def join(args, outs, chunk_defs, chunk_outs):
if do_not_make_cloupe(args):
outs.output_for_cloupe = None
return
reference = ReferenceManager(args.reference_path)
contig_info_fn = martian.make_path("contig_info.json")
with open(contig_info_fn, 'w') as outfile:
contig_info = get_contig_info(args.reference_path)
json.dump(contig_info, outfile)
gem_group_index_json = get_gem_group_index_json(args, outs)
call = [
"crconverter",
args.sample_id,
args.pipestance_type,
"--matrix",
args.feature_barcode_matrix,
"--analysis",
args.analysis,
"--output",
outs.output_for_cloupe,
"--description",
'"' + args.sample_desc + '"',
"--peaks",
args.peaks,
"--fragmentsindex",
args.fragments_index,
"--geneannotations",
reference.genes,
"--contiginfo",
contig_info_fn,
]
if args.metrics_json is not None:
call.extend(["--metrics", args.metrics_json])
if args.aggregation_csv is not None:
call.extend(["--aggregation", args.aggregation_csv])
if gem_group_index_json is not None:
call.extend(["--gemgroups", gem_group_index_json])
transcript_gene_types = get_annotation_gene_types(args)
if transcript_gene_types is not None:
call.extend(["--geneannotationtypes", ",".join(transcript_gene_types)])
# the sample desc may be unicode, so send the whole
# set of args str utf-8 to check_output
unicode_call = [arg.encode('utf-8') for arg in call]
# but keep the arg 'call' here because log_info inherently
# attempts to encode the message... (TODO: should log_info
# figure out the encoding of the input string)
martian.log_info("Running crconverter: %s" % " ".join(call))
try:
results = tk_subproc.check_output(unicode_call)
martian.log_info("crconverter output: %s" % results)
except subprocess.CalledProcessError as e:
outs.output_for_cloupe = None
martian.throw("Could not generate .cloupe file: \n%s" % e.output)
def main(args, outs):
if do_not_make_cloupe(args):
outs.output_for_cloupe = None
return
gem_group_index_json = get_gem_group_index_json(args, outs)
call = [
"crconverter", args.sample_id, args.pipestance_type, "--matrix",
args.filtered_gene_bc_matrices_h5, "--analysis",
get_analysis_h5_path(args), "--output", outs.output_for_cloupe,
"--description", args.sample_desc
]
if args.metrics_json:
call.extend(["--metrics", args.metrics_json])
if args.aggregation_csv:
call.extend(["--aggregation", args.aggregation_csv])
if gem_group_index_json:
call.extend(["--gemgroups", gem_group_index_json])
# the sample desc may be unicode, so send the whole
# set of args str utf-8 to check_output
unicode_call = [arg.encode('utf-8') for arg in call]
# but keep the arg 'call' here because log_info inherently
# attempts to encode the message... (TODO: should log_info
# figure out the encoding of the input string)
martian.log_info("Running crconverter: %s" % " ".join(call))
try:
results = tk_subproc.check_output(unicode_call)
martian.log_info("crconverter output: %s" % results)
except subprocess.CalledProcessError, e:
outs.output_for_cloupe = None
martian.throw("Could not generate .cloupe file: \n%s" % e.output)
def record_package_versions():
for package in cr_constants.PACKAGE_VERSION_CMDS:
name = package['name']
cmd = package['cmd']
version = tk_subproc.check_output(cmd, shell=True)
print '%s: %s' % (name, version)
def join(args, outs, chunk_defs, chunk_outs):
contig_info = get_contig_info(args)
with open(outs.contig_info_json, 'w') as outfile:
json.dump(contig_info, outfile)
call = [
"dlconverter", args.sample_id, "--output", outs.output_for_dloupe,
"--description", args.sample_desc, "--node-profile-h5",
args.normalized_node_profiles, "--contig-info-json",
outs.contig_info_json, "--merged-bed", args.node_cnv_calls,
"--tree-data", args.tree_data, "--tracks", args.tracks,
"--per-cell-summary", args.per_cell_summary_metrics
]
gene_annotation_path = tk_ref.get_loupe_genes(args.reference_path)
if os.path.exists(gene_annotation_path):
call.extend(["--gene-annotations", gene_annotation_path])
# the sample desc may be unicode, so send the whole
# set of args str utf-8 to check_output
unicode_call = [arg.encode('utf-8') for arg in call]
martian.log_info("Running dlconverter: %s" % " ".join(call))
try:
results = tk_subproc.check_output(unicode_call)
martian.log_info("dlconverter output: %s" % results)
except subprocess.CalledProcessError, e:
outs.output_for_dloupe = None
martian.throw("Could not generate .dloupe file: \n%s" % e.output)
def main(args, outs):
if args.pipestance_type != "count" and args.pipestance_type != "aggr":
martian.exit("The type argument must be one of: count, aggr")
if args.pipestance_type == "count":
pname = "SC_RNA_COUNTER_CS"
if args.pipestance_type == "aggr":
pname = "SC_RNA_AGGREGATOR_CS"
pipestance_exists = os.path.exists(args.pipestance_path)
if not pipestance_exists:
martian.exit("Invalid pipestance path: %s" % args.pipestance_path)
# check to see if an analysis file exists. If it doesn't, then
# this is likely a barnyard sample, and we cannot generate a
# .loupe file (CELLRANGER-773);
analysis_h5_path = os.path.join(args.pipestance_path, "outs/analysis/analysis.h5")
# 1.2.0 location only
internal_count_h5_path = os.path.join(
args.pipestance_path,
"SC_RNA_COUNTER_CS/SC_RNA_COUNTER/SC_RNA_ANALYZER/SUMMARIZE_ANALYSIS/fork0/files/analysis/analysis.h5"
)
internal_aggr_h5_path = os.path.join(
args.pipestance_path,
"SC_RNA_AGGREGATOR_CS/SC_RNA_AGGREGATOR/SC_RNA_ANALYZER/SUMMARIZE_ANALYSIS/fork0/files/analysis/analysis.h5"
)
if not os.path.exists(analysis_h5_path) \
and not os.path.exists(internal_count_h5_path) \
and not os.path.exists(internal_aggr_h5_path):
martian.exit("Could not find single-species analysis HDF5 file. " +
"Loupe Cell Browser files are not generated for multi-species experiments.")
