R/boxplot.R

In markdunning/beadarray: Quality assessment and low-level analysis for Illumina BeadArray data

# setMethod("boxplot",
#     signature(x = "beadLevelData"),
#     function (x, transFun=logGreenChannelTransform,...) 
#     {
# 
#   	arraynms = sectionNames(x)
#   	narrays = length(arraynms)
# 
# 	inten <- unlist(lapply(1:narrays, function(i) transFun(x, array=i)))
# 	section <- NULL
# 
# 	for(i in 1:narrays){
# 
# 		section <- c(section, rep(sectionNames(x)[i], numBeads(x)[i]))
# 	}
# 
# 	df <- data.frame(Value = inten, Section = section)
# 
# 	p <- ggplot(df, aes(x = factor(Section), y = Value, fill=factor(Section))) + geom_boxplot(outlier.shape=NA)
# 	p
# }
# )
setGeneric("boxplot", function(x,...) standardGeneric("boxplot"))

setMethod("boxplot",
    signature(x = "beadLevelData"),
    function (x, transFun=logGreenChannelTransform,...) 
    {
       tmp = list()
        arraynms = sectionNames(x)
        narrays = length(arraynms)


        for(i in 1:narrays){
            tmp[[arraynms[i]]] = transFun(x, array=i)
        }
        boxplot(tmp,...)   
    }
)




setMethod("boxplot",
    signature(x = "ExpressionSetIllumina"),
    function (x, what="exprs",probeFactor = NULL, SampleGroup=NULL, facet = NULL,plot=FALSE,...) 
    {
	
	addedSampleGroup <- addedProbeFactor <- FALSE
	

	data <- melt(assayDataElement(x, what))

	if(!is.null(SampleGroup)){

		if(SampleGroup %in% colnames(pData(x))){
		
		data <- data.frame(data, SampleGroup = pData(x)[match(data[,"Var2"], rownames(pData(x))),SampleGroup])
		addedSampleFactor <- TRUE

	

		}

		else message("Could not find a phenoData column called " , SampleGroup)
	}	

	if(!is.null(probeFactor)){
		if(probeFactor %in% colnames(fData(x))){
		
		data <- data.frame(data, probeFactor = fData(x)[match(data[,"Var1"], rownames(fData(x))),probeFactor])
		addedProbeFactor <- TRUE
		}
    
		else message("Could not find a featureData column called " , probeFactor)
	}	


	##Traditional boxplot of all probes on every array vs array.	
	
	if(!(addedSampleGroup) & !(addedProbeFactor)){
	
	p <- ggplot(data, aes(x = factor(Var2), y = value)) + geom_boxplot(outlier.shape=NA,fill="steelblue") 

	}

	##Change x axis to vary with the new probe factor	

	else if(addedProbeFactor){
		
		p <- ggplot(data, aes(x = factor(probeFactor), y = value, fill=factor(probeFactor))) + geom_boxplot(outlier.shape=NA) +  scale_fill_discrete(name=probeFactor)

		if(addedSampleGroup){

			p <- p + facet_wrap(~SampleGroup)
		}

		else p <- p + facet_wrap(~Var2)

	}
	

	##change x axis to vary with new sample factor
	else if(addedSampleGroup){

		p <- ggplot(data, aes(x = factor(SampleGroup), y = value, fill=factor(SampleGroup))) + geom_boxplot(outlier.shape=NA) + scale_fill_discrete(name=SampleGroup)
			
		

	}
	
	p + theme(axis.title.x = element_blank(),axis.text.x  = element_text(angle=90),axis.title.y=element_text(angle=0)) + ylab(what)

}
)