make blastdb using python

Question

[luoterolb]

Hi, I want to run some blast locally. I have some queries and a lot of fasta files that would be my databases. But I need to convert from fasta to "database" format. I know how to do it with commandline (makeblastdb), but they are several fasta files, so I want to write a script and do it with (bio)python. I have been looking how to do it, but I haven't find out how. Can any one help me, please? Thanks!

[Ram]

Using BioPython to makeblastdb a bunch of files is neither a smart learning process nor an efficient way to actually get your task done. Why use anything but the shell?

[GenoMax]

While you could use biopython, why not do that using a simple shell script?

[luoterolb]

Thanks for your answers!

[Ram]

Do not add an answer unless you're answering the question. This should be a comment reply. I'm moving it to a comment on the top level post now.

[carloshgcosta]

Well, I know I'm preety late for the topc conversation but. If you're up to UNIX, or to any other system that you can install BLAST local.

#!/usr/bin/env python
import sys
import os

#If you're using nt data in your working dir
blastn_db = "makeblastdb -in <sequence.fasta> -dbtype nucl -input_type fasta -out <database>"
os.system(blastn_db)

Instead of using bioblast or any other module, you'll simply call the os to run in the background of the shell. If found this a little bit more efficient than using bioblast/biopython, however in terms of speed I can't say if it's faster or not. Hope it helps

Answer 1

[st.ph.n]

just use shell as @genomax2 suggested;

Put your fasta files you want to make into dbs into a dir:

for file in *.fasta; do /path/to/blast_install/bin/makeblastdb -in $file -dbtype prot/nucl; done

if you need to use python:

#!/usr/bin/env python
import glob, subprocess
for file in glob.glob("*.fasta"):
    subprocess.call("/path/to/blast_install/bin/makeblastdb -in " + file + " -db type prot/nucl", shell=True)

Answer 2

[Joe]

I agree with this being a much better idea through a standard shell script but if you're adamant about doing it through python (perhaps incorporating it as part of a much longer and more complicated pipeline), you need to do something like this:

Use the subprocess module:

import subprocess 

blastdb_cmd = 'makeblastdb -in {0} -dbtype nucl -title temp_blastdb'.format(fasta)
DB_process = subprocess.Popen(blastdb_cmd,
                              shell=True,
                              stdin=subprocess.PIPE,
                              stdout=subprocess.PIPE,
                              stderr=subprocess.PIPE)
DB_process.wait()

This is taken from my script here if you want to 'see it in action'.