[Bioperl-l] possible bug printing GenBank feature qualfiers

Chris Fields cjfields at uiuc.edu
Sat Apr 1 04:26:53 UTC 2006


Hilmar,

I figured it out.  The reason the test case wasn't parsing correctly  
was the spacing in the feature table.

This works:
=========================

LOCUS       MY_LOCUS                  10 aa            linear   UNK
DEFINITION  my description.
ACCESSION   12345
FEATURES             Location/Qualifiers
      misc_feature    1..10
                      /foo="0"
ORIGIN
          1 atggagaact
//
=========================

Removing or adding a space before 'misc_feature', so that the line  
does not start with exactly 5 spaces, causes the warning and  
misparses the feature.  The spacing after is not as important.   
Should it be that inflexible?

As for the fix, Scott's addition didn't change anything unless  
Heikki's last CVS fix to Bio::Annotation::SimpleValue, which changes  
the overloaded operator, is also reverted.  That makes sense when  
reading Heikki's commit message:

"fix to allow 0 values despite operator overload (Paul Mooney)"

Since the problem seems to be solved and both fixes (Scott's and  
Heikki's) are redundant and essentially get the same results, one of  
them should be rolled back.  I'll go ahead and roll the  
Bio::SeqIO::genbank commit back sometime in the next few days, unless  
you think there might be a better way to go about this?

On Mar 31, 2006, at 5:02 PM, Hilmar Lapp wrote:

> Note that GenBank requires a 'source' feature. The GenBank parser uses
> it to get the NBCI taxon ID (as that is the feature where it will be
> given as a db_xref tag). I thought the parser wouldn't mandate the
> feature but maybe at some point it assumes that it's there. Need to
> check, just a speculation.
>
> On 3/31/06, Brian Osborne <osborne1 at optonline.net> wrote:
>> Chris,
>>
>> Not OS-specific, I also see "no feature key!" on Mac OS.
>>
>> Brian O.
>>
>>
>> On 3/31/06 3:27 PM, "Chris Fields" <cjfields at uiuc.edu> wrote:
>>
>>> I wouldn't be a bit surprised if it is a WinXP-specific issue, so  
>>> I'll give
>>> it a try this weekend on Mac OS X using the latest CVS to see  
>>> what happens.
>>
>>
>>
>
>
> --
> ----------------------------------------------------------
> : Hilmar Lapp -:- San Diego, CA -:- hlapp at gmx dot net :
> ----------------------------------------------------------
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l

Christopher Fields
Postdoctoral Researcher
Lab of Dr. Robert Switzer
Dept of Biochemistry
University of Illinois Urbana-Champaign






More information about the Bioperl-l mailing list