Replace GOenrichment_proj.R

scripts/GOenrichment_proj.R

+.libPaths(c('/local/data/course/project/araformatics/pipeline/test_R',.libPaths()))
 #!/usr/bin/env Rscript
-
 #BIF30806 Project group: Araformatic
 #(2021)
 #GO enrichment
@@ -7,11 +7,11 @@
 #expression analysis
 # Scrip written based on:
 #    https://ucdavis-bioinformatics-training.github.io/2018-June-RNA-Seq-Workshop/friday/enrichment.html
-#    Documentation available of topGO usage
-#    https://www.r-bloggers.com/2015/09/passing-arguments-to-an-r-script-from-command-lines/
+#    Documentation available on topGO usage
 #Usage: Rscript GOenrichment_proj.R input_file.txt
 #    input_file: txt file, genes as rows, leafs as columns and values
 #    are adjusted p value from DGE
+#Output: txt file and a pdf GO graph for each leaf(sample) in the input file
 
 #Take arguments from command line
 args = commandArgs(trailingOnly = TRUE)
@@ -26,10 +26,10 @@ if (length(args)==0) {
 
 #Install packages from source.
 
-source("https://bioconductor.org/biocLite.R")
-biocLite(c("topGO", "KEGGREST", "org.At.tair.db", "Rgraphviz"))
-if (!requireNamespace("BiocManager", quietly = TRUE))
-  install.packages("BiocManager")
+#source("https://bioconductor.org/biocLite.R")
+#biocLite(c("topGO", "KEGGREST", "org.At.tair.db", "Rgraphviz"))
+#if (!requireNamespace("BiocManager", quietly = TRUE))
+#  install.packages("BiocManager")
 
 #BiocManager::install("topGO")
 #BiocManager::install("AnnotationDbi")
@@ -44,9 +44,8 @@ library(Rgraphviz)
 library(multtest, verbose = FALSE)
 
 ####### ADJUSTED P-VALUE #http://rileylab.org/wp-content/uploads/2016/08/multtest.pdf
-#useful info: #multiple_dexp[,c(1,which(colnames(multiple_dexp) == current_leaf))]
 
-#GO enrichment on multiple leafs(samples)
+#GO enrichment on multiple leaves(samples)
 
 #list with samples names
 multiple_dexp<- read.delim(args[1],stringsAsFactors = F) 
@@ -74,7 +73,7 @@ for(leaf in sample_names[2:length(sample_names)]){
                 geneSelectionFun = function(x)(x == 1),
                 annot = annFUN.org, mapping = "org.At.tair.db")
   
-  #Run the Fishers exact Test (raw p-vals)
+  #Run the Fishers exact Test (raw p-vals)
   fisher_sample <- runTest(GOdata_sample, algorithm = "elim", statistic = "fisher")
   
   sample_result <- GenTable(GOdata_sample, raw.p.value = fisher_sample,