[Bioperl-l] Seq error/RemoteBlast question

Wiepert, Mathieu Wiepert.Mathieu@mayo.edu
Thu, 22 Aug 2002 10:49:45 -0500


Hi,

I noticed that RemoteBlast output seems to be missing the Query information in the output.  I noticed that this behavior is the same at NCBI if I don't put in a Fasta header line in the sequence field, along with the sequence.

So, I hacked RemoteBlast, like this

$header{'QUERY'} = ">".$seq->primary_id()." ".$seq->desc()."\n".$seq->seq();

which then got me the Query name filled out in the ResultI compliant object.  Just what I wanted.

I then was curious what would happen if the id and desc were blank or null, would that kill the whole thing (i.e. should I be checking that $seq->primary_id() and $seq->desc() values exist, or can I just concatenate like I did safely and not worry about the string becoming null).  I made my Fasta file look like

>
LIRPIGEKLP

Which did a strange thing.  The $seq object created from the above had this...
<stuff>
'display_id' => 'LIRPIGEKLP'
'primary_id' => 'LIRPIGEKLP'
'seq' => ''
</stuff>

So no sequence, but the actual sequence was stuffed into the primary_id.  Is this correct behavior?  Blast sure didn't like it, got no results, and error:

Use of uninitialized value in concatenation (.) or string at /home/mxw02/bioperl

That error makes sense (should be checking that the seq->desc field is filled in for sure, not sure if it safe to assume that the seq->seq is, would have thought so?

-Mat