From bddff5337923650da50956aed5b3e050083c54d4 Mon Sep 17 00:00:00 2001
From: Aurelien Brionne <aurelien.brionne@inrae.fr>
Date: Wed, 15 Jul 2020 14:42:48 +0200
Subject: [PATCH] merge_enrich_terms pvalue cutoff upgrade
---
DESCRIPTION | 2 +-
NEWS | 1 +
R/enrich_GO_terms.R | 38 +++++++++++++++--------
R/merge_enrich_terms.R | 50 ++++++++++++++++++-------------
README.md | 2 +-
vignettes/mouse_bionconductor.Rmd | 14 +++++----
6 files changed, 68 insertions(+), 39 deletions(-)
diff --git a/DESCRIPTION b/DESCRIPTION
index e2b177a..0748244 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,5 +1,5 @@
Package: ViSEAGO
-Version: 1.3.5
+Version: 1.3.6
Title: ViSEAGO: a Bioconductor package for clustering biological functions using Gene Ontology and semantic similarity
Author: Aurelien Brionne [aut, cre],
Amelie Juanchich [aut],
diff --git a/NEWS b/NEWS
index 8dd0ac6..4f5f07d 100644
--- a/NEWS
+++ b/NEWS
@@ -71,4 +71,5 @@ CHANGES IN VERSION 1.3
o Ensembl2GO() biomart update
o show_heatmap() upgrade
o upset print update
+ o merge_enrich_terms upgrade pvalue cutoff
diff --git a/R/enrich_GO_terms.R b/R/enrich_GO_terms.R
index cfb8b3d..8fa5d88 100644
--- a/R/enrich_GO_terms.R
+++ b/R/enrich_GO_terms.R
@@ -32,19 +32,33 @@ setMethod(
# Extract pvalues
Data<-Data[,grep("\\.pvalue",names(Data)),with=FALSE]
- # count significant pvalues by condition
- Data<-Data[,lapply(.SD,function(x){sum(x<0.01,na.rm = TRUE)}),.SDcols=seq_len(ncol(Data))]
+ # pvalues threshlod according condition
+ p<-vapply(slot(object,"topGO"),function(x){
- # melt the table
- Data<-melt.data.table(
- Data,
- measure.vars=names(Data),
- variable.name = "conditions",
- value.name = "significant GO terms number"
- )
+ # unlist
+ x=unlist(x)
+
+ # extract pvalue threshold
+ as.numeric(
+ sub(
+ "^.+<",
+ "",
+ x[grep("test_name",names(x))]
+ )
+ )
+ },0)
- # remove .pvalue in conditions column
- Data[,"conditions":=gsub("\\.pvalue","",Data$conditions)]
+ # count significant pvalues by condition
+ Data<-lapply(seq_len(ncol(Data)),function(x){
+
+ data.table(
+ conditions=sub("\\.pvalue","",names(Data)[x]),
+ `significant GO terms number`=sum(Data[,x,with=FALSE]<p[x],na.rm=TRUE)
+ )
+ })
+
+ # bind results
+ Data<-rbindlist(Data)
# get topGO information
topGO<-slot(object,"topGO")
@@ -86,7 +100,7 @@ setMethod(
"\n- input:\n ", paste(paste(names(object@input),
vapply(slot(object,"input"),function(x){paste(x,collapse=", ")},""),sep=": "),collapse="\n "),
"\n- topGO summary:\n ", topGO,
- "\n- enrich GOs data.table (p<0.01 in at least one list): ",nrow(slot(object,"data"))," GO terms of ",nrow(Data)," conditions.",
+ "\n- enrich GOs (in at least one list): ",nrow(slot(object,"data"))," GO terms of ",nrow(Data)," conditions.",
paste("\n ",Data$conditions,":",Data$`significant GO terms number`,"terms"),sep=""
)
}
diff --git a/R/merge_enrich_terms.R b/R/merge_enrich_terms.R
index d1f4c18..27e29f6 100644
--- a/R/merge_enrich_terms.R
+++ b/R/merge_enrich_terms.R
@@ -154,7 +154,7 @@ setMethod(
# extract quering objects names
x=Input[[x]]
- # check existance
+ # check existence
values<-ls(envir=envir)
# check if available
@@ -171,7 +171,7 @@ setMethod(
# objects type
obj.type=vapply(x,class,"")
- # extract ontoloy type
+ # extract ontology type
vapply(seq_along(x),function(y){
# extract ontology slot
@@ -265,8 +265,8 @@ setMethod(
# scored GOs
GO_scored=length(slot(x[[y]],"score")),
- # significant GOs
- GO_significant=table(slot(x[[y]],"score")<0.01)[2],
+ # significant GOs according cutOff
+ GO_significant=table(slot(x[[y]],"score")<as.numeric(sub("^.+[[:space:]]","",slot(x[[y]],"testName"))))[2],
# feasibles genes
feasible_genes=slot(x[[y]],"geneData")[1],
@@ -326,7 +326,7 @@ setMethod(
# tested algorithm
algorithms<-Data[pos]
- # extract significvant pvalues results
+ # extract significant pvalues results
unlist(
lapply(algorithms,function(y){
@@ -334,7 +334,11 @@ setMethod(
pvalues<-topGO::score(y)
# extract names of enrich terms
- as.vector(names(pvalues[pvalues<0.01]))
+ as.vector(
+ names(
+ pvalues[pvalues<as.numeric(sub("^.+[[:space:]]","",slot(y,"testName")))]
+ )
+ )
})
)
})
@@ -375,7 +379,7 @@ setMethod(
stop("No enrich GO terms available in at least one condition")
}
- # initialyse input
+ # initialize input
input=list()
# combine results
@@ -709,23 +713,23 @@ setMethod(
pvalues=data.table(do.call("cbind",pvalues))
# algoritms
- algorithms=vapply(algorithms,function(x){slot(x,"algorithm")},"")
+ algo=vapply(algorithms,function(x){slot(x,"algorithm")},"")
# if use of different algorithms
- if(length(algorithms)>1){
+ if(length(algo)>1){
# add names
- names(pvalues)[-1]<-paste(rep(algorithms,each=2),names(pvalues)[-1],sep=".")
