R/compute_data_table.R

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

#' @title Compute data Table
#' @description This function computes the data table that will be used
#'              for the analysis. The data_table is a tidy and normalized
#'              version of the original count_table and will be used
#'              throughout the analysis.
#' @param screenR_Object The ScreenR object obtained using the
#'                       \code{\link{create_screenr_object}}
#'
#' @return ScreenR_Object with the data_table filed containing the table.
#' @export
#' @importFrom rlang .data
#' @importFrom stringr str_split_fixed
#' @importFrom  tidyr pivot_longer
#' @importFrom  dplyr left_join
#' @concept compute
#' @examples
#' object <- get0("object", envir = asNamespace("ScreenR"))
#' object <- compute_data_table(object)
#' head(slot(object, "data_table"))
compute_data_table <- function(screenR_Object) {
    # First the table is created with the join of the annotation and the count
    # table
    data <- screenR_Object@normalized_count_table
    table <- data %>%
        tidyr::pivot_longer(!.data$Barcode,
            names_to = "Sample",
            values_to = "Frequency"
        ) %>%
        dplyr::mutate(Barcode = as.factor(.data$Barcode)) %>%
        dplyr::left_join(screenR_Object@annotation_table, by = "Barcode") %>%
        select(
            .data$Barcode, .data$Gene, .data$Sample, .data$Frequency,
            .data$Sequence, .data$Library, .data$Gene_ID
        ) %>%
        mutate(
            Day = str_split_fixed(.data$Sample, pattern = "_", n = 3)[, 1],
            Treatment = gsub(
                ".*_", "",
                gsub("(.*)_\\w+", "\\1", .data$Sample)
            )
        )

    table <- dplyr::mutate(table, Sample = factor(
        x = .data$Sample,
        levels = unique(.data$Sample)
    ))

    # Then the table is put into the object
    screenR_Object@data_table <- table
    return(screenR_Object)
}