R : Get Gene Symbol From Refseq Id - Biomart Problem

Question

[Nicolas Rosewick]

Hi,

I've several refseq id that I want to convert to gene symbol. So in R with biomaRt:

library(biomaRt)
mart<- useDataset("hsapiens_gene_ensembl", useMart("ensembl"))
refseq <- c("NM_001271003")
getBM(filters="refseq_mrna", attributes="external_gene_id", values=refseq, mart=mart)

gives me logical(0)

But the gene is well defined in pubmed (TFPI2): http://www.ncbi.nlm.nih.gov/nuccore/NM_001271003

The gene is also in ensembl: http://www.ensembl.org/Homo_sapiens/Gene/Matches?db=core;g=ENSG00000105825;r=7:93514709-93520303 but in external references, the refseq id: NM_001271003 is not reported..

So is there an another way to get the gene symbol from refseq id using R ?

Thanks

Answer 1

[Emily]

The Ref Seq ID that you refer to is indeed very similar to the Ensembl transcript ENST00000222543, however ENST00000222543 actually aligns to a different RefSeq mRNA, NM_006528. If you look at these alignments of NM_001271003 and NM_006528, you can see that NM_006528 perfectly corresponds to the exons of ENST00000222543 whereas NM_001271003 differs by 33bp in one exon and 18bp in another exon. Ensembl chose the RefSeq which was the best fit to our model.