class CoverageParser(object):
def __init__(
self,
sample,
panel_file_paths,
kmer,
force,
seq=None,
ctx=None,
threads=2,
memory="1GB",
panels=None,
verbose=True,
tmp_dir="tmp/",
skeleton_dir="atlas/data/skeletons/",
):
self.sample = sample
self.seq = seq
self.ctx = ctx
self.kmer = kmer
self.force = force
self.covgs = {"variant": {}, "presence": {}}
self.variant_covgs = self.covgs["variant"]
self.gene_presence_covgs = self.covgs["presence"]
self.mc_cortex_runner = None
self.verbose = verbose
self.skeleton_dir = skeleton_dir
self.tmp_dir = tmp_dir
self.panel_file_paths = panel_file_paths
self.panels = []
self.threads = threads
self.memory = memory
for panel_file_path in self.panel_file_paths:
panel = Panel(panel_file_path)
self.panels.append(panel)
if self.seq and self.ctx:
raise ValueError("Can't have both -1 and -c")
def run(self):
self._run_cortex()
self._parse_covgs()
def _run_cortex(self):
self.mc_cortex_runner = McCortexGenoRunner(
sample=self.sample,
panels=self.panels,
seq=self.seq,
ctx=self.ctx,
kmer=self.kmer,
force=self.force,
threads=self.threads,
memory=self.memory,
panel_name=self.panel_name,
tmp_dir=self.tmp_dir,
skeleton_dir=self.skeleton_dir,
)
self.mc_cortex_runner.run()
def estimate_depth(self):
depth = []
for variant_covg in self.variant_covgs.values():
if variant_covg.reference_coverage.median_depth > 0:
depth.append(variant_covg.reference_coverage.median_depth)
for spcs in self.gene_presence_covgs.values():
__median_depth = median(
[spc.median_depth for spc in spcs.values()])
if __median_depth > 0:
depth.append(__median_depth)
_median = median(depth)
if _median < 1:
return 1
else:
return _median
def remove_temporary_files(self):
self.mc_cortex_runner.remove_temporary_files()
@property
def panel_name(self):
return "-".join([panel.name for panel in self.panels])
def _parse_summary_covgs_row(self, row):
try:
return (
row[0],
int(row[2]),
int(row[3]),
100 * float(row[4]),
int(row[5]),
int(row[6]),
)
except ValueError:
logger.warning("Failed to parse %s" % str(row))
return row[0], 0, 0, 0.0, 0, 0
def _parse_covgs(self):
with open(self.mc_cortex_runner.covg_tmp_file_path, "r") as infile:
self.reader = csv.reader(infile, delimiter="\t")
for row in self.reader:
allele, median_depth, min_depth, percent_coverage, k_count, klen = self._parse_summary_covgs_row(
row)
allele_name = allele.split("?")[0]
if self._is_variant_panel(allele_name):
self._parse_variant_panel(row)
else:
self._parse_seq_panel(row)
def _is_variant_panel(self, allele_name):
try:
alt_or_ref = allele_name.split("-")[0]
return alt_or_ref in ["ref", "alt"]
except ValueError:
return False
def _parse_seq_panel(self, row):
allele, median_depth, min_depth, percent_coverage, k_count, klen = self._parse_summary_covgs_row(
row)
probe_coverage = ProbeCoverage(
percent_coverage=percent_coverage,
median_depth=median_depth,
min_depth=min_depth,
k_count=k_count,
klen=klen,
)
allele_name = allele.split("?")[0]
params = get_params(allele)
panel_type = params.get("panel_type", "presence")
name = params.get("name")
version = params.get("version", "1")
if panel_type in ["variant", "presence"]:
sequence_probe_coverage = SequenceProbeCoverage(
name=name,
probe_coverage=probe_coverage,
version=version,
length=params.get("length"),
)
try:
self.covgs[panel_type][name][version] = sequence_probe_coverage
except KeyError:
self.covgs[panel_type][name] = {}
self.covgs[panel_type][name][version] = sequence_probe_coverage
else:
# Species panels are treated differently
l = int(params.get("length", -1))
try:
self.covgs[panel_type][name]["total_bases"] += l
if percent_coverage > 75 and median_depth > 0:
self.covgs[panel_type][name]["percent_coverage"].append(
percent_coverage)
self.covgs[panel_type][name]["length"].append(l)
self.covgs[panel_type][name]["median"].append(median_depth)
except KeyError:
if panel_type not in self.covgs:
self.covgs[panel_type] = {}
self.covgs[panel_type][name] = {}
self.covgs[panel_type][name]["total_bases"] = l
if percent_coverage > 75 and median_depth > 0:
self.covgs[panel_type][name]["percent_coverage"] = [
percent_coverage
]
self.covgs[panel_type][name]["length"] = [l]
self.covgs[panel_type][name]["median"] = [median_depth]
else:
self.covgs[panel_type][name]["percent_coverage"] = []
self.covgs[panel_type][name]["length"] = []
self.covgs[panel_type][name]["median"] = []
def _parse_variant_panel(self, row):
