def _remove_pcr_duplicates(self, config, prefix, bams, strategy):
rmdup_cls = {"collapsed" : FilterCollapsedBAMNode,
"normal" : MarkDuplicatesNode}
keep_duplicates = False
if isinstance(strategy, types.StringTypes) and (strategy.lower() == "mark"):
keep_duplicates = True
# Indexing is required if we wish to calulate per-region statistics,
index_required = (bool(prefix.get("RegionsOfInterest")) or
# or if we wish to run GATK, but only if we don't
# use a downstream rescaled BAM as input for GATK
(self.options["Features"]["RealignedBAM"] and not
self.options["RescaleQualities"]))
results = {}
for (key, files_and_nodes) in bams.items():
output_filename = self.folder + ".rmdup.%s.bam" % key
node = rmdup_cls[key](config = config,
input_bams = files_and_nodes.keys(),
output_bam = output_filename,
keep_dupes = keep_duplicates,
dependencies = files_and_nodes.values())
validated_node = index_and_validate_bam(config, prefix, node,
create_index=index_required)
results[key] = {output_filename : validated_node}
return results
def _mapdamage_rescale(self, config, destination, prefix, files_and_nodes):
output_filename = self.folder + ".rescaled.bam"
# Generates basic plots / table files
plot = self._mapdamage_plot(config=config,
destination=destination,
prefix=prefix,
files_and_nodes=files_and_nodes)
# Builds model of post-mortem DNA damage
model = MapDamageModelNode.customize(reference=prefix["Reference"],
directory=destination,
dependencies=plot)
apply_options(model.command, self.options["mapDamage"])
model = model.build_node()
# Rescales BAM quality scores using model built above
input_files = files_and_nodes.keys()
scale = MapDamageRescaleNode.customize(config=config,
reference=prefix["Reference"],
input_files=input_files,
output_file=output_filename,
directory=destination,
dependencies=model)
apply_options(scale.command, self.options["mapDamage"])
scale = scale.build_node()
# Grab indexing and validation nodes, required by ROIs and GATK
index_required = bool(prefix.get("RegionsOfInterest")) \
or self.options["Features"]["RealignedBAM"]
validate = index_and_validate_bam(config, prefix, scale,
create_index=index_required)
return {output_filename: validate}, (model,)
def _build_realigned_bam(self, config, prefix, bams):
output_filename = os.path.join(
self.folder, "%s.%s.realigned.bam" % (self.target, prefix["Name"]))
intervals_filename = os.path.join(self.folder, self.target,
prefix["Name"] + ".intervals")
validated_filename = os.path.join(
self.folder, self.target, prefix["Name"] + ".realigned.validated")
trainer = gatk.GATKIndelTrainerNode(config=config,
reference=prefix["Reference"],
infiles=bams.keys(),
outfile=intervals_filename,
threads=config.gatk_max_threads,
dependencies=self.datadup_check)
aligner = gatk.GATKIndelRealignerNode(config=config,
reference=prefix["Reference"],
infiles=bams.keys(),
intervals=intervals_filename,
outfile=output_filename,
dependencies=trainer)
validated_node = index_and_validate_bam(config=config,
prefix=prefix,
node=aligner,
log_file=validated_filename)
return {output_filename: validated_node}
def _mapdamage_rescale(self, config, destination, prefix, files_and_nodes):
model = self._mapdamage_model(config=config,
destination=destination,
prefix=prefix,
files_and_nodes=files_and_nodes)
# Rescales BAM quality scores using model built above
input_files = files_and_nodes.keys()
output_filename = self.folder + ".rescaled.bam"
scale = MapDamageRescaleNode.customize(config=config,
reference=prefix["Reference"],
input_files=input_files,
output_file=output_filename,
directory=destination,
dependencies=model)
apply_options(scale.command, self.options["mapDamage"])
scale = scale.build_node()
# Grab indexing and validation nodes, required by ROIs and GATK
index_required = bool(prefix.get("RegionsOfInterest")) \
