Antwort: network USA

[David.Bauer at SCHERING.DE]

Hi,

I use for this a workaround which uses method app calling scripts which use two
urls at ncbi.
In emboss.default I have two entries for nucleotide and protein, which call an
external script.
############
DB ncbin [ type: N method: app format: genbank
  app: "/bips/bin/emboss/ncbi_fetchn %s"
  comment: "NCBI GenBank Nucleotide" ]

DB ncbip [ type: P method: app format: genbank
  app: "/bips/bin/emboss/ncbi_fetchp %s"
  comment: "NCBI GenBank Protein" ]
##################

The script is unfortunately not very portable as it uses a modified perl LWP
module to work with our firewall.
Basic idea is to use the

"http://www.ncbi.nlm.nih.gov/entrez/utils/pmqty.fcgi?db=nucleotide&term
=$id&dopt=genbank"
resp. "http://www.ncbi.nlm.nih.gov/entrez/utils/pmqty.fcgi?db=protein&term
=$id&dopt=genpept" for protein

to get the gid where $id is locus or acc.
If there are different gid for one acc, then all of them are returned.
What I have observed is that the gid of the most recent version is returned
first (but I'm not sure if this is always true).
So I just grab the first gid which comes and then use the same url you already
mentioned:

"http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=text&db=nucleotide&uid
=$gid&dopt=GenBank"
("http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=text&db=protein&uid
=$gid&dopt=genpept")

to return the whole entry.

So the user can just use
ncbin:<some id or acc.>
with  seqret or entret to get the sequence or genbank entry.


Hope this helps,
Ciao, David.