[BioPython] Running local blast with aa

[Jeffrey Chang]

Hi Sebastian,

Let me see if I got this correct.  You have a sequence that you're
using to BLAST against a database.  If you run it by hand using local
BLAST, you get some hits back.  However, if you run it from Biopython,
you're getting no hits back.

This would be very unusual because biopython does not do any
interpretation on the inputs or outputs to BLAST.  It just calls it
and returns whatever the program prints.  Are you sure that you're
calling it with the same parameters?  For example, is one of them
filtered for low complexity sequences?

To see how the program is getting called in biopython, insert a line
into NCBIStandalone.py:
    for attr in keywds.keys():
        params.extend([att2param[attr], str(keywds[attr])])
--> print repr(params)
    r, w, e = popen2.popen3(' '.join([blastcmd] + params))
    w.close()

Jeff


On Fri, Oct 25, 2002 at 12:56:08PM -0300, Sebastian Bassi wrote:
> Hello,
> 
> I have a problem I think is a Python problem since when I trie to do it 
> manually with the blast executable it works.
> I´m running local BLAST (the standalone version under WinNT).
> I did the test on the blast readme file and went OK. (the test was to 
> blastn a ecoli sequence)
> Then I tried to blastX an ecoli nucleitide sequence against the ecoli.aa 
> database (I did formated the ecolia.aa with formatdb of course). The 
> nucleotide sequence I used to test was:
>  >test de prot
> TATGAGCATACTTTGATGGCTTTGGAGGCTGGTTGTCATGTTATGTGTGAGAAGCCTCCTGCTATGACTCCTGAGCAGGC
> TCGTGAGATGTGTGATACTGCTCGTAAGTTGGGTAAGGTTTTGGCTTATGATTTTCATCATCGTTTTGCTTTGGATACTC
> AGCAGTTGCGTGAGCAGGTTACTAATGGTGTTTTGGGTGAGATTTATGTTACTACTGCTCGTGCT
> This is a backtranslation from a part of an actual ecoli protein 
> (AAC74397). (was backtranslated with python using table bacterial table, 
>   11)
> When I run this blast on the NCBI site, I get the original protein as a 
> first hit (AAC74397). When I do it locally from the command line BLAST 
> it also works fine.
> But the problem is that when I do it locally using Python (I mean, the 
> blastX against the ecoli.aa database) I get 0 hits. I´m using all 
> standard/default values.
> Here´s the output I get:
> 
> BLASTX 2.2.4 [Aug-26-2002]
> 
> 
> Reference: Altschul, Stephen F., Thomas L. Madden, Alejandro A. Schaffer,
> Jinghui Zhang, Zheng Zhang, Webb Miller, and David J. Lipman (1997),
> "Gapped BLAST and PSI-BLAST: a new generation of protein database search
> programs",  Nucleic Acids Res. 25:3389-3402.
> 
> Query= test de prot
>           (225 letters)
> 
> Database: //server/bioinfo/blast/data/ecoli.aa
>             4289 sequences; 1,358,990 total letters
> 
> 
> 
>   ***** No hits found ******
> 
>    Database: //server/bioinfo/blast/data/ecoli.aa
>      Posted date:  Oct 25, 2002  8:30 AM
>    Number of letters in database: 1,358,990
>    Number of sequences in database:  4289
> 
> If you want to see my input file:
> 
>  >test de prot
> TATGAGCATACTTTGATGGCTTTGGAGGCTGGTTGTCATGTTATGTGTGAGAAGCCTCCTGCTATGACTC
> CTGAGCAGGCTCGTGAGATGTGTGATACTGCTCGTAAGTTGGGTAAGGTTTTGGCTTATGATTTTCATCA
> TCGTTTTGCTTTGGATACTCAGCAGTTGCGTGAGCAGGTTACTAATGGTGTTTTGGGTGAGATTTATGTT
> ACTACTGCTCGTGCT
> 
> And here is my little python program (BTW, it works fine for blastn, so 
> I asume the program logic is OK).
> 
> 
> from Bio.Blast import NCBIStandalone
> import os
> import string
> import re
> 
> pathdb="//server/bioinfo/blast/data/ecoli.aa"
> blastexe="//server/bioinfo/blast/blastall.exe"
> pathin="D:\\projects\\bioinfo-adv\\set-cd-small\\"
> filesentr="D:\\projects\\bioinfo-adv\\set-cd-small"
> pathout="blastbatch\\outblast"
> 
> print "OK 1"
> cont = 0
> 
> # mando todos los files de set-cd-complete a una lista
> 
> listaentrada=os.listdir(filesentr)
> for x in listaentrada:
>     cont = cont + 1
>     blast_out, error_info = NCBIStandalone.blastall(blastexe, 'blastx', 
> pathdb, pathin+x)
>     salida=open(pathout+x+".txt","w")
>     salida.writelines(blast_out)
>     salida.close()
>     print "Blast numero "+`cont`+" "+x
> print "FIN"
> 
> print "OK 2 re ok"
> 
> 
> _______________________________________________
> BioPython mailing list  -  BioPython@biopython.org
> http://biopython.org/mailman/listinfo/biopython