refSeq gene tables -- all in one file?

[memeri]

Does NCBI have a single, downloadable file containing all refSeq-annotated exons for all genes? This information is available on a browser, one gene at a time, for example:

http://www.ncbi.nlm.nih.gov/gene/3939?report=gene_table&format=text

I'd like to maintain a local database with this information.

[mastal]

See the ftp downloads site:

ftp://ftp.ncbi.nlm.nih.gov/refseq/release/

[memeri]

Thank you, mastal.
I was hoping for something a bit higher level than this, though. If I want to compile an up-to-date list of all annotated exons mapped to the current human reference genome assembly, I need to search through all 468 '...genomic.gbff' files in the 'vertebrate_mammalian' directory. In a somewhat random search of about 20 of these files, I didn't come across one with an entry from the current build (GRCh38.p2), and I don't see an index that may help me limit my search. I suspect that there may be a more direct route to this information; I'm sure NCBI does not search through all this every time they serve up a gene table.

[memeri]

Thanks, blancha. These are, of course, browser based methods -- one gene at a time, butnice to know!

[blancha]

Sorry, @memeri, I already deleted my last post, since I wasn't sure it answered your question. I'm putting the information back, since it probably does answer your question.

These are not one gene at a time methods.
You will get the entire annotation for the whole genome in one file.
You will have the location of each exon for all the genes.

The refSeq annotation of the exons can easily be downloaded from the UCSC Table Browser, using the RefSeq Track.
You can select the fields you want, if you don't want all the fields, as well as the format of the output file.

An alternative is to use Ensembl. You can just download a GTF file, or use biomaRt.

Here are the fields available in the refSeq track.

name
chrom
strand
txStart
txEnd
cdsStart
cdsEnd
exonCount
exonStarts
exonEnds
score
name2
cdsStartStat
cdsEndStat
exonFrames

You may have to work a bit to get exactly the format you want, but all the information is available from the UCSC Table Browser, Ensembl, and probable GenCode.

I think that answers your question.
If it doesn't, I'll let someone else try.
And, I won't delete the post this time.

[memeri]

So, I've found the answer, for those who may be interested. The file 'GCF_000001405.28_knownrefseq_alignments.gff3' (or the most recent version) in the directory 'ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/alignments/' maps every refSeq exon to the current human genome build. Plus you need the file 'gene2accession.gz' in 'ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/' to map the accession numbers to the geneIDs.

Thanks again.

[memeri]

Thanks for your second note, blancha. I didn't see it until after I posted my most recent note. I'll look at it more closely to see if it's better than what I came up with. It looks like it returns less irrelevant information in a single download, which is better.
Mark