I’m currently running a gene set enrichment analysis using topGO. I’m interested in the case with a predefined list of interesting genes and with custom rabbit annotations (coming from biomart).
When I prepare my data with (gene_2_GO being the custom rabbit annotations):
GOdata <- new('topGOdata', ontology='CC', allGenes = geneList, annot = annFUN.gene2GO, gene2GO = gene_2_GO)
I get obsolete GO terms in further analysis (a classic fisher for example). For example, one obsolete GO term is the “GO:0044420”, it appears in the environment GOCCTerm { "GO:0044420" %in% ls(GOCCTerm) }.
The obsolete GO terms don’t appear in the custom annotation gene_2_GO, so my questions are:
- Why terms that are not in the custom annotation appear after creation of the topGOdata-class ?
- How can I prevent obsolete GO terms from appearing in my results ?
- Will obsolete GO terms eventually go away in updates ? of a data-bases ? of the package itself ?
Thanks a lot for your help