[EMBOSS] db records retrieve

[Peter Rice]

On 08/04/2011 11:27, Alessandro Bruselles wrote:
> Hi all,
> I'm not able to retrieve sequences starting from official gene symbols (e.g.
> BCL2L10),
> what db do I have to use?

You need a database you can search by gene name. If the database has 
access through a website you can use the URL access method ... providing 
a URL that returns with the sequence in a format that EMBOSS can read.

For UniProt:

DB uniprotgene [
   type: "protein"
   format: "swiss"
   methodquery: "url"
   url: 
"http://srs.ebi.ac.uk/srs7bin/cgi-bin/wgetz?-e+[uniprot-gen:%s]+-ascii"
]

(the URL should be all in one line)

We will have other ways to do this in EMBOSS 6.4.0 when we enable gen 
and other database search fields.

> This is the output of whichdb:
>
> *$ whichdb BCL2L10 -debug
> Search all sequence databases for an entry and retrieve it
> Output file [outfile.whichdb]: stdout
> Warning: Cannot open division file '<null>' for database 'tsw'
> Warning: seqCdQry failed
> Warning: Cannot open division file '<null>' for database 'emblnew'
> Warning: seqCdQry failed
> Warning: Cannot open division file '<null>' for database 'trembl'
> Warning: seqCdQry failed
> Warning: Cannot open division file '<null>' for database 'nbrf'
>
>     EMBOSS An error in ajseqdb.c at line 5350:
> seqCdQryOpen failed*

You need to fix or comment out the tsw, emblnew, trembl and nbrf 
databases to avoid those messages. whichdb tries all sequence databases 
and is a good way to catch database that are not working (servers down, 
data or index files not in the expected directory)

Hope this helps

Peter