moved up qc section

dge_workflow/featcounts_deseq.R

@@ -69,7 +69,7 @@ genesAfter <- nrow(countMatrix)
 
 
 ########################################################################################################################
-#' ## Perform Differential Expression Analysis
+#' ## Data Preparation
 
 
 #' DESeq is an R package to analyse count data from high-throughput sequencing assays such as RNA-Seq and test for differential expression. The latest version, DESeq2, is used.
@@ -120,31 +120,6 @@ dds <- DESeqDataSetFromMatrix(countMatrix, colData, formula(~ condition))
 dds <- estimateSizeFactors(dds)
 
 
-#' For details about size factor normalziation and calculation see https://www.biostars.org/p/79978/
-
-##' Size Factors estimated by DESeq
-sizeFactors(dds) %>%
-    set_names(colnames(countMatrix)) %>% melt() %>%
-    rownames2column("sample") %>%
-    ggplot(aes(sample, value)) + geom_bar(stat="identity") + ggtitle("deseq size factors") + coord_flip()
-
-##' From the DESeq docs about how size factors are used: The sizeFactors vector assigns to each column of the count matrix a value, the size factor, such that count values in the columns can be brought to a common scale by dividing by the corresponding size factor.
-
-##' This means that counts are divied by size factors. So let's now load the lambda libraies and replace the predefined size factors with our custom ones
-
-#' From DESeq manual: The regularized log transformation is preferable to the variance stabilizing transformation if the size factors vary widely.
-
-#' Run Deseq Test
-#dds <- DESeq(dds, fitType='local', betaPrior=FALSE)
-#dds <- DESeq(dds, fitType='local')
-dds <- DESeq(dds)
-
-#res <- results(dds)
-#resultsNames(dds)
-
-#' Model Overview:
-summary(results(dds))
-
 ########################################################################################################################
 #' ## Quality Control
 
@@ -193,7 +168,34 @@ heatmap.2(distMatrix, labRow=colnames(labelcntData), labCol=colnames(labelcntDat
           #key=F, col=colorpanel(100, "black", "white"),
           margin=c(8, 8), main="Sample Distance Matrix")
 
+
 ########################################################################################################################
+#' # Perform Differential Expression Analysis
+
+#' For details about size factor normalziation and calculation see https://www.biostars.org/p/79978/
+
+##' Size Factors estimated by DESeq
+sizeFactors(dds) %>%
+    set_names(colnames(countMatrix)) %>% melt() %>%
+    rownames2column("sample") %>%
+    ggplot(aes(sample, value)) + geom_bar(stat="identity") + ggtitle("deseq size factors") + coord_flip()
+
+##' From the DESeq docs about how size factors are used: The sizeFactors vector assigns to each column of the count matrix a value, the size factor, such that count values in the columns can be brought to a common scale by dividing by the corresponding size factor.
+
+##' This means that counts are divied by size factors. So let's now load the lambda libraies and replace the predefined size factors with our custom ones
+
+#' From DESeq manual: The regularized log transformation is preferable to the variance stabilizing transformation if the size factors vary widely.
+
+#' Run Deseq Test
+#dds <- DESeq(dds, fitType='local', betaPrior=FALSE)
+#dds <- DESeq(dds, fitType='local')
+dds <- DESeq(dds)
+
+#res <- results(dds)
+#resultsNames(dds)
+
+#' Model Overview:
+summary(results(dds))
 
 
 ## extract all de-sets according to our contrasts model