Fixes issue #22 with getGOsize reporting wrong GO term size

imbforge · Dec 15, 2023 · cd08892 · cd08892
1 parent 35e2117
commit cd08892
Show file tree

Hide file tree

Showing 64 changed files with 96 additions and 73 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,7 +1,7 @@
 Package: rrvgo
 Type: Package
 Title: Reduce + Visualize GO
-Version: 1.13.4
+Version: 1.15.1
 Authors@R: c(person("Sergi", "Sayols", email = "[email protected]", role = c("aut", "cre")), person("Sara", "Elmeligy", role = "ctb"))
 Description: Reduce and visualize lists of Gene Ontology terms by identifying redudance based on semantic similarity.
 URL: https://www.bioconductor.org/packages/rrvgo, https://ssayols.github.io/rrvgo/index.html

diff --git a/R/rrvgo.R b/R/rrvgo.R
@@ -91,6 +91,8 @@ calculateSimMatrix <- function(x,
 #'   orgdb object itself)
 #' @param keytype keytype passed to AnnotationDbi::keys to retrieve GO terms 
 #'   associated to gene ids in your orgdb
+#' @param children when retrieving GO term size, include genes in children terms.
+#'   (based on relationships in the GO DAG hierarchy). Defaults to TRUE
 #' @return a data.frame identifying the different clusters of terms, the parent
 #' term representing the cluster, and some metrics of importance describing how
 #' unique and dispensable a term is.
@@ -102,7 +104,7 @@ calculateSimMatrix <- function(x,
 #' @importFrom stats cutree hclust
 #' @export
 reduceSimMatrix <- function(simMatrix, scores=c("uniqueness", "size"),
-                            threshold=0.7, orgdb, keytype="ENTREZID") {
+                            threshold=0.7, orgdb, keytype="ENTREZID", children=TRUE) {
 
   # check function arguments
   if(is(scores, "character")) {
@@ -115,7 +117,7 @@ reduceSimMatrix <- function(simMatrix, scores=c("uniqueness", "size"),
   cluster <- cutree(hclust(as.dist(1 - simMatrix)), h=threshold)
 
   # get category size and term uniqueness, and use it as scores if they were not provided
-  sizes    <- getGoSize(rownames(simMatrix), orgdb, keytype)
+  sizes    <- getGoSize(rownames(simMatrix), orgdb, keytype, children)
   termUniq <- getTermUniq(simMatrix, cluster)
   if(is(scores, "character")) {
     scores <- switch(scores,
@@ -194,27 +196,22 @@ getTermDisp <- function(simMatrix, cluster, clusterRep) {
 #'   package itself)
 #' @param keytype keytype passed to AnnotationDbi::keys to retrieve GO terms 
 #'   associated to gene ids in your orgdb
+#' @param children include genes in children terms (based on relationships in
+#'  the GO DAG hierarchy)
 #' @importFrom AnnotationDbi select keys
 #' @importFrom stats setNames
 #' @importFrom methods is
 #' @return number of genes associated with each term
-getGoSize <- function(terms, orgdb, keytype) {
+getGoSize <- function(terms, orgdb, keytype, children) {
   if(all(is(orgdb) != "OrgDb")) {
     orgdb <- loadOrgdb(orgdb)
   }
 
-  # get all GO terms with genes associated
-  go <- suppressMessages(
-          AnnotationDbi::select(orgdb,
-                                keytype=keytype,
-                                columns=c("GO", "ONTOLOGY"),
-                                keys=AnnotationDbi::keys(orgdb, keytype=keytype)))
-  go <- go[!is.na(go$GO), ]
-  go <- go[go$GO %in% terms, ]
+  # retrieve genes per term and count unique genes within each term
+  go <- AnnotationDbi::select(orgdb, keytype=if(children) "GOALL" else "GO", keys=terms, columns=keytype)
 
   # count
-  counts   <- table(go$GO)
-  go <- go[go$GO %in% terms, ]
+  counts   <- tapply(go$ENTREZID, go$GO, function(x) length(unique(x)))
   empty    <- terms[!(terms %in% names(counts))]
   nocounts <- setNames(rep(0, length(empty)), empty)
 

diff --git a/docs/404.html b/docs/404.html
diff --git a/docs/LICENSE-text.html b/docs/LICENSE-text.html
diff --git a/docs/articles/index.html b/docs/articles/index.html
diff --git a/docs/articles/rrvgo.html b/docs/articles/rrvgo.html
diff --git a/docs/articles/rrvgo_files/figure-html/unnamed-chunk-4-1.png b/docs/articles/rrvgo_files/figure-html/unnamed-chunk-4-1.png
diff --git a/docs/articles/rrvgo_files/figure-html/unnamed-chunk-6-1.png b/docs/articles/rrvgo_files/figure-html/unnamed-chunk-6-1.png
diff --git a/docs/authors.html b/docs/authors.html
diff --git a/docs/index.html b/docs/index.html
diff --git a/docs/pkgdown.yml b/docs/pkgdown.yml
@@ -3,5 +3,5 @@ pkgdown: 2.0.7
 pkgdown_sha: ~
 articles:
   rrvgo: rrvgo.html
-last_built: 2023-10-17T12:00Z
+last_built: 2023-12-15T12:18Z
 
diff --git a/docs/reference/Rplot001.png b/docs/reference/Rplot001.png
diff --git a/docs/reference/calculateSimMatrix.html b/docs/reference/calculateSimMatrix.html
diff --git a/docs/reference/getGoSize.html b/docs/reference/getGoSize.html
diff --git a/docs/reference/getGoTerm.html b/docs/reference/getGoTerm.html
diff --git a/docs/reference/getTermDisp.html b/docs/reference/getTermDisp.html
diff --git a/docs/reference/getTermUniq.html b/docs/reference/getTermUniq.html
diff --git a/docs/reference/gg_color_hue.html b/docs/reference/gg_color_hue.html