- }
+ names(pvalues)[-1]<-paste(rep(algo,each=2),names(pvalues)[-1],sep=".")
+ }
- # return input params
- assign(
- "input",
- c(input,list(algorithms)),
- inherits=TRUE
- )
+ # return input params
+ assign(
+ "input",
+ c(input,list(algo)),
+ inherits=TRUE
+ )
- ## combine results#
+ ## combine results
# all results in list
Results<-list(
@@ -751,8 +755,14 @@ setMethod(
# remove NA in GO.Id column
Results<-Results[!is.na(Results$GO.ID)]
+ # extract pvalue threshold
+ p<-unique(vapply(algorithms,function(x){ as.numeric(sub("^.+[[:space:]]","",slot(x,"testName")))},0))
+
+ # stop if more than one pvalue threshold among comparison
+ if(length(p)>1){stop("Only one pvalue theshold is allowed by list element.")}
+
# Remove gene ID and symbol if GO term not significant
- Results[Results$pvalue>=0.01,`:=`(Significant_genes=NA,Significant_genes_symbol=NA)]
+ Results[Results$pvalue>=p,`:=`(Significant_genes=NA,Significant_genes_symbol=NA)]
if(!is.null(names(Input))){
@@ -817,7 +827,7 @@ setMethod(
allResults[,"GO.ID":=NULL]
)
- # rename the fisrt 3 columns
+ # rename the first 3 columns
names(allResults)[seq_len(3)]<-c("GO.ID","term","definition")
# significant results in at least one condition
diff --git a/README.md b/README.md
index f078327..5c35505 100644
--- a/README.md
+++ b/README.md
@@ -27,7 +27,7 @@ remotes::install_gitlab(
devtools::build("ViSEAGO")
# install package (from R console)
- install.packages("ViSEAGO_1.3.5.tar.gz", repos = NULL, type = "source")
+ install.packages("ViSEAGO_1.3.6.tar.gz", repos = NULL, type = "source")
```
## Citation
diff --git a/vignettes/mouse_bionconductor.Rmd b/vignettes/mouse_bionconductor.Rmd
index 8e035c3..c9ff44e 100644
--- a/vignettes/mouse_bionconductor.Rmd
+++ b/vignettes/mouse_bionconductor.Rmd
@@ -179,23 +179,26 @@ Now, we perform the GO enrichment tests for BP category with Fisher's exact test
<u>NB</u>: p-values of enriched GO terms are not adjusted and considered significant if below 0.01.
```{r Enrichment_data_tests}
-# perform TopGO tests
+# perform topGO tests
elim_BP_PregnantvsLactate<-topGO::runTest(
BP_PregnantvsLactate,
algorithm ="elim",
- statistic = "fisher"
+ statistic = "fisher",
+ cutOff=0.01
)
elim_BP_VirginvsLactate<-topGO::runTest(
BP_VirginvsLactate,
algorithm ="elim",
- statistic = "fisher"
+ statistic = "fisher",
+ cutOff=0.01
)
elim_BP_VirginvsPregnant<-topGO::runTest(
BP_VirginvsPregnant,
algorithm ="elim",
- statistic = "fisher"
+ statistic = "fisher",
+ cutOff=0.01
)
```
@@ -206,7 +209,8 @@ The printed table contains for each enriched GO terms, additional columns includ
```{r Enrichment_merge}
# merge topGO results
-BP_sResults<-ViSEAGO::merge_enrich_terms(
+ViSEAGO::
+BP_sResults<-merge_enrich_terms(
Input=list(
PregnantvsLactate=c(
"BP_PregnantvsLactate",
--
GitLab