Adding MAE wrapper functions

.gitignore

@@ -4,3 +4,4 @@
 .Ruserdata
 doc
 Meta
+inst/doc

NAMESPACE

@@ -12,6 +12,7 @@ export(compileFeatureScores)
 export(compileFeatures)
 export(confusionMatrix)
 export(convertProfileToNetworks)
+export(convertToMAE)
 export(countIntType)
 export(countIntType_batch)
 export(countPatientsInNet)

R/buildPredictor.R

@@ -247,8 +247,13 @@ for (k in seq_len(length(exprs))) {
 	tmp <- exprs[[k]]
 	df <- sampleMap(dataList)[which(sampleMap(dataList)$assay==names(exprs)[k]),]
 	colnames(tmp) <- df$primary[match(df$colname,colnames(tmp))]
-	tmp <- as.matrix(assays(tmp)[[1]]) # convert to matrix
-	datList2[[names(exprs)[k]]]<- tmp	
+	if ("matrix" %in% class(tmp)) {
+		datList2[[names(exprs)[k]]] <- tmp
+	} else {
+		tmp <- as.matrix(assays(tmp)[[1]]) # convert to matrix
+		datList2[[names(exprs)[k]]]<- tmp	
+	}
+
 }
 if ("clinical" %in% names(groupList)) {
 	tmp <- colData(dataList)

R/convertToMAE.R

@@ -0,0 +1,84 @@
+#' Wrapper that converts an input list into a MultiAssayExperiment object
+#' 
+#' @details This function takes in a list of key-value pairs (keys: data types,
+#' values: matrices/dataframes) and calls the necessary functions from the
+#' MultiAssayExperiment package to incorporate the values from the input list 
+#' into a MultiAssayExperiment object, transforming the values according to the 
+#' keys. If duplicate sample names are found in the assay data, only the first
+#' instance is kept.
+#' @param dataList  (list) input key-value pairs (keys: data types, values: 
+#' data in the form of matrices/dataframes); must have a key-value pair that
+#' corresponds to patient IDs/metadata labelled pheno.
+#' @return MAE (MultiAssayExperiment) data from input list incorporated into a
+#' MultiAssayExperiment object, compatible with further analysis using the 
+#' netDx algorithm.
+#' @examples
+#' data(xpr, pheno)
+#' 
+#' # Generate random proteomic data
+#' prot <- matrix(rnorm(100*20), ncol=20)
+#' colnames(prot) <- sample(pheno$ID, 20)
+#' rownames(prot) <- sprintf("protein%i",1:100)	
+#' 
+#' myList <- list(rna = xpr, proteomic = prot, pheno = pheno)
+#' 
+#' MAE <- convertToMAE(myList)
+#' @export
+
+
+convertToMAE <- function(dataList) {
+
+  # Check input data:
+  if (class(dataList) != "list") {
+    stop("dataList must be a list. \n")
+  }
+  if (is.null(dataList$pheno)) {
+    stop("dataList must have key-value pair labelled pheno.\n")
+  }
+  if (length(dataList) == 1) {
+    stop("dataList must have assay data to incorporate into a 
+         MultiAssayExperiment object")
+  }
+
+  # Note that a MultiAssayExperiment object requires an ExperimentList and 
+  # colData (sampleMap optional if each assay uses the same colnames)
+
+  # Possible elements for ExperimentList:
+  # - base::matrix (gene expression, microRNA, metabolomics, microbiome data)
+  # - SummarizedExperiment::SummarizedExperiment (same as matrix, but capable
+  #   of storing additional assay-level metadata)
+  # - Biobase::ExpressionSet (legacy representation, use SummarizedExperiment)
+  # - SummarizedExperiment::RangedSummarizedExperiment (range-based datasets; 
+  #   gene expression, methylation, data types that refer to genomic positions)
+  # - RaggedExperiment::RaggedExperiment (range-based datasets; copy number and
+  #   mutation data, measurements by genomic positions)
+
+  # Assumes that pheno is a DataFrame (or coerceable to be a DataFrame)
+  patientPheno <- dataList$pheno
+
+  # Generate ExperimentList from input dataList
+  tmp <- NULL
+  track <- c()
+  datType <- names(dataList)
+  for (k in 1:length(dataList)) {
+    # For key-value pairs that aren't labelled pheno, transform into 
+    # objects compatible with input into MultiAssayExperiment object
+    if (names(dataList[k]) != "pheno") {
+
+      # Remove duplicated columns (we keep the first column) in the assay data
+      if (sum(duplicated(colnames(dataList[[k]]))) != 0) {
+        dataList[[k]] <- dataList[[k]][,!duplicated(colnames(dataList[[k]]))]
+      }
+
+      # Assumes that data is of matrix class 
+      # *(maybe implement matrix conversion into SummarizedExperiment in future)
+      track <- c(track, k)
+      tmp <- c(tmp, list(dataList[[k]]))
+    }
+  }
+  names(tmp) <- datType[track]
+
+  MAE <- MultiAssayExperiment(experiments = tmp, colData = patientPheno)
+
+  return(MAE)
+}

