def print_run_bidirectional_blast(reference, other_genome, dbtype, outdir):
"""Write torque submission files for running bidirectional blast on a server and print execution command.
Args:
reference (str): Path to "reference" genome, aka your "base strain"
other_genome (str): Path to other genome which will be BLASTed to the reference
dbtype (str): "nucl" or "prot" - what format your genome files are in
outdir (str): Path to folder where Torque scripts should be placed
"""
# TODO: add force_rerun option
if dbtype == 'nucl':
command = 'blastn'
elif dbtype == 'prot':
command = 'blastp'
else:
raise ValueError('dbtype must be "nucl" or "prot"')
r_folder, r_name, r_ext = utils.split_folder_and_path(reference)
g_folder, g_name, g_ext = utils.split_folder_and_path(other_genome)
# Reference vs genome
r_vs_g_name = r_name + '_vs_' + g_name
r_vs_g = r_vs_g_name + '_blast.out'
if op.exists(op.join(
outdir, r_vs_g)) and os.stat(op.join(outdir, r_vs_g)).st_size != 0:
log.debug('{} vs {} BLAST already run'.format(r_name, g_name))
else:
cmd = '{} -query {} -db {} -outfmt 6 -out {}'.format(
command, reference, g_name, r_vs_g)
utils.write_torque_script(command=cmd,
err=r_vs_g_name,
out=r_vs_g_name,
name=r_vs_g_name,
outfile=op.join(outdir, r_vs_g_name) + '.sh',
walltime='00:15:00',
queue='regular')
# Genome vs reference
g_vs_r_name = g_name + '_vs_' + r_name
g_vs_r = g_vs_r_name + '_blast.out'
if op.exists(op.join(
outdir, g_vs_r)) and os.stat(op.join(outdir, g_vs_r)).st_size != 0:
log.debug('{} vs {} BLAST already run'.format(g_name, r_name))
else:
cmd = '{} -query {} -db {} -outfmt 6 -out {}'.format(
command, other_genome, r_name, g_vs_r)
utils.write_torque_script(command=cmd,
err=g_vs_r_name,
out=g_vs_r_name,
name=g_vs_r_name,
outfile=op.join(outdir, g_vs_r_name) + '.sh',
walltime='00:15:00',
queue='regular')
def run_makeblastdb(infile, dbtype, outdir=''):
"""Make the BLAST database for a genome file.
Args:
infile (str): path to genome FASTA file
dbtype (str): "nucl" or "prot" - what format your genome files are in
outdir (str): path to directory to output database files (default is original folder)
Returns:
Paths to BLAST databases.
"""
# TODO: add force_rerun option
# TODO: rewrite using utils function command
# Output location
og_dir, name, ext = utils.split_folder_and_path(infile)
if not outdir:
outdir = og_dir
outfile_basename = op.join(outdir, name)
# Check if BLAST DB was already made
if dbtype == 'nucl':
outext = ['.nhr', '.nin', '.nsq']
elif dbtype == 'prot':
outext = ['.phr', '.pin', '.psq']
else:
raise ValueError('dbtype must be "nucl" or "prot"')
outfile_all = [outfile_basename + x for x in outext]
db_made = True
for f in outfile_all:
if not op.exists(f):
db_made = False
# Run makeblastdb if DB does not exist
if db_made:
log.debug(
'BLAST database already exists at {}'.format(outfile_basename))
return outfile_all
else:
retval = subprocess.call(
'makeblastdb -in {} -dbtype {} -out {}'.format(
infile, dbtype, outfile_basename),
shell=True)
if retval == 0:
log.debug('Made BLAST database at {}'.format(outfile_basename))
return outfile_all
else:
log.error('Error running makeblastdb, exit code {}'.format(retval))
def run_bidirectional_blast(reference, other_genome, dbtype, outdir=''):
"""BLAST a genome against another, and vice versa.
This function requires BLAST to be installed, do so by running:
sudo apt install ncbi-blast+
Args:
reference (str): path to "reference" genome, aka your "base strain"
other_genome (str): path to other genome which will be BLASTed to the reference
dbtype (str): "nucl" or "prot" - what format your genome files are in
outdir (str): path to folder where BLAST outputs should be placed
Returns:
Paths to BLAST output files.
(reference_vs_othergenome.out, othergenome_vs_reference.out)
"""
# TODO: add force_rerun option
if dbtype == 'nucl':
command = 'blastn'
elif dbtype == 'prot':
command = 'blastp'
else:
raise ValueError('dbtype must be "nucl" or "prot"')
r_folder, r_name, r_ext = utils.split_folder_and_path(reference)
g_folder, g_name, g_ext = utils.split_folder_and_path(other_genome)
# make sure BLAST DBs have been made
run_makeblastdb(infile=reference, dbtype=dbtype, outdir=r_folder)
run_makeblastdb(infile=other_genome, dbtype=dbtype, outdir=g_folder)
# Reference vs genome
r_vs_g = r_name + '_vs_' + g_name + '_blast.out'
r_vs_g = op.join(outdir, r_vs_g)
if op.exists(r_vs_g) and os.stat(r_vs_g).st_size != 0:
log.debug('{} vs {} BLAST already run'.format(r_name, g_name))
else:
cmd = '{} -query {} -db {} -outfmt 6 -out {}'.format(
command, reference, op.join(g_folder, g_name), r_vs_g)
log.debug('Running: {}'.format(cmd))
retval = subprocess.call(cmd, shell=True)
if retval == 0:
log.debug('BLASTed {} vs {}'.format(g_name, r_name))
else:
log.error('Error running {}, exit code {}'.format(command, retval))
# Genome vs reference
g_vs_r = g_name + '_vs_' + r_name + '_blast.out'
g_vs_r = op.join(outdir, g_vs_r)
if op.exists(g_vs_r) and os.stat(g_vs_r).st_size != 0:
log.debug('{} vs {} BLAST already run'.format(g_name, r_name))
else:
cmd = '{} -query {} -db {} -outfmt 6 -out {}'.format(
command, other_genome, op.join(r_folder, r_name), g_vs_r)
log.debug('Running: {}'.format(cmd))
retval = subprocess.call(cmd, shell=True)
if retval == 0:
log.debug('BLASTed {} vs {}'.format(g_name, r_name))
else:
log.error('Error running {}, exit code {}'.format(command, retval))
return r_vs_g, g_vs_r