problem with wildcard in emboss applications.

[Ted Chiang]

Alan:

Thank you for you helpful suggestions.  I tried out your test by using one
of the test databases included in the package
(blah/emboss2.1.0/share/EMBOSS/test) and discovered
that the wildcard command works.

	seqret "test:*" 

Here's what's in the emboss.default:

-------------------------------------------------------------
set emboss_acdroot /tools/emboss2.1.0/share/EMBOSS/acd
set emboss_database_dir /data
set EMBOSS_DATA /tools/emboss2.1.0/share/EMBOSS/data
set emboss_logfile /tools/emboss2.1.0/logs/emboss.log


DB test [
   method: emblcd
   format: EMBL
   dir: /tools/emboss2.1.0/share/EMBOSS/test/embl
   file: "*.dat"
   comment: "test"
   type: N
   release: 126.0
   indexdir: /tools/emboss2.1.0/share/EMBOSS/test/embl
]


DB testuna127 [
   method: emblcd
   format: GB
   dir: /data/genbank/una.127
   file: "*.seq"
   comment: "Genbank Unannotated Sequences"
   type: N
   release: 126.0
   indexdir: /data/emboss/una.127
]
------------------------------------------------------------


However the bottom DB is where our problem is.  The command:

	%seqret testuna127:*
	Reads and writes (returns) sequences
	Error: Unable to read sequence 'testuna127:*'


fails even though with our previous version of EMBOSS-1.13.3, we always
got it to work. Interestingly, in retrieving a single database entry,
there is no problem, such as:


	%seqret testuna127:AA005656
	Reads and writes (returns) sequences
	Output sequence [af005656.fasta]:


Here is the debug output:

	%seqret -debug testuna127:* 
	EMBOSS An error in ajmess.c at line 1463:
	Cannot open debug file seqret.dbg


I am pretty convinced that the actual seqret program is working fine since
it works with the test database that came with the EMBOSS-2.1.0 package,
but I have run out of ideas why it can't run properly on our databases.
Do you have any suggestions?  Can it have something to do with the PLPLOT
or the configuration problems we had in installing it on our IRIX
machines?

thanks.

-Ted


=====================================
Ted Chiang
Bioinformatics Supercomputing Centre
Hospital for Sick Children, Toronto
ext. 7028
tchiang at bioinfo.sickkids.on.ca



On Fri, 4 Jan 2002 ableasby at hgmp.mrc.ac.uk wrote:

> I can neither reproduce this on an O2 nor your colleagues link
> strangeness in plplot.
> 
> Try setting up the emboss_tempdata definition in emboss.default
> (or your .embossrc) to point to blah/EMBOSS-2.1.0/test
> and then try   seqret "tembl:*"
> If that works examine your db definitions. If there are still
> problems you should send a few sample entries from the database
> you're trying to access, your database definition for it, how
> you indexed it and the results of a 'seqret -debug' to
> emboss-bug
> 
> If the above doesn't work then, after rechecking the 'set' line
> for emboss_tempdata, then again send all the info you can including a
> seqret -debug output to emboss-bug
> 
> Cheers
> Alan
> 
>