goseq package in R “missing value where TRUE/FALSE needed” error

Question

I am attempting to run a GO Analysis in R (I have never done this analysis, so I am trying different packages), and I am struggling to find the problem with my code in the goseq package.

I start with this code which produces a list of the differentially expressed gene names:

 de.genes <- rownames(res)[ which(res$padj < fdr.threshold & !is.na(res$padj)) ]

Then I try to run this code (based on page 7 of the vignette ( https://bioconductor.org/packages/devel/bioc/vignettes/goseq/inst/doc/goseq.pdf )

 pwf <- nullp(de.genes, "hg38","geneSymbol")

but I get the following error:

 Can't find hg38/geneSymbol length data in genLenDataBase...
 Found the annotation package, TxDb.Hsapiens.UCSC.hg38.knownGene
 Trying to get the gene lengths from it.
 Error in if (matched_frac == 0) { : missing value where TRUE/FALSE needed
 In addition: Warning message:
 In grep(txdbPattern, installedPackages):argument 'pattern' has length > 1 and only the first element will be used

I found this forum: https://support.bioconductor.org/p/38580/ that says I need an "indicator variable" but I do not know what this is.

Any help with this error would be greatly appreciated, or if you know of any other GO packages that are easy to learn. Thanks!

Answer 1

You can check the supported databases, hg38 is not one of them:

library(org.Hs.eg.db)
library(goseq)

supported[grep("hg38|hg19",supported$Genome),]
   Genome         Id  Id Description Lengths in geneLeneDataBase
4    hg19  knownGene  Entrez Gene ID                        TRUE
36   hg19    ensGene Ensembl gene ID                        TRUE
81   hg19 geneSymbol     Gene Symbol                        TRUE
98   hg38                                                  FALSE
   GO Annotation Available
4                     TRUE
36                    TRUE
81                    TRUE
98                    TRUE

You can get a rough idea of what it looks like by using hg19, you will have some missing or unmatched by should be ok. You need to have a binary vector and it should be named, for example:

set.seed(111)
allgenes = keys(org.Hs.eg.db,keytype="SYMBOL")
de.genes = rbinom(100,1,0.3)
names(de.genes) = sample(allgenes,100)

It looks like this:

  GALNT5        TPRKB         CD48       OR52R1 LOC105372708 LOC112163649 
       0            1            0            0            0            0 

LOC105369203 LOC110121115 LOC105377654 LOC105371502 LOC101929964 HPC14 0 0 0 0 0 0 IGHD4-17 LOC101927993 HINT1 BCC3 RPL18P3 LOC108281192 0 0 0 0 0 1 RNU6-793P JUN 0 0

This will be ok:

res = nullp(de.genes,"hg19","geneSymbol")