R/marker_intensity_boxplot.R

In TrigosTeam/SPIAT: Spatial Image Analysis of Tissues

#' marker_intensity_boxplot
#'
#' @description Produces boxplots of marker levels for cells phenotyped as being
#'   positive for the marker, and those that where phenotyped as being negative.
#' @param spe_object SpatialExperiment object in the form of the output of
#'   \code{\link{format_image_to_spe}}.
#' @param marker String. Marker being queried.
#' @param feature_colname String. Column containing marker information
#' @import dplyr
#' @import ggplot2
#' @return A plot is returned
#' @examples
#' marker_intensity_boxplot(SPIAT::simulated_image, "Immune_marker1")
#' @export

marker_intensity_boxplot <- function(spe_object, marker, feature_colname = "Phenotype"){
    # setting these variables to NULL as otherwise get "no visible binding for global variable" in R check
    intensity <- NULL

    formatted_data <- bind_info(spe_object)

    #selecting cells that do express the marker
    intensity_true <- formatted_data[grepl(marker, formatted_data[[feature_colname]]), ]

    #selecting cells that do not contain the marker
    intensity_false <- formatted_data[!grepl(marker, formatted_data[[feature_colname]]), ] #for multiple entries that does not contain marker

    #select the specific marker and add a boolean of intensity
    if (nrow(intensity_true) != 0) {
        intensity_true$intensity <- "P"
    } else{
        stop(sprintf("There are no cells positive for %s", marker))
    }

    if (nrow(intensity_false) != 0){
        intensity_false$intensity <- "N"
    } else{
        stop(sprintf("There are no cells negative for %s", marker))
    }

    #bind the 2 dataframes together
    intensity_by_marker <- rbind(intensity_true, intensity_false)
    intensity_by_marker$intensity <- factor(intensity_by_marker$intensity, levels=c("P", "N"))
    
    #plot boxplot
    title <- paste("Level of ", marker, sep="")

    #code from ggplot2 to show the number in each group
    give.n <- function(x){
        return(c(y = max(x)+1, label = length(x)))
    }

    p <- ggplot(intensity_by_marker, aes(x = intensity, y = intensity_by_marker[, marker]))
    p <- p + geom_boxplot()
    p <- p + labs(title = title, x = "Marker status (Positive/Negative cell)", y = "Marker level")
    p <- p + stat_summary(fun.data = give.n, geom = "text")
    p <- p + theme_bw()

    # methods::show(p)
    return(p)
}