R/scaleByGroup.R

In wolski/quantable: Streamline Descriptive Analysis of Quantitative Data Matrices

#' scale data given group 
#' 
#' @param data matrix
#' @param protGroup names matching rownames of data
#' @param plot plot diagnostics
#' @param scale scale default TRUE
#' @param center center default TRUE
#' @export
#' 
scaleByGroup <- function(data , protGroup, plot=FALSE, scale=TRUE, center=TRUE){
  reference = data[rownames(data) %in% protGroup,]
  noReference = data[!rownames(data) %in% protGroup,]
  
  referenceScaled = robustscale(reference,scale=scale, center=center)
  
  if(center){
    noReference = sweep(noReference,2,referenceScaled$medians,"-")
  }
  if(scale){
    noReference = sweep(noReference,2,referenceScaled$mads,"/")
  }
  if(plot){
    graphics::par(mfrow=c(1,2))
    graphics::boxplot(noReference,main="noReference",ylim=c(-8,6), pch=".", las=2,cex.axis=0.5)
    abline(h=0,col=2)
    graphics::boxplot(referenceScaled$data,main="reference",ylim=c(-8,6),las=2,pch=".",cex.axis=0.5)
    abline(h=0,col=2)
  }
  return(list(reference = referenceScaled$data, noReference = noReference))
}