[Biopython-dev] Possible problem with NCBIStandalone.blastall

[Peter]

On Fri, Nov 7, 2008 at 6:56 AM, Sebastian Bassi wrote:
> When I run a command line blast with these parameters:
>
> /root/blast-2.2.18/bin/blastall -p blastn -d /var/www/blast/db/UniVec
> -q -5 -G 3 -E 3 -F "m D" -e 700 -Y 1.75e12 -i tmpsq
>
> I find a match (with evalue of 18).
> But when I do it from biopyhon I can't find any match:
>
> rh, eh = NCBIStandalone.blastall(blast_exe, "blastn", db,
>                                     fin, nuc_mismatch='-5',
>                                     gap_open = '3',
>                                     gap_extend = '3',
>                                     search_length = '1.75e12',
>                                     expectation='20')

You are not using exactly the same arguments, so its not surprising
you get different results:

-q -5 =>nuc_mismatch = -5 (or as a string)
-G 3 => gap_open = 3 (or as a string)
-E 3 => gap_extend = 3 (or as a string)
-F "m D" => filter="m D" (MISSING!)
-e 700 => expectation=700 (or as a string)
-Y = 1.75e12 => search_length = '1.75e12' (or as a float)

Your expectation cut off is more generous in the Biopython version
(700) than the commanline line version (20), but that wouldn't explain
the difference.  Its probably due to omitting the filter option (-F).
If that doesn't resolve the difference then there is something very
strange going on...

Peter