R/plot.fdr_table.R

In peterblattmann/SWATH2stats: Transform and Filter SWATH Data for Statistical Packages

#' S3 plot function for results of class "fdr_table" as produced by e.g. the
#' function assess_fdr_overall()
#'
#' This function created standard plots from results of class "fdr_table" as
#' produced by e.g. the function assess_fdr_overall() visualizig ID numbers in
#' dependence of estimated FDR and also estimated FDR in dependence of m_score
#' cutoff.
#'
#' @param x  List of class "fdr_table" as produced e.g. by the function
#'   assess_fdr_overall() from this package.
#' @param output Choose output type. "pdf_csv" creates the output as files in
#'   the working directory, "Rconsole" triggers delivery of the output to the
#'   console enabling further computation or custom plotting / output.
#' @param filename  Basename for output files to be created (if output =
#'   "pdf_csv" has been selected).
#' @param ...  Extra arguments passed on to functions inside this.
#' @return  Plots in Rconsole or report files.
#' @author Moritz Heusel
#' @examples{
#'  data("OpenSWATH_data", package="SWATH2stats")
#'  data("Study_design", package="SWATH2stats")
#'  data <- sample_annotation(OpenSWATH_data, Study_design)
#'  x <- assess_fdr_overall(data, FFT=0.7, output="Rconsole", plot=FALSE)
#'  plot(x, output="Rconsole", filename="Assess_fdr_overall_testplot")
#'  }
#' @importFrom graphics par lines legend mtext points axis grid abline
#' @importFrom grDevices pdf dev.off
#' @export
plot.fdr_table <- function(x, 
                           output = "Rconsole", 
                           filename = "FDR_report_overall",
                           ...) {
    ## Plot and create output from ID-FDR report (x) ## Plot 1: Target/true target
    ## curves as estimated by decoy counting + FFT-correction

    # Save previous par(mfrow) settings and restore upon exit (BioConductor
    # suggestion)
    par.mfrow.old <- par(mfrow = par()$mfrow)
    on.exit(par(par.mfrow.old), add = TRUE)
    par.mar.old <- par(mar = par()$mar)
    on.exit(par(par.mar.old), add = TRUE)

    # Open pdf if output as pdf was desired by user
    if (output == "pdf_csv") {
        pdf(file = paste0(filename, ".pdf"), height = 6, width = 10)
        par(mfrow = c(1, 3), mar = c(13, 4, 15, 2) + 0.1)
    }

    # plot 1.1 Assay-level sensitivity
    plot(x$assay.fdr, x$target.assays, 
         xlab = "assay FDR", ylab = "# of assays",
         type = "l", lty = 2, 
         xlim = c(0, max(x$assay.fdr, na.rm = TRUE)), 
         ylim = (c(0, max(x$target.assays))))
    lines(x$assay.fdr, x$true.target.assays, xlim = c(0, max(x$assay.fdr)))
    legend("topleft", legend = c("all targets", "true targets"), 
           cex = 0.5, lty = c(2,1))

    # plot 1.2 Peptide-level sensitivity
    plot(x$peptide.fdr, x$target.peptides, 
         xlab = "peptide FDR", ylab = "# of peptides",
         type = "l", lty = 2, 
         xlim = c(0, max(x$peptide.fdr, na.rm = TRUE)), 
         ylim = (c(0, max(x$target.peptides))))
    lines(x$peptide.fdr, x$true.target.peptides, 
          xlim = c(0, max(x$peptide.fdr)))
    legend("topleft", legend = c("all targets", "true targets"), 
           cex = 0.5, lty = c(2, 1))

    if (output == "pdf_csv") {
        mtext("SWATH2stats global FDR & sensitivity report of OpenSWATH/pyProphet results",
            line = 8, cex = 1.5)
        mtext("Overall sensitivity:", line = 3, adj = 0.5, cex = 1.1)
    }

    # plot 1.3 Protein-level sensitivity
    plot(x$protein.fdr, x$target.proteins, 
         xlab = "protein FDR", ylab = "# of proteins",
         type = "l", lty = 2, 
         xlim = c(0, max(x$protein.fdr, na.rm = TRUE)),
         ylim = (c(0, max(x$target.proteins))))
    lines(x$protein.fdr, x$true.target.proteins, 
          xlim = c(0, max(x$protein.fdr)))
    legend("topleft", legend = c("all targets", "true targets"), 
           cex = 0.5, lty = c(2, 1))

    # Plot 2: Global m_score adjustment and connectivity to global FDR quality 
    # levels as estimated by decoy counting + FFT-correction
    xlimit <- -1 * (length(x$assay.fdr) + 1)

    par(mar = c(5, 8, 4, 8) + 0.1, mfrow = c(1, 1))
    plot(log10(x$mscore_cutoff), x$assay.fdr, axes = FALSE, 
         ylim = c(0, 1.1 * max(x$assay.fdr, na.rm = TRUE)), 
        main = "Global m-score cutoff connectivity to FDR quality",
        xlab = "", ylab = "", 
        type = "l", col = "black", 
        xlim = c(xlimit, 0))
    points(log10(x$mscore_cutoff), x$assay.fdr, pch = 20, col = "black")
    axis(2, ylim = c(0, 1.1 * max(x$assay.fdr, na.rm = TRUE)), 
         col = "black", lwd = 2)
    
    mtext(2, text = "Assay FDR", line = 2)
    par(new = TRUE)
    plot(log10(x$mscore_cutoff), x$peptide.fdr, 
         axes = FALSE, 
         ylim = c(0, 1.1 * max(x$peptide.fdr, na.rm = TRUE)), 
         xlab = "", ylab = "", type = "l", lwd = 2, col = "red", main = "",
         xlim = c(xlimit, 0), lty = 2)
    points(log10(x$mscore_cutoff), x$peptide.fdr, pch = 20, col = "red")
    axis(2, ylim = c(0, 1.1 * max(x$peptide.fdr, na.rm = TRUE)), 
         col = "red", lwd = 2, line = 3.5)
    
    mtext(2, text = "Peptide FDR", col = "red", line = 5.5)
    par(new = TRUE)
    plot(log10(x$mscore_cutoff), x$protein.fdr, axes = FALSE, 
         ylim = c(0, 1.1 * max(x$protein.fdr, na.rm = TRUE)), 
         col = "blue", xlab = "", ylab = "", type = "l", lty = 3,
         main = "", xlim = c(xlimit, 0), lwd = 2)
    points(log10(x$mscore_cutoff), x$protein.fdr, col = "blue", pch = 20)
    axis(4, ylim = c(0, 1.1 * max(x$protein.fdr, na.rm = TRUE)), 
         col = "blue", lwd = 2)
    mtext(4, text = "Protein FDR", col = "blue", line = 2)
    axis(1, at = seq(min(log10(x$mscore_cutoff)), max(log10(x$mscore_cutoff))))
    grid()
    abline(v = seq(min(log10(x$mscore_cutoff)), max(log10(x$mscore_cutoff))), 
           col = "grey", lty = 3)
    mtext(1, text = "log10(m_score cutoff)", col = "black", line = 2)
    
    if (output == "pdf_csv") {
        dev.off()
        message(filename, ".pdf written to working folder", "\n")
    }
}