[BioPython] Running Blast locally

[Peter]

Muthuraman, Manickam wrote:
> Dear Christof
> 
> Your command also works  separately but my question was how to intergrate blast in biopython script.
> 
> in biopython tutorial and cookbook they have the follwoing code where i need to provide the path to database ,file to blast and blast_exe.
> 
> I am not clear how to set the path for seq_file,db and exe.
> 
> import os
> my_blast_db=os.path.join(os.getcwd(),'at-est','a-cds-10-7.fasta')
> my_blast_file=os.path.join(os.getcwd(),'at-est','test_blast','sorghum_est-test.fasta')
> my_blast_exe=os.path.join(os.getcwd(),'blast','/home/manickam/blast/blastall')

Try typing this at the python prompt:

import os
help(os.path.join)

Are you familiar with relative paths etc?  You might find something like 
this easier to understand:

my_blast_db   = '/home/manickam/db/at-est/a-cds-10-7.fasta')
my_blast_file = '/home/manickam/sorghum_est-test.fasta')
my_blast_exe  = '/home/manickam/blast/blastall'

Or, based on you previous email you were using:

 > here is the command :
 > ./blastall -d db/swissprot -i /home/manickam/Documents/m_cold.fasta
 > -p blastp

Maybe something like this:

my_blast_db   = '/home/manickam/blast/db/swissprot')
my_blast_file = '/home/manickam/Documents/m_cold.fasta')
my_blast_exe  = '/home/manickam/blast/blastall'

It all depends on where you installed the blast program, where you put 
the blast databases, and where you are going to have your inputfile.

> here is the whole script
01> import os
02> my_blast_db=os.path.join(os.getcwd(),'at-est','a-cds-10-7.fasta')
03> 
my_blast_file=os.path.join(os.getcwd(),'at-est','test_blast','sorghum_est-test.fasta')
04> 
my_blast_exe=os.path.join(os.getcwd(),'blast','/home/manickam/blast/blastall')
05> from Bio.Blast import NCBIStandalone
06> 
blast_out,error_info=NCBIStandalone.blastall(my_blast_exe,'blastp',my_blast_db,my_blast_file)

At this point, some example scripts will save the output to a file, and 
then reload it and carry on.  This is very helpful if you have problems 
because you can open the file by hand and look at it.

07> b_parser=NCBIStandalone.BlastParser()
08> b_iterator=NCBIStandalone.Iterator(blast_out,b_parser)
09> b_record=b_iterator.next()
10> while 1:
11>     b_record=b_iterator.next()
12>     if b_record is None:
13>         break
14>     for alignment in b_record.alignments:
15>         print "inside 2 loop"
16>         for hsp in alignment.hsps:
17>             print "inside 1 loop"
18>             print 'seq:',alignment.title
> 
> it runs but b_record is None so it comes out of the while loop at first time itself. so it mean i am not getting out put of the blast.

Notice that at line 9, you set b_record to the first set of results 
(i.e. from the first sequence in your FASTA file).

Then, inside the look, at line 11 set b_record to the SECOND set of 
results and try and look at it.

I suggest you comment out line 9, and it should work better.

Finally, this code is using the "plain text" blast output, which can 
sometimes cause BioPython trouble.  I would recommend the XML parser but 
as you might know from the mailing list, it looks like they have changed 
the file format for multiple results in XML output...

Peter