R/boxPlot.R

In zzxxyui/metadarclean: Metabolomics data analysis

setGeneric("boxPlot", signature=c("Object", "covariate1", "covariate2"),
  function(Object, covariate1, covariate2, file="boxplots.pdf")
    standardGeneric("boxPlot"))

setMethod("boxPlot", signature=c("Dataset", "character", "missing"),
  function(Object, covariate1, file="boxplots.pdf") {
    pdf(file)
    for(i in seq(nrow(Object))) {
      fit <- aov(x~f, data=data.frame("x"=exprs(Object)[i,],
        "f"=factor(pData(Object)[,covariate1])
      ))
      boxplot(exprs(Object)[i,]~factor(pData(Object)[,covariate1]),
        main=paste("Name:", featureNames(Object)[i],
          "\nAnova P-value:",
          signif(summary(fit)[[1]]["f","Pr(>F)"], 2)))
    }
    dev.off()
  })

setMethod("boxPlot", signature=c("Dataset", "character", "character"),
  function(Object, covariate1, covariate2, file="boxplots.pdf") {
    pdf(file)
    previous <- par(mai=c(0.5,0.5,1,0.2))
    for(i in seq(nrow(Object))) {
      fit <- aov(x~f1*f2, data=data.frame("x"=exprs(Object)[i,],
        "f1"=factor(pData(Object)[,covariate1]),
        "f2"=factor(pData(Object)[,covariate2])
      )
      )
      p1 <- signif(summary(fit)[[1]][1,"Pr(>F)"], 2)
      p2 <- signif(summary(fit)[[1]][2,"Pr(>F)"], 2)
      p3 <- signif(summary(fit)[[1]][3,"Pr(>F)"], 2)
      
      boxplot(exprs(Object)[i,]~factor(paste(pData(Object)[,covariate1],
        pData(Object)[,covariate2])),
        main=paste("Name:", featureNames(Object)[i],
          "\n", covariate1, "p-value", p1,
          "\n", covariate2, "p-value", p2,
          "\n", covariate1, covariate2, "interaction p-value", p3))   
    }
    par(previous)
    dev.off()
  })