TopGO: how to select specific genes from the bacgkround for the analysis

Question

[Oli]

Hello community,

I need some help to make TopGO do what i need. I have intersected the differentially expressed genes in my RNA-seq experiment with a list of genes I'm interesed to for other reasons. When i create the TopGO object i have some troubles with the selection function. For the whole list of differentially expressed genes i select only those with a p-value < 0.05. What if i want to perform the enrichment analysis on the genes from the second list using the DGE from my experiment as the background?

My code is this:

background <- geneuniverse$PValue
names(background) <- geneuniverse$ensembl_gene_id
selection <- function(allScore){ return(allScore < 0.05)}
GoData <- new("topGOdata",
              ontology = "BP",
              allGenes = background,
              geneSel = selection,
              annot = annFUN.org,
              mapping = "org.Hs.eg.db", 
              ID = "ensembl",
              nodeSize = 3)

If i read my genes of interest in a vector by doing:

interesting_genes <- scan('interesting_genes.txt', header=TRUE)

How can I get TRUE/FALSE for matches between interesting genes and background? I have already tried match, %in% and grepl, but geneSel can only be a function. Any help?

I dont' know if it's of any help, but heres how the data look:

head(background)
ENSG00000183242 ENSG00000248347 ENSG00000172482 ENSG00000120068 ENSG00000137558 
   1.81e-15        1.15e-08        5.59e-09        5.33e-07        4.86e-06      
head(interesting_genes)
[1] "ENSG00000114779" "ENSG00000143322" "ENSG00000123130" "ENSG00000103740" "ENSG00000164398"

Answer 1

[e.rempel]

Hi Oli,

you could specify allGenes as a factor (length equals number of significant genes in the background). This factor equals 1 if the gene is also in * interesting_genes* and 0 otherwise.

background_genes = names(background)[background <= 0.05]
int.genes <- factor(as.integer(background_genes %in% interesting_genes))
names(int.genes) = background_genes
GoData <- new("topGOdata",
          ontology = "BP",
          allGenes = int.genes,
          annot = annFUN.org,
          mapping = "org.Hs.eg.db", 
          ID = "ensembl",
          nodeSize = 3)

So you are just limiting universe of genes to background_genes.