tests/testthat/test-statistics.R

In EmanuelSoda/ScreenR: Package to Perform High Throughput Biological Screening

data("count_table", package = "ScreenR")
data("annotation_table", package = "ScreenR")
data <- count_table
annotaion <- annotation_table

groups <- factor(c(
    "T1/T2", "T1/T2", "Treated", "Treated", "Treated",
    "Control", "Control", "Control", "Treated", "Treated", "Treated",
    "Control", "Control", "Control"
))


palette <- c(
    "#1B9E75", "#1B9E75", "#D95F02", "#D95F02", "#D95F02",
    "#7570B3", "#7570B3", "#7570B3", "#E7298A", "#E7298A", "#E7298A",
    "#66A61E", "#66A61E", "#66A61E"
)

create_test_object <- function() {
    data <- data %>%
        dplyr::filter(Barcode != "*")

    colnames(data) <- c(
        "Barcode", "T1", "T2", "Time3_TRT_A", "Time3_TRT_B", "Time3_TRT_C",
        "Time3_A", "Time3_B", "Time3_C", "Time4_TRT_A", "Time4_TRT_B",
        "Time4_TRT_C", "Time4_A", "Time4_B", "Time4_c"
    )
    obj <- create_screenr_object(
        table = data,
        annotation = annotaion, groups = groups, replicates = c("")
    )
    obj <- normalize_data(obj)
    obj <- compute_data_table(obj)

    obj@data_table <- obj@data_table %>%
        dplyr::filter(Gene %in% paste0("Gene_", seq(1, 10)))

    obj@normalized_count_table <- obj@normalized_count_table %>%
        dplyr::filter(Barcode %in% obj@data_table$Barcode)

    obj@count_table <- obj@count_table %>%
        dplyr::filter(Barcode %in% obj@data_table$Barcode)

    obj@annotation_table <- obj@annotation_table %>%
        dplyr::filter(Barcode %in% obj@data_table$Barcode)

    return(obj)
}
test_that("ROAST", {
    object <- create_test_object()
    matrix <- model.matrix(~ object@groups)
    colnames(matrix) <- c("Control", "T1/T2", "Treated")

    roast_hit <- suppressWarnings(find_roast_hit(
        screenR_Object = object,
        matrix_model = matrix, contrast = "Treated"
    ))
    expect_equal(class(roast_hit)[[1]], "tbl_df")
})


test_that("Camera", {
    object <- create_test_object()

    matrix <- model.matrix(~ object@groups)
    colnames(matrix) <- c("Control", "T1/T2", "Treated")
    camera_hit <- suppressWarnings(find_camera_hit(
        screenR_Object = object,
        matrix_model = matrix, contrast = "Treated"
    ))
    expect_equal(class(camera_hit)[1], "tbl_df")
})



test_that("Hit Z-score per giorno", {
    object <- create_test_object()

    # In order to speed up the test we will compute the metrics only for a
    # subset of the genes
    genes <- c("Gene_1", "Gene_5")
    object@data_table <- object@data_table[object@data_table$Gene %in% genes, ]

    table <- compute_metrics(object,
        treatment = "TRT", control = "Time3",
        day = "Time3"
    )

    hit_table <- find_zscore_hit(table, number_barcode = 2)
    expect_equal(class(hit_table)[1], "tbl_df")
})


test_that("Find_Zscore_hit mean", {
    object <- create_test_object()

    genes <- c("Gene_1", "Gene_5")
    object@data_table <- object@data_table[object@data_table$Gene %in% genes, ]

    table <- compute_metrics(object,
        control = "Time3", treatment = "TRT",
        day = c("Time3")
    )

    hit_zscore <- find_zscore_hit(table, number_barcode = 2, metric = "mean")

    expect_equal(class(hit_zscore)[[1]], "tbl_df")
})

test_that("Find_Score_hit median ", {
    object <- create_test_object()

    genes <- c("Gene_1", "Gene_5")
    object@data_table <- object@data_table[object@data_table$Gene %in% genes, ]

    table <- compute_metrics(object,
        control = "Time3", treatment = "TRT",
        day = c("Time3")
    )
    hit_zscore <- find_zscore_hit(table, number_barcode = 4)

    expect_equal(class(hit_zscore)[[1]], "tbl_df")
})

test_that("find_robust_zscore_hit median ", {
    object <- create_test_object()

    genes <- c("Gene_1", "Gene_5")
    object@data_table <- object@data_table[object@data_table$Gene %in% genes, ]

    table <- compute_metrics(object,
        control = "Time3", treatment = "TRT",
        day = c("Time3")
    )
    hit_zscore_R <- find_robust_zscore_hit(table, number_barcode = 2)
    expect_equal(class(hit_zscore_R)[[1]], "grouped_df")
})