or self.options["Features"]["RealignedBAM"]
validate = index_and_validate_bam(config=config,
prefix=prefix,
node=scale,
create_index=index_required)
return {output_filename: validate}, (model, )
def _remove_pcr_duplicates(self, config, prefix, bams, strategy):
rmdup_cls = {
"collapsed": FilterCollapsedBAMNode,
"normal": MarkDuplicatesNode
}
keep_duplicates = False
if isinstance(strategy, types.StringTypes) and (strategy.lower()
== "mark"):
keep_duplicates = True
# Indexing is required if we wish to calulate per-region statistics,
index_required = (
bool(prefix.get("RegionsOfInterest")) or
# or if we wish to run GATK, but only if we don't
# use a downstream rescaled BAM as input for GATK
(self.options["Features"]["RealignedBAM"]
and not self.options["Features"]["mapDamage"] == 'rescale'))
results = {}
for (key, files_and_nodes) in bams.items():
output_filename = self.folder + ".rmdup.%s.bam" % key
node = rmdup_cls[key](config=config,
input_bams=files_and_nodes.keys(),
output_bam=output_filename,
keep_dupes=keep_duplicates,
dependencies=files_and_nodes.values())
validated_node = index_and_validate_bam(
config=config,
prefix=prefix,
node=node,
create_index=index_required)
results[key] = {output_filename: validated_node}
return results
def _build_realigned_bam(self, config, prefix, bams):
output_filename = os.path.join(
self.folder, "%s.%s.realigned.bam" % (self.target, prefix["Name"]))
intervals_filename = os.path.join(
self.folder, self.target, prefix["Name"] + ".intervals")
validated_filename = os.path.join(
self.folder, self.target, prefix["Name"] + ".realigned.validated")
trainer = gatk.GATKIndelTrainerNode(config=config,
reference=prefix["Reference"],
infiles=bams.keys(),
outfile=intervals_filename,
threads=config.gatk_max_threads,
dependencies=self.datadup_check)
aligner = gatk.GATKIndelRealignerNode(config=config,
reference=prefix["Reference"],
infiles=bams.keys(),
intervals=intervals_filename,
outfile=output_filename,
dependencies=trainer)
validated_node = index_and_validate_bam(config=config,
prefix=prefix,
node=aligner,
log_file=validated_filename)
return {output_filename: validated_node}
def _build_raw_bam(self, config, prefix, files_and_bams):
output_filename = os.path.join(self.folder, "%s.%s.bam" % (self.target, prefix["Name"]))
validated_filename = os.path.join(self.folder, self.target, prefix["Name"] + ".validated")
node = MergeSamFilesNode(config = config,
input_bams = files_and_bams.keys(),
output_bam = output_filename,
dependencies = self.datadup_check)
validated_node = index_and_validate_bam(config, prefix, node, validated_filename)
return {output_filename : validated_node}
def _build_raw_bam(self, config, prefix, files_and_bams):
output_filename = os.path.join(
self.folder, "%s.%s.bam" % (self.target, prefix["Name"]))
validated_filename = os.path.join(self.folder, self.target,
prefix["Name"] + ".validated")
node = MergeSamFilesNode(config=config,
input_bams=files_and_bams.keys(),
output_bam=output_filename,
dependencies=self.datadup_check)
validated_node = index_and_validate_bam(config=config,
prefix=prefix,
node=node,
log_file=validated_filename)
return {output_filename: validated_node}
def _finalize_nodes(self, config, prefix, parameters, node):
self._set_rg_tags(node.commands["convert"])
min_quality = self.options["Aligners"]["BWA"]["MinQuality"]
node.commands["convert"].set_option('-q', min_quality)
if self.options["Aligners"]["BWA"]["FilterUnmappedReads"]:
node.commands["convert"].set_option('-F', "0x4")
index_required = self._is_indexing_required(prefix)
validated_node = index_and_validate_bam(config=config,
prefix=prefix,
node=node.build_node(),
create_index=index_required)
return validated_node