def rsem(inputs, outputs):
"""
:params inputs: a tuple of 1 or 2 fastq.gz files, e.g.
('/path/to/rsem_output/homo_sapiens/GSE50599/GSM1224499/SRR968078_1.fastq.gz',
'/path/to/rsem_output/homo_sapiens/GSE50599/GSM1224499/SRR968078_2.fastq.gz')
"""
inputs = [_ for _ in inputs if not _.endswith('.sra2fastq.COMPLETE')]
# this is equivalent to the sample.outdir or GSM dir
outdir = os.path.dirname(inputs[0])
# the names of parameters are the same as that in gen_qsub_script, but
# their values are more or less different, so better keep them separate
fastq_gz_input = gen_fastq_gz_input(inputs)
res = re.search(PATH_RE, outdir)
gse = res.group('GSE')
species = res.group('species')
gsm = res.group('GSM')
reference_name = config['LOCAL_REFERENCE_NAMES'][species]
sample_name = '{outdir}/{gsm}'.format(**locals())
n_jobs = options.j_rsem
flag_file = outputs[-1]
cmd = config['CMD_RSEM'].format(n_jobs=n_jobs,
fastq_gz_input=fastq_gz_input,
reference_name=reference_name,
sample_name=sample_name,
output_dir=outdir)
misc.execute_log_stdout_stderr(cmd,
flag_file=flag_file,
debug=options.debug)
def gen_qsub_script(inputs, outputs):
"""generate qsub script, usually named 0_submit.sh"""
inputs = [_ for _ in inputs if not _.endswith('.sra2fastq.COMPLETE')]
# SHOULD DO TRY EXCEPT IN CASE THE PREVIOUS STEP DIDN'T FINISH SUCCESSFULLY
outdir = os.path.dirname(inputs[0])
# only need the basename since the 0_submit.sh will be executed in the
# GSM dir
fastq_gz_input = gen_fastq_gz_input([os.path.basename(_) for _ in inputs])
res = re.search(PATH_RE, outdir)
gse = res.group('GSE')
species = res.group('species')
gsm = res.group('GSM')
reference_name = config['REMOTE_REFERENCE_NAMES'][species]
sample_name = '{gsm}'.format(gsm=gsm)
qsub_script = os.path.join(outdir, '0_submit.sh')
# TEMPLATES_DIR: the standard templates directory
# os.getcwd(): the current working directory
# use both for looking for the template
jinja2_env = Environment(
loader=FileSystemLoader([TEMPLATES_DIR, os.getcwd()]))
template = jinja2_env.get_template(options.qsub_template)
with open(qsub_script, 'wb') as opf:
content = template.render(**locals())
opf.write(content)
logger.info('templated {0}'.format(qsub_script))
def gen_qsub_script(inputs, outputs):
"""generate qsub script, usually named 0_submit.sh"""
inputs = [_ for _ in inputs if not _.endswith('.sra2fastq.COMPLETE')]
# SHOULD DO TRY EXCEPT IN CASE THE PREVIOUS STEP DIDN'T FINISH SUCCESSFULLY
outdir = os.path.dirname(inputs[0])
# only need the basename since the 0_submit.sh will be executed in the
# GSM dir
fastq_gz_input = gen_fastq_gz_input(
[os.path.basename(_) for _ in inputs])
res = re.search(PATH_RE, outdir)
gse = res.group('GSE')
species = res.group('species')
gsm = res.group('GSM')
reference_name = config['REMOTE_REFERENCE_NAMES'][species]
sample_name = '{gsm}'.format(gsm=gsm)
qsub_script = os.path.join(outdir, '0_submit.sh')
# TEMPLATES_DIR: the standard templates directory
# os.getcwd(): the current working directory
# use both for looking for the template
jinja2_env = Environment(loader=FileSystemLoader([TEMPLATES_DIR, os.getcwd()]))
template = jinja2_env.get_template(options.qsub_template)
with open(qsub_script, 'wb') as opf:
content = template.render(**locals())
opf.write(content)
logger.info('templated {0}'.format(qsub_script))
def rsem(inputs, outputs):
"""
:params inputs: a tuple of 1 or 2 fastq.gz files, e.g.
('/path/to/rsem_output/homo_sapiens/GSE50599/GSM1224499/SRR968078_1.fastq.gz',
'/path/to/rsem_output/homo_sapiens/GSE50599/GSM1224499/SRR968078_2.fastq.gz')
"""
inputs = [_ for _ in inputs if not _.endswith('.sra2fastq.COMPLETE')]
# this is equivalent to the sample.outdir or GSM dir
outdir = os.path.dirname(inputs[0])
# the names of parameters are the same as that in gen_qsub_script, but
# their values are more or less different, so better keep them separate
fastq_gz_input = gen_fastq_gz_input(inputs)
res = re.search(PATH_RE, outdir)
gse = res.group('GSE')
species = res.group('species')
gsm = res.group('GSM')
reference_name = config['LOCAL_REFERENCE_NAMES'][species]
sample_name = '{outdir}/{gsm}'.format(**locals())
n_jobs = options.j_rsem
flag_file = outputs[-1]
cmd = config['CMD_RSEM'].format(
n_jobs=n_jobs,
fastq_gz_input=fastq_gz_input,
reference_name=reference_name,
sample_name=sample_name,
output_dir=outdir)
misc.execute_log_stdout_stderr(cmd, flag_file=flag_file, debug=options.debug)
def test_gen_fastq_gz_input(self):
self.assertEqual(rsem.gen_fastq_gz_input(
[
'path/to/SRR000000_1.fastq.gz',
'path/to/SRR000000_2.fastq.gz'
]), '--paired-end <(/bin/zcat path/to/SRR000000_1.fastq.gz) <(/bin/zcat path/to/SRR000000_2.fastq.gz)')
self.assertEqual(rsem.gen_fastq_gz_input(
[
'path/to/SRR000000_1.fastq.gz',
'path/to/SRR111111_1.fastq.gz',
'path/to/SRR000000_2.fastq.gz',
'path/to/SRR111111_2.fastq.gz'
]), ('--paired-end '
'<(/bin/zcat path/to/SRR000000_1.fastq.gz path/to/SRR111111_1.fastq.gz) '
'<(/bin/zcat path/to/SRR000000_2.fastq.gz path/to/SRR111111_2.fastq.gz)'))
self.assertEqual(rsem.gen_fastq_gz_input(['path/to/SRR000000_1.fastq.gz']),
'<(/bin/zcat path/to/SRR000000_1.fastq.gz)')
self.assertEqual(rsem.gen_fastq_gz_input(['path/to/SRR000000_2.fastq.gz']),
'<(/bin/zcat path/to/SRR000000_2.fastq.gz)')
self.assertIsNone(rsem.gen_fastq_gz_input(['invalid_fastq_gz_name']))