tests/testthat/test_convertToMAE.R

@@ -0,0 +1,75 @@
+# test convertToMAE.R
+
+test_that("convertToMAE works", {
+    # 20 patients, 10 case, 10 control
+	pheno <- data.frame(ID=sprintf("PAT%i",1:20),
+		STATUS=rep(c("case","control"),each=10))
+	# 100 dummy genes
+	rna <- matrix(rnorm(100*20),nrow=100); 
+	colnames(rna) <- pheno$ID
+	rownames(rna) <- sprintf("gene%i",1:100)	
+	# 2 dummy clin variables
+	clin <- t(data.frame(AGE=runif(20,10,50)))
+	colnames(clin) <- pheno$ID
+    clin <- t(clin)
+
+	# netDx files
+	dataList <- list(rna=rna,pheno=clin)
+
+    x <- convertToMAE(dataList)
+    expect_is(x, "MultiAssayExperiment")
+})
+
+test_that("convertToMAE works with more than one assay", {
+	# 20 patients, 10 case, 10 control
+	pheno <- data.frame(ID=sprintf("PAT%i",1:20),
+		STATUS=rep(c("case","control"),each=10))
+	# 100 dummy genes
+	rna <- matrix(rnorm(100*20),nrow=100); 
+	colnames(rna) <- pheno$ID
+	rownames(rna) <- sprintf("gene%i",1:100)
+	# 200 dummy proteins
+	prot <- matrix(rnorm(200*20), nrow = 200);
+	colnames(prot) <- pheno$ID
+	rownames(prot) <- sprintf("protein%i",1:200) 	
+	# 2 dummy clin variables
+	clin <- t(data.frame(AGE=runif(20,10,50)))
+	colnames(clin) <- pheno$ID
+    clin <- t(clin)
+
+	# netDx files
+	dataList <- list(rna = rna, proteomics = prot, pheno = clin)
+
+	x <- convertToMAE(dataList)
+	expect_is(x, "MultiAssayExperiment")
+})
+
+test_that ("convertToMAE removes duplicated sample", {
+	# 20 patients, 10 case, 10 control
+	pheno <- data.frame(ID=sprintf("PAT%i",1:20),
+                    STATUS=rep(c("case","control"),each=10))
+	# 100 dummy genes, with first sample duplicated
+	rna <- matrix(rnorm(100*20),nrow=100); 
+	colnames(rna) <- pheno$ID
+	rownames(rna) <- sprintf("gene%i",1:100)
+	rna <- cbind(rna, rna[,1])
+	colnames(rna)[21] <- colnames(rna)[1]
+	# 200 dummy proteins
+	prot <- matrix(rnorm(200*20), nrow = 200);
+	colnames(prot) <- pheno$ID
+	rownames(prot) <- sprintf("protein%i",1:200) 	
+	# 2 dummy clin variables
+	clin <- t(data.frame(AGE=runif(20,10,50)))
+	colnames(clin) <- pheno$ID
+	clin <- t(clin)
+
+	# netDx files
+	dataList <- list(rna = rna, proteomics = prot, pheno = clin)
+
+	x <- convertToMAE(dataList)
+	expect_is(x, "MultiAssayExperiment")
+	# number of samples in rna assay vs colData should differ by 1
+	expect_equal((dim(rna)[2] - dim(colData(x))[1]), 1)
+	# number of samples in metadata should agree with colData
+	expect_equal((dim(clin)[1] - dim(colData(x))[1]), 0)
+})