))
})

test_that("filter_blank() does not throw a warning if group = .data$Group", {
  expect_no_warning(toy_metaboscape %>% join_metadata(toy_metaboscape_metadata) %>% filter_blank(blank_samples = "blank", blank_as_group = T, group_column = .data$Group))
})


test_that("returns expected results if > 1 group is provided", {
  test_filters_group <- test_filters %>%
    dplyr::select("Sample") %>%
    dplyr::mutate(Group = dplyr::case_when(Sample %in% paste0("s", 0:6) ~ "sg1",
                                           Sample %in% paste0("s", 7:12) ~ "sg2",
                                           .default = "b")) %>%
    dplyr::distinct()

  filtered_features <- test_filters %>%
    join_metadata(test_filters_group) %>%
    # only look at "samples"
    dplyr::filter(!(Sample %in% c("q1", "q2", "q3")))

  filtered1 <- filtered_features %>%
    filter_blank(blank_samples = c("sg2", "b"), min_frac = 3, blank_as_group = T, group_column = Group)

  filtered2 <- filtered_features %>%
    filter_blank(blank_samples = c("sg2", "b"), min_frac = 2, blank_as_group = T, group_column = Group)

  filtered3 <- filtered_features %>%
    filter_blank(blank_samples = c("sg2", "b"), min_frac = 1.6, blank_as_group = T, group_column = Group)

  filtered4 <- filtered_features %>%
    filter_blank(blank_samples = c("sg2", "b"), min_frac = 1.4, blank_as_group = T, group_column = Group)

  filtered5 <- filtered_features %>%
    filter_blank(blank_samples = c("sg2", "b"), min_frac = 1, blank_as_group = T, group_column = Group)

  expected1 <- filtered_features %>%
    dplyr::filter(Feature %in% c("f1", "f2"))

  expected2 <- filtered_features %>%
    dplyr::filter(Feature %in% c("f1", "f2", "f3", "f4"))

  expected3 <- filtered_features %>%
    dplyr::filter(Feature %in% c("f1", "f2", "f3", "f4", "f5"))

  expected4 <- filtered_features %>%
    dplyr::filter(Feature %in% c("f1", "f2", "f3", "f4", "f5", "f6"))

  expected5 <- filtered_features %>%
    dplyr::filter(Feature != "f0")

  expect_equal(filtered1, expected1)
  expect_equal(filtered2, expected2)
  expect_equal(filtered3, expected3)
  expect_equal(filtered4, expected4)
  expect_equal(filtered5, expected5)
})

test_that("row & column order stays unchanged", {
  filtered_df <- test_filters %>%
    dplyr::mutate(Intensity = 1) %>%
    filter_blank(blank_samples = "b1", blank_as_group = F, min_frac = 1)

  expect_equal(filtered_df, dplyr::mutate(test_filters, Intensity = 1))
})