[BiO BB] RefSeq Accession Numbers

[pmr at ebi.ac.uk]

Hi Hamid,

> I tried this :
> http://srs.ebi.ac.uk/srsbin/cgi-bin/wgetz?-e+[refseqp-acc:AP_000714]>uniprot+-ascii
> but this error has been shown:
> Error:No Entries found
> when I remove the last part (>uniprot)
> http://srs.ebi.ac.uk/srsbin/cgi-bin/wgetz?-e+[refseqp-acc:AP_000714]+-ascii
> it retrieves the result for AP_000714.
> but for your accession number:uniprot-acc:P13569 the reverse retrieves
> NP_000483 and when I try:
> http://srs.ebi.ac.uk/srsbin/cgi-bin/wgetz?-e+[refseqp-acc:AP_000483]>uniprot+-ascii
> I get the result ... It is strange!

Interesting ... it depends on how the links are built. One weakness in SRS
is that there is only one default path for links, but you can also make
them explicit.

I may have been wrong in assuming links will go through UniParc. We had an
example last year where we found links - not at EBI but at another SRS
server - between Genbank and RefSeq defaulted to going through the
taxonomy database (start with one human gene, find them all!!!)

I suspect in this case the route is not through UniParc. It could be
entrezgene or IPI, both of which fail for your entry. We can check and
maybe find a better preferred link that will include bacterial proteins.

You can force the route through UniParc with an extra redirect:

http://srs.ebi.ac.uk/srsbin/cgi-bin/wgetz?-e+[refseqp-acc:AP_000714]>uniparc>uniprot+-ascii

(If email mangles that, I added ">uniparc" before ">uniprot" in your
refseq to uniprot link).

You can also just link to UniParc and choose your own links.

Hope that helps,

Peter Rice