Goseq Error: Only One Go Id Is In Enriched.Go List Multiple Times

Question

[pajailwala]

I am using GOseq on "mm9" as the genome and "geneSymbol" as the ID in my DE.genes and All.genes list. The input files are loaded fine, the named vector is fine and so is the calculated pwf and the nullp plot. The wallenius approximation also fetches the over-represented GO categories amongst the DE genes:

head(GO.wall)
category over_represented_pvalue under_represented_pvalue
 1878 GO:0016491            2.278057e-09                1.0000000
 220  GO:0003674            2.524681e-07                1.0000000
 264  GO:0003824            3.039433e-07                1.0000000
 159  GO:0001671            1.629699e-04                0.9999960
 1911 GO:0016616            2.039002e-04                0.9999346
 1909 GO:0016614            2.229153e-04                0.9999252

However, when I find the report GO categoris over enriched using the .05 FDR cutoff using the enriched.GO command, only the top over-represented GO ID (GO: 0016491) is reported repeatedly in the output:

head(enriched.GO)
[1] "GO:0016491" "GO:0016491" "GO:0016491" "GO:0016491" "GO:0016491" "GO:0016491"

I cant seem to debug what I am doing here here as I have followed the instructions from the vignette.

Here is the complete code for reference:

library(goseq) 
assayed.genes<-scan("Allgenes.txt", what=character()) 
de.genes<-scan("DE.txt", what=character()) 
gene.vector=as.integer(assayed.genes %in% de.genes) 
names(gene.vector)=assayed.genes 
print(table(gene.vector)) 
pwf=nullp(gene.vector,"mm9", "geneSymbol") 
print(head(pwf)) 
GO.MF=goseq(pwf, "mm9","geneSymbol", test.cats=c("GO:MF")) 
print(head(GO.MF)) 
enriched.GO=GO.MF$category[p.adjust(GO.MF$over_represented_pvalue,method="BH")] 
print(head(enriched.GO))

I will really appreciate any tips for those who have had more experience with GOseq.

Thanks!

[raffaelladalessandro]

I have the same problem using this function in multiple datasets. Have you find which was the error?

Thanks!

Answer 1

[sgujja]

Hi,

I have a similar issue with GOseq.

Can you please let us know the solution you've found for getting more enriched GO terms.

Thanks