Replaced probe ids by probe names for Agilent data processing.

R/preprocess_data_agilent_limma.R

@@ -40,6 +40,9 @@ preprocess_data_agilent_limma <- function(input_data_dir, output_data_file,
                                 green.only = TRUE,
                                 verbose    = TRUE)
   batch_data <- log2(batch$E)
+
+  # We change the probe ids (1 to 45015) to be the probe names instead (A_XX_PXXXX).
+  rownames(batch_data) <- (batch$genes)$ProbeName
   remove(raw_data_input_dir, batch)
 
   # We run the LIMMA pre-processing method on the data.
@@ -61,6 +64,11 @@ preprocess_data_agilent_limma <- function(input_data_dir, output_data_file,
     batch_data_norm <- batch_data_norm[, samples]
   }
 
+  # We remove the duplicate rows (based on the row names only).
+  probe_id_counts  <- table(row.names(batch_data))
+  unique_probe_ids <- setdiff(names(probe_id_counts), names(probe_id_counts[probe_id_counts > 1]))
+  batch_data_norm  <- batch_data_norm[unique_probe_ids, ]
+
   # We save the eset data as TSV file.
   eset <- methods::new("ExpressionSet", exprs = as.matrix(batch_data_norm))
   rm(batch_data_bg)

R/read_eset.R

@@ -7,18 +7,21 @@
 #' @param data_file A string representing the file that contains the expression data.
 #' @param header A boolean indicating whether the file contains a header (default to TRUE).
 #' @param sep A string that is used as a field separator to read the data (default to
-#'  tab for TSV files).
+#' tab for TSV files).
 #' @param row_names The index of the row names (default to 1).
 #' @param as_is A boolean indicating whether R should keep the data as they are in the
-#'  file (default to TRUE).
+#' file (default to TRUE).
+#' @param as_eset A boolean indicating whether the returned object should be an eset. A matrix
+#' is returned otherwise. Default to TRUE.
 #' @param verbose A boolean representing whether the function should display log information. This
-#'  is TRUE by default.
-#' @return An ExpressionSet containing the preprocessed expression data.
+#' is TRUE by default.
+#' @return An object containing the expression data. Either an ExpressionSet or a Matrix.
 read_eset <- function(data_file,
                       header    = TRUE,
                       sep       = "\t",
                       row_names = 1,
                       as_is     = TRUE,
+                      as_eset   = TRUE,
                       verbose   = TRUE) {
 
   # We load the matrix and creates the associated eset object.
@@ -30,11 +33,14 @@ read_eset <- function(data_file,
   exprs_eset <- Biobase::ExpressionSet(assayData = exprs_mat)
 
   # We clean up and log information.
-  rm(exprs_mat)
   if (verbose == TRUE) {
     message(paste0("[", Sys.time(), "] Expression data read."))
   }
 
-  # We return the eset.
-  return(exprs_eset)
+  # We return the object.
+  if (as_eset) {
+    return(exprs_eset)
+  } else {
+    return(exprs_mat)
+  }
 }