def runCallerWorkflow(self, sample):
patientId = sample['PatientID']
callset = self._getRawCallset()
#pprint(callset)
#process_output_folder = getPipelineFolder('somaticseq-' + self.mode, patientId)
vjsd_output_folder = os.path.join(self.output_folder,
"0_fixed_variants")
merged_output_folder = os.path.join(self.output_folder,
"1_merged_variants")
filtered_output_folder = os.path.join(self.output_folder,
"2_filtered_variants")
table_output_folder = os.path.join(self.output_folder, "3_tables")
filetools.checkDir(vjsd_output_folder)
filetools.checkDir(merged_output_folder)
filetools.checkDir(filtered_output_folder)
filetools.checkDir(table_output_folder)
processed_callset = self._processVJSDFiles(
callset,
#self.output_folder,
vjsd_output_folder,
patientId)
merged_raw_variant_file = self._mergeVariantFiles(
processed_callset,
#self.output_folder,
merged_output_folder,
patientId)
filtered_raw_variant_file = self._filterVariantTargets(
merged_raw_variant_file,
#self.output_folder,
filtered_output_folder,
patientId)
self.trained_snp_table = self._generateCovariateTable(
sample,
callset,
filtered_raw_variant_file,
#self.output_folder
table_output_folder)
if self.mode == 'table':
pass #The trained_snp_table is the final result.
elif self.mode == 'trainer':
self.classifier = self.buildTrainer(self.trained_snp_table)
elif self.mode == 'prediction':
self.classifier = None
prediction_table = self.runPredictor(self.trained_snp_table)
prediction_vcf = self._convertToVcf(sample, prediction_table)
else:
message = "'{}' is not a supported mode for SomaticSeq! ('trainer', 'prediction', 'table')".format(
self.mode)
raise ValueError(message)
return self.trained_snp_table
def __init__(self, sample, truthset_options, callset_type, truthset_type,
**kwargs):
"""
Parameters
----------
sample: dict
truthset_options: Settings
callset_type: {'snp', 'indel'}
truthset_type = {'rna', 'intersection'}
"""
########################### Define Common Attributes ##################
self.debug = False
# Parameters to use when generating the truthsets.
if isinstance(sample, str): patientId = sample
else:
patientId = sample['PatientID']
self.truthset_type = truthset_type
self.indel_intersection = kwargs.get('indel_intersection', 2)
self.snp_intersection = kwargs.get('snp_intersection', 5)
self.min_tumor_vaf = 0.08
self.max_normal_vaf = 0.03
self.gatk_program = truthset_options['Programs']['GATK']
self.picard_program = truthset_options['Programs']['picard']
self.reference = truthset_options['Reference Files'][
'reference genome']
self.output_folder = truthset_options.getPipelineFolder('truthset')
self.temp_folder = os.path.join(self.output_folder, 'callsets',
patientId)
filetools.checkDir(self.temp_folder, True)
######################## Generate the Truthset ########################
self.filename = self.runWorkflow(patientId=patientId,
callset_type=callset_type,
truthset_type=truthset_type,
truthset_options=truthset_options,
**kwargs)
def getPipelineFolder(self, step, patientId=None, caller_name=None):
if step == 'callset':
subfolders = ["3_called_variants", patientId]
elif step == 'variants-somatic':
subfolders = ["3_called_variants", patientId, caller_name]
elif step == 'variants-copynumber':
subfolders = ["4_called_cnvs", patientId, caller_name]
elif step == 'temporary':
subfolders = ['5_temporary_files', patientId]
elif step == 'bam-files':
subfolders = []
elif step == 'reference':
return "/home/upmc/Documents/Reference/"
elif step == 'variants-rna':
subfolders = ['7_rna_variants', patientId]
elif step == 'truthset':
subfolders = ['truthset']
elif step == 'somaticseq':
subfolders = ['somaticseq']
elif step == 'somaticseq-callset':
subfolders = ['somaticseq', 'callsets', patientId]
elif step == 'somaticseq-training':
subfolders = ['somaticseq', 'training']
elif step == 'somaticseq-prediction':
subfolders = ['somaticseq', 'prediction']
elif step == 'somaticseq-table':
subfolders = ['somaticseq', 'tables']
else:
message = "'{}' is not a valid step in the pipeline!".format(step)
raise ValueError(message)
pipeline_folder = os.path.join(self.base_pipeline_folder, *subfolders)
filetools.checkDir(pipeline_folder, True)
return pipeline_folder
def _overwriteExistingFiles(self):
""" Deletes any existing files """
shutil.rmtree(self.output_folder)
filetools.checkDir(self.output_folder)
def __init__(self, sample, options):
##### Define commonly-used variables
self.caller_name = self.__class__.__name__
self.sample_log = options.getPipelineFile('sample log')
self.targets = sample['ExomeTargets']
self.reference = options['Reference Files']['reference genome']
self.dbSNP = options['Reference Files']['dbSNP']
self.cosmic = options['Reference Files']['COSMIC']
self.verbose_level = options['globals']['verbose']
self.program = options['Programs'].get(self.caller_name.lower())
self.gatk_program = options['Programs']['GATK']
self.max_cpu_threads = options['Parameters']['MAX_CORES']
self.max_memory_usage = options['Parameters']['JAVA_MAX_MEMORY_USAGE']
self.min_base_quality = options['Parameters']['MIN_NUCLEOTIDE_QUALITY']
self.min_mapping_quality = options['Parameters']['MIN_MAPPING_QUALITY']
self.min_somatic_quality = options['Parameters']['SOMATIC_QUALITY']
self.min_coverage = options['Parameters']['MIN_COVERAGE']
##### Define the paths and common partial filenames
self.output_folder = options['variants-somatic', sample['PatientID'],
self.caller_name]
#self.output_folder = options['Pipeline Options']['somatic pipeline folder']
self.base_prefix = "{normal}_vs_{tumor}.{prefix}".format(
tumor=sample['SampleID'],
normal=sample['NormalID'],
prefix=self.caller_name.lower())
self.abs_prefix = os.path.join(self.output_folder, self.base_prefix)
self.temp_folder = options['temporary', sample['PatientID']]
#self.temp_folder = options['Pipeline Options']['temporary folder']
self.temp_files = list()
self.full_output = []
self.setCustomEnvironment(sample, options)
filetools.checkDir(self.output_folder, True)
filetools.checkDir(self.temp_folder, True)
self.console_file = os.path.join(self.output_folder,
self.caller_name + ".console_log.txt")
self.readme_filename = os.path.join(self.output_folder,
self.caller_name + ".readme.txt")
if options['globals']['overwrite'] and os.path.exists(
self.output_folder):
shutil.rmtree(self.output_folder)
print("\tRunning ", self.caller_name)
if False:
print("\tprogram location: ", self.program)
print("\toutput folder: ", self.output_folder)
print("\ttemp folder: ", self.temp_folder)
print("\tcaller output prefix: ", self.base_prefix)
self.createReadmeFile(sample)
self.runCallerWorkflow(sample)
self.renameOutputFiles()
self.verifyOutputFiles(self.full_output)