How to get GO terms for E. Coli genes in R?

Question

[mamillerpa]

I just did a differential gene expression analysis with cufflinks and imported it into R with cummerbund.

My reference genome and GFF annotation were from Ensembl, and the Ensembl IDs for genes were carried through instead of the mnemonic gene symbols. (see below)

Is there a way to get GO terms for each of my E. coli genes within R/bioconductor? Can I do it with gene identifiers like "gene:b3281", or do I need to translate those to some other identifiers like "aroE" ?

biolinux@biolinux-VirtualBox[biolinux] grep -C 1 aroE Escherichia_coli_str_k_12_substr_mg1655.GCA_000005845.2.22.gff3
Chromosome ensembl transcript 3429766 3430023 . - . ID=transcript:AAC76305;Parent=gene:b3280;biotype=protein_coding;external_name=yrdB-1;logic_name=ena
Chromosome ensembl gene 3430020 3430838 . - . ID=gene:b3281;biotype=protein_coding;description=dehydroshikimate reductase%2C NAD(P)-binding;external_name=aroE;logic_name=ena
Chromosome ensembl transcript 3430020 3430838 . - . ID=transcript:AAC76306;Parent=gene:b3281;biotype=protein_coding;external_name=aroE-1;logic_name=ena
Chromosome ensembl gene 3430843 3431415 . - . ID=gene:b3282;biotype=protein_coding;description=tRNA(ANN) t(6)A37 threonylcarbamoyladenosine modification protein%2C threonine-dependent ADP-forming ATPase;external_name=tsaC;logic_name=ena


biolinux@biolinux-VirtualBox[fromoffice] grep b3281 gene_exp.diff
gene:b3281 gene:b3281 - Chromosome:3429235-3430838 q1 q2 OK 24.697 33.6117 0.444627 0.672035 0.3566 0.504885 no

[mamillerpa]

I have just parsed the "b numbers" and gene symbols out of the Ensmebl GFF file. Then I can get symbol:GO term mappings from topGO

Does this look even remotely reasonable?

EcExpGFF <- read.delim("Escherichia_coli_str_k_12_substr_mg1655.GCA_000005845.2.22.gff3", header=FALSE, stringsAsFactors=FALSE)

#I'm only interested in genes. Column 9 contains semicolon delimited key:value pairs, including the "b number" and the gene symbol "external gene id"
EcGFFgenes <- EcExpGFF[EcExpGFF$V3=="gene", 9]

#split the semicolon delimited list
EcLookupFromGFF <- read.csv(text = as.character(EcGFFgenes), header = FALSE, sep=";", stringsAsFactors = FALSE)

#remove the Key: portion of the text.

EcGenes<-substr(EcLookupFromGFF$V1,4,max(length(EcLookupFromGFF$V1)))
EcSyms <-substr(EcLookupFromGFF$V4,15,max(length(EcLookupFromGFF$V4)))

#now get the mapping between GO terms and the gene symbols. Use topGO library.

GO_BP2sym_list <- annFUN.org(whichOnto = "BP", EcSyms, mapping = "org.EcK12.eg.db", ID = "symbol")

Answer 1

[oganm]

You can use biomart package for this purpose. But it will not be an enrichment analysis and won't convey much useful information.

mart = useMart("ensembl", dataset="mmusculus_gene_ensembl") #replace mus musculus with your favorite species
goTerms = getBM(attributes = c("ensembl_gene_id", "goslim_goa_accession",
                               "goslim_goa_description"),
                filters = "mgi_symbol",
                values = geneIDs,
                mart = mart) #geneIDs is a vector containing your ensemble gene IDs

[mamillerpa]

Thanks, oganm.

> mart = useMart('ensembl') ; 
> listDatasets(mart)

doesn't show anything for E. coli. I was under the impression that Ensembl Bacterial didn't have a mart.