[Bioperl-l] blastxml oddity

[Chris Fields]

NCBI's BLAST team is working on this; it occurs only if you use an GI/ 
accession instead of a full sequence, and only for XML.

chris

On Oct 4, 2007, at 2:30 PM, Chris Fields wrote:

> Just noticed an oddity from BLASTXML output from the NCBI server; I'm
> cc'ing this to NCBI so maybe they can explain.  BTW, the following
> doesn't occur via URLAPI.
>
> When running a standard BLAST query using the NCBI web page, if
> requesting XML output after the run I get the entire query seq masked
> out and no midline.  This occurs with all default settings except
> output type (set to XML).  Can anyone replicate this?
>
> Here's a sample:
>
> <Hit_hsps>
>              <Hsp>
>                <Hsp_num>1</Hsp_num>
>                <Hsp_bit-score>320.472</Hsp_bit-score>
>                <Hsp_score>820</Hsp_score>
>                <Hsp_evalue>2.17708e-86</Hsp_evalue>
>                <Hsp_query-from>1</Hsp_query-from>
>                <Hsp_query-to>181</Hsp_query-to>
>                <Hsp_hit-from>1</Hsp_hit-from>
>                <Hsp_hit-to>181</Hsp_hit-to>
>                <Hsp_query-frame>0</Hsp_query-frame>
>                <Hsp_hit-frame>0</Hsp_hit-frame>
>                <Hsp_identity>0</Hsp_identity>
>                <Hsp_positive>0</Hsp_positive>
>                <Hsp_gaps>0</Hsp_gaps>
>                <Hsp_align-len>181</Hsp_align-len>
>
> <Hsp_qseq>XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 
> XX
> XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX 
> XX
> XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX</Hsp_qseq>
>
> <Hsp_hseq>MNQKAVILDEQAIRRALTRIAHEMIERNKGMNNCILVGIKTRGIYLAKRLAERIEQIEGN 
> PV
> TVGEIDITLYRDDLSKKTSNDEPLVKGADIPVDITDQKVILVDDVLYTGRTVRAGMDALVDVGRPSSIQL 
> AV
> LVDRGHRELPIRADYIGKNIPTSKSEKVMVQLDEVDQNDLVAIYENE</Hsp_hseq>
>
> <Hsp_midline>
>
>                                                     </Hsp_midline>
>              </Hsp>
>            </Hit_hsps>
>
>
>
> <notblast>chris</notblast>
> _______________________________________________
> 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