bug fixes

common/cp_enrichment.R

@@ -153,10 +153,10 @@ cp_all <- function(geneIds){
 #enrResults <- glMapped %>% do(cp_all(.$entrez_gene_id))
 
 ## run the actual enrichment test for all clusters and all ontologies
-library(foreach); library(doMC); registerDoMC(cores=10)
-enrResults <- glMapped %>%
-#    ddply(.(cluster), function(curGroup) cp_all(curGroup$entrez_gene_id), .progress="text", .parallel=T) %>%
-    ddply(groups(glMapped) %>% ac, function(curGroup) cp_all(curGroup$entrez_gene_id), .progress="text", .parallel=T) %>%
+#library(foreach); library(doMC); registerDoMC(cores=3)
+enrResults <- ddply(glMapped, groups(glMapped) %>% ac, function(curGroup){
+             cp_all(curGroup$entrez_gene_id)
+         }, .progress="text", .parallel=T) %>%
     cache_it(paste0("enrdata_", digest(glMapped)))
 
 
@@ -164,8 +164,6 @@ enrResults <- glMapped %>%
 #enrResults %<>% rename(Category=Description)
 #enrResults %<>% rename(ontology=Category)
 
-## restablish the grouping
-#enrResults %<>% group_by_(.dots=groups(geneLists) %>% ac)
 
 
 write.delim(enrResults, file=paste0(resultsBaseName, "enrResults.txt"))
@@ -180,8 +178,12 @@ facetSpecs <- paste("~", groups(geneLists) %>% ac %>% paste(collapse=" + "))
 enrResults %>% ggplot(aes(ontology)) + facet_wrap(as.formula(facetSpecs)) + geom_bar() + rot_x_lab() + ggtitle("enriched term counts by cluster")
 
 
+#' Keep at max 100 terms for visualzation per group
 #if(table(enrResults$cluster))
-enrResults %<>% slice(1:100)
+## restablish the grouping and limit results per group to 100
+enrResults %<>% group_by_(.dots=groups(geneLists) %>% ac)
+enrResults %<>% sample_frac(1) %>% filter((row_number()<100))
+
 
 #+ include=FALSE, eval=FALSE
 ## plot the enrichment results
@@ -275,7 +277,7 @@ l_ply(term_barplot_files$file, function(pngFile){ cat(paste0("<img src='", pngFi
 
 ########################################################################################################################
 # ' ## Enriched KEGG Pathways
-#+ eval=nrow(sigEnrResults %>% filter(Category=="KEGG_PATHWAY")) >0
+#+ eval=nrow(enrResults %>% filter(ontology=="kegg")) >0
 
 #' To understand spatio-temporal changes in gene expression better we now overlay enriched kegg pathways with the -log10(q_value) of each contrast. The direction of the expression changes is encoded as color, whereby red indicates that sample_1 is overexpressed. Because we have multiple contrasts of interest, this defines a slice-color barcode for each gene. To relate the barcode to contrasts we define the following slice order:
 
@@ -435,4 +437,4 @@ pathwayPlots %>% l_ply(function(plotData){
 })
 
 ## respin it for cild inclusion
-# require(knitr); setwd("/Volumes/projects/bioinfo/scripts/ngs_tools/dev/common"); spin("david_enrichment.R", knit=F)
\ No newline at end of file
+# require(knitr); setwd("/Volumes/projects/bioinfo/scripts/ngs_tools/dev/common"); spin("cp_enrichment.R", knit=F)