Parsing Blast Output Biopython Error

Question

[Ankur]

Hi, I have the following code

 def runBLAST(self):
        print "Running BLAST .........."
        cmd=subprocess.Popen("blastp -db nr -query repeat.txt -out out.faa -evalue 0.001 -gapopen 11 -gapextend 1 -matrix BLOSUM62 -remote -outfmt 5",shell=True)
        cmd.communicate()[0]
        f1=open("out.faa")
        blast_records = NCBIXML.parse(f1)
        save_file = open("my_fasta_seq.fasta", 'w')
        for blast_record in blast_records[:10]:
            for alignment in blast_record.alignments:
                for hsp in alignment.hsps:
                    save_file.write('>%s\n' % (alignment.hseq,))
        save_file.close()
        f1.close()
        f2=open("my_fasta_seq.fasta")
        for record in SeqIO.parse(f2,"fasta"):
            f=open("tempBLAST1.txt","w")
            f.write(">"+"\n"+strrecord.name)+"\n"+str(record.seq)+"\n")
            f.close()

I get the error on TypeError: for blastrecord in blastrecords[:10]: saying 'generator' object is not subscriptable. I am looking to get top 10 blast hits (sequences)

Answer 1

[Michael Kuhn]

This is not a specific BioPython problem, but a general Python question, answered e.g. on StackOverflow. It might be that BioPython only parses the next result on demand, in this case you might be better off with:

for i, blast_record in enumerate(blast_records):
    if i == 10: break
    ...

[Neilfws]

It's also a follow-up to the previous question and perhaps should have continued there instead. It's fine to edit your questions and discuss answers in the comments, rather than starting a new question for every variation of the same problem.

[Peter]

As Michael says, blast_records is a generator/iterator. You can loop over it or iterate explicitly by calling next(), but you cannot access records by index. This is a general design pattern for coping with very large files composed of multiple smaller records, also used in the the Biopython SeqIO parse function etc.