def program_args(self):
name = 'wespipeline' if not utils.GlobalParams(
).exp_name else utils.GlobalParams().exp_name
rg = '@RG\\tID:' + name + '\\tSM:' + name + '\\tPL:illumina'
args = [
'bwa',
'mem',
'-M',
'-R',
rg,
'-t',
FastqAlign().cpus,
self.input()['reference']['fa'].path,
self.input()['fastq']['fastq1'].path,
]
if 'fastq_2' in self.input()['fastq']:
args.append(self.input()['fastq']['fastq2'].path)
args.append('-o')
args.append(self.output().path)
return args
def output(self):
outputs = set()
outputs.add(
luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + ".fa.amb")))
outputs.add(
luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + ".fa.ann")))
outputs.add(
luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + ".fa.bwt")))
outputs.add(
luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + ".fa.pac")))
outputs.add(
luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + ".fa.sa")))
return outputs
def output(self):
return {
'bam' : luigi.LocalTarget( \
path.join(utils.GlobalParams().base_dir, utils.GlobalParams().exp_name+'_nodup.bam')),
'metrics' : luigi.LocalTarget( \
path.join(utils.GlobalParams().base_dir, utils.GlobalParams().exp_name+'_MD.matrix'))
}
def output(self):
get_name = lambda x: os.path.splitext(
os.path.basename(os.path.realpath(x)))[0]
output_filename = get_name(
self.input()['fastq']['fastq1'].path) + '.sam'
if utils.GlobalParams().exp_name:
output_filename = os.path.join(
utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + ".sam")
return luigi.LocalTarget(output_filename)
def requires(self):
if self.reference_local_file != '':
dependencies = {
'program': utils.GetProgram(),
'file': utils.LocalFile(file=self.reference_local_file),
}
else:
file = os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + '.2bit')
dependencies = {
'program': GetProgram(),
'file': utils.Wget(url=self.ref_url, output_file=file),
}
return dependencies
def requires(self):
if self.from2bit == True:
dependency = TwoBitToFa(
reference_local_file=self.reference_local_file,
ref_url=self.ref_url)
else:
if self.reference_local_file != '':
dependency = utils.LocalFile(self.reference_local_file)
else:
out_file = luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + '.fa'))
dependency = utils.Wget(url=self.ref_url, output_file=out_file)
return dependency
def output(self):
filename = lambda f: path.split(f)[-1].split('.')[0]
return luigi.LocalTarget(
path.join(
utils.GlobalParams().base_dir,
''.join([filename(self.vcf1), '_vs_',
filename(self.vcf2)]) + '.diff.sites_in_files'))
def requires(self):
if self.accession_number != '' and self.fastq1_local_file == '' and self.fastq1_local_file == '':
return SraToolkitFastq(accession_number=self.accession_number, paired_end=self.paired_end)
dependencies = dict()
if self.compressed == True:
if self.fastq1_local_file != '':
dependencies.update({'fastq1' : UncompressFastqgz(
fastq_local_file=self.fastq1_local_file,
fastq_url=self.fastq1_url,
output_file='.'.join(self.fastq1_local_file.split('.')[:-1]))})
else:
dependencies.update({'fastq1' : UncompressFastqgz(
fastq_local_file=self.fastq1_local_file,
fastq_url=self.fastq1_url,
output_file=path.join(utils.GlobalParams().base_dir, 'hg19_1.fastq'))})
else:
if self.fastq1_local_file != '':
dependencies.update({'fastq1' : utils.LocalFile(file=self.fastq1_local_file)})
else:
dependencies.update({'fastq1' : utils.Wget(url=self.fastq1_url, output_file=path.join(utils.GlobalParams().base_dir, 'hg19_1.fastq'))})
if self.paired_end == True:
if self.compressed == True:
if self.fastq2_local_file != '':
dependencies.update({'fastq2' : UncompressFastqgz(
fastq_local_file=self.fastq2_local_file,
fastq_url=self.fastq2_url,
output_file='.'.join(self.fastq2_local_file.split('.')[:-1]))})
else:
dependencies.update({'fastq2' : UncompressFastqgz(
fastq_local_file=self.fastq2_local_file,
fastq_url=self.fastq2_url,
output_file=path.join(utils.GlobalParams().base_dir, 'hg19_2.fastq'))})
else:
if self.fastq2_local_file != '':
dependencies.update({'fastq2' : utils.LocalFile(file=self.fastq2_local_file)})
else:
dependencies.update({'fastq2' : utils.Wget(url=self.fastq2_url, output_file=path.join(utils.GlobalParams().base_dir, 'hg19_2.fastq'))})
return dependencies
def program_args(self):
filename = lambda f: path.split(f)[-1].split('.')[0]
output_filename = path.join(utils.GlobalParams().base_dir, \
''.join([filename(self.vcf1), '_vs_', filename(self.vcf2)]))
return ['bash', '-c', \
f"vcftools --vcf {self.vcf2} --diff {self.vcf1} --diff-site " + \
"--chr chr1 --chr chr2 --chr chr3 --chr chr4 --chr chr5 --chr chr6 " + \
"--chr chr7 --chr chr8 --chr chr9 --chr chr10 --chr chr11 --chr chr12 " + \
"--chr chr13 --chr chr14 --chr chr15 --chr chr16 --chr chr17 --chr chr18 " + \
f"--chr chr20 --chr chr21 --chr chr22 --chr chrX --chr chrY --out {output_filename} 2> {output_filename}.log" ]
def output(self):
return luigi.LocalTarget(
os.path.join(utils.GlobalParams().base_dir,
utils.GlobalParams().exp_name + '.fa'))
def output(self):
program_file = os.path.join(utils.GlobalParams().base_dir,
'twoBitToFa')
return luigi.LocalTarget(program_file)
def requires(self):
program_file = os.path.join(utils.GlobalParams().base_dir,
'twoBitToFa')
program_url = 'ftp://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/twoBitToFa'
return utils.Wget(url=program_url, output_file=program_file)