# has to be 1.2 or higher
cellranger_pd_before_1_2_path = os.path.join(args.pipestance_path, "CELLRANGER_PD")
cellranger_cs_before_1_2_path = os.path.join(args.pipestance_path, "CELLRANGER_CS")
if os.path.exists(cellranger_pd_before_1_2_path) or os.path.exists(cellranger_cs_before_1_2_path):
martian.exit("mkloupe is only supported for Cell Ranger 1.2 and later.")
call = ["crconverter",
args.sample_id,
pname,
"--pipestance", args.pipestance_path,
"--output", outs.output_for_cloupe]
martian.log_info("Running crconverter: %s" % " ".join(call))
try:
results = tk_subproc.check_output(call)
martian.log_info("crconverter output: %s" % results)
except subprocess.CalledProcessError, e:
outs.output_for_cloupe = None
martian.throw("Could not generate .cloupe file: \n%s" % e.output)
def get_version():
# NOTE: this makes assumptions about the directory structure
script_dir = os.path.join(os.path.dirname(os.path.abspath(__file__)), '..',
'..', '..', 'bin')
version_fn = os.path.join(script_dir, '..', '.version')
if os.path.exists(version_fn):
with open(version_fn, 'r') as f:
output = f.read()
else:
output = tk_subproc.check_output(
['git', 'describe', '--tags', '--always', '--dirty'],
cwd=script_dir)
return output.strip()
def main(args, outs):
"""
Run the vlconverter executable with inputs that should be available in the outs
folder at the end of the pipeline run. This will generate "output_for_vloupe.vloupe"
in the stage folder.
Memory usage not expected to be excessive with this (thus no custom split/join
as of yet); it will need to load a few full files (bam.bai, fasta.fai) into memory.
"""
if args.concat_ref_bam is None or not os.path.isfile(args.concat_ref_bam) or \
args.consensus_bam is None or not os.path.isfile(args.consensus_bam) or \
args.contig_bam_bai is None or not os.path.isfile(args.contig_bam_bai):
martian.log_info(
'One or more bam files missing - cannot make vloupe file')
return
call = [
"vlconverter", args.sample_id, args.pipestance_type, "--output",
outs.output_for_vloupe, "--reference-bam", args.concat_ref_bam,
"--reference-bam-index", args.concat_ref_bam_bai, "--reference-fasta",
args.concat_ref_fasta, "--reference-fasta-index",
args.concat_ref_fasta_fai, "--reference-annotations",
args.concat_ref_annotations_json, "--clonotypes", args.clonotypes_csv,
"--consensus-bam", args.consensus_bam, "--consensus-bam-index",
args.consensus_bam_bai, "--consensus-annotations",
args.consensus_annotations_json, "--consensus-fasta",
args.consensus_fasta, "--consensus-fasta-index",
args.consensus_fasta_fai, "--contig-bam-relative-path",
args.contig_bam_relative_path, "--contig-bam-index",
args.contig_bam_bai, "--contig-annotations",
args.contig_annotations_json, "--contig-bed",
args.contig_annotations_bed, "--contig-fasta", args.contig_fasta,
"--contig-fasta-index", args.contig_fasta_fai, "--description",
args.sample_desc
]
# the sample desc may be unicode, so send the whole
# set of args str utf-8 to check_output
unicode_call = [arg.encode('utf-8') for arg in call]
# but keep the arg 'call' here because log_info inherently
# attempts to encode the message... (TODO: should log_info
# figure out the encoding of the input string)
martian.log_info("Running vlconverter: %s" % " ".join(call))
try:
results = tk_subproc.check_output(unicode_call)
martian.log_info("vlconverter output: %s" % results)
except subprocess.CalledProcessError, e:
outs.output_for_vloupe = None
martian.throw("Could not generate .vloupe file: \n%s" % e.output)
def get_unmapped_read_count_from_indexed_bam(bam_file_name):
"""
Get number of unmapped reads from an indexed BAM file.
Args:
bam_file_name (str): Name of indexed BAM file.
Returns:
int: number of unmapped reads in the BAM
Note:
BAM must be indexed for lookup using samtools.
"""
index_output = tk_subproc.check_output('samtools idxstats %s' %
bam_file_name,
shell=True)
return int(index_output.strip().split('\n')[-1].split()[-1])
