Topgo Error

Question

[johnstantongeddes]

I'm using topGO and getting a couple odd (maybe related?) errors.

The default test

resultWeight01 <- runTest(BP_GOdata, statistic = 'fisher')

runs fine.

But, when I try:

resultKS <- runTest(BP_GOdata, algorithm = 'classic', statistic = 'ks')

I get the message

Error in seq_len(N)[-x.a] : only 0's may be mixed with negative subscripts

and when I run

resultKS.elim <- runTest(BP_GOdata, algorithm = 'elim', statistic = "ks")

I get the error message:

Error in ks.test(x.a, seq_len(N)[-x.a], alternative = "greater") : not enough 'y' data

The default results may be adequate, but curious why the other methods don't work.

My topGOdata:

BP_GOdata <- new("topGOdata", description = "BP gene set analysis", ontology = "BP", allGenes = geneList, geneSel = selectFDR, nodeSize = 10, annot = annFUN.gene2GO, gene2GO = geneID2GO)

Answer 1

[johnstantongeddes]

Solved my own question...

I had NA values in my 'geneList'. These didn't cause problems with the default (gene counts) but did cause problems with the 'ks' tests based on gene scores.

Answer 2

[tao2013]

The same probelm. Thanks for your sharing