probe, median_depth, min_depth, percent_coverage, k_count, klen = self._parse_summary_covgs_row(
row)
params = get_params(probe)
probe_type = probe.split("-")[0]
if "var_name" in params:
var_name = (params.get("gene", "") + "_" + params.get("mut", "") +
"-" + params.get("var_name", ""))
else:
var_name = allele.split("?")[0].split("-")[1]
if not var_name in self.variant_covgs:
variant_probe_coverage = VariantProbeCoverage(
reference_coverages=[],
alternate_coverages=[],
var_name=probe,
params=params,
)
self.variant_covgs[var_name] = variant_probe_coverage
probe_coverage = ProbeCoverage(
min_depth=min_depth,
k_count=k_count,
percent_coverage=percent_coverage,
median_depth=median_depth,
klen=klen,
)
if probe_type == "ref":
self.variant_covgs[var_name].reference_coverages.append(
probe_coverage)
self.variant_covgs[
var_name].best_reference_coverage = self.variant_covgs[
var_name]._choose_best_reference_coverage()
elif probe_type == "alt":
self.variant_covgs[var_name].alternate_coverages.append(
probe_coverage)
self.variant_covgs[
var_name].best_alternate_coverage = self.variant_covgs[
var_name]._choose_best_alternate_coverage()
else:
raise ValueError("probe_type must be ref or alt")
class CoverageParser(object):
def __init__(self,
sample,
panel_file_paths,
kmer,
force,
seq=None,
ctx=None,
threads=2,
memory="1GB",
panels=None,
verbose=True,
tmp_dir='tmp/',
skeleton_dir='atlas/data/skeletons/',
mccortex31_path="mccortex31"):
self.sample = sample
self.seq = seq
self.ctx = ctx
self.kmer = kmer
self.force = force
self.covgs = {"variant": {}, "presence": {}}
self.variant_covgs = self.covgs["variant"]
self.gene_presence_covgs = self.covgs["presence"]
self.mc_cortex_runner = None
self.verbose = verbose
self.skeleton_dir = skeleton_dir
self.tmp_dir = tmp_dir
self.panel_file_paths = panel_file_paths
self.panels = []
self.mccortex31_path = mccortex31_path
self.threads = threads
self.memory = memory
for panel_file_path in self.panel_file_paths:
panel = Panel(panel_file_path)
self.panels.append(panel)
if self.seq and self.ctx:
raise ValueError("Can't have both -1 and -c")
def run(self):
self._run_cortex()
self._parse_covgs()
def _run_cortex(self):
self.mc_cortex_runner = McCortexGenoRunner(
sample=self.sample,
panels=self.panels,
seq=self.seq,
ctx=self.ctx,
kmer=self.kmer,
force=self.force,
threads=self.threads,
memory=self.memory,
panel_name=self.panel_name,
tmp_dir=self.tmp_dir,
skeleton_dir=self.skeleton_dir,
mccortex31_path=self.mccortex31_path)
self.mc_cortex_runner.run()
def estimate_depth(self):
depth = []
for variant_coverages in self.variant_covgs.values():
for variant_covg in variant_coverages:
if variant_covg.reference_coverage.median_depth > 0:
depth.append(variant_covg.reference_coverage.median_depth)
for spcs in self.gene_presence_covgs.values():
__median_depth = median(
[spc.median_depth for spc in spcs.values()])
if __median_depth > 0:
depth.append(__median_depth)
_median = median(depth)
if _median < 1:
return 1
else:
return _median
def remove_temporary_files(self):
self.mc_cortex_runner.remove_temporary_files()
@property
def panel_name(self):
return "-".join([panel.name for panel in self.panels])
def _parse_summary_covgs_row(self, row):
try:
return row[0], int(row[2]), int(row[3]), 100 * float(row[4]), int(
row[5])
except ValueError:
logger.warning("Failed to parse %s" % str(row))
return row[0], 0, 0, 0.0, 0
def _parse_covgs(self):
with open(self.mc_cortex_runner.covg_tmp_file_path, 'r') as infile:
self.reader = csv.reader(infile, delimiter="\t")
for row in self.reader:
allele, median_depth, min_depth, percent_coverage, k_count = self._parse_summary_covgs_row(
row)
allele_name = allele.split('?')[0]
if self._is_variant_panel(allele_name):
self._parse_variant_panel(row)
else:
self._parse_seq_panel(row)
def _is_variant_panel(self, allele_name):
try:
alt_or_ref = allele_name.split('-')[0]
return alt_or_ref in ["ref", "alt"]
except ValueError:
return False
def _parse_seq_panel(self, row):
allele, median_depth, min_depth, percent_coverage, k_count = self._parse_summary_covgs_row(
row)
probe_coverage = ProbeCoverage(percent_coverage=percent_coverage,
median_depth=median_depth,
min_depth=min_depth,
k_count=k_count)
allele_name = allele.split('?')[0]
params = get_params(allele)
panel_type = params.get("panel_type", "presence")
name = params.get('name')
version = params.get('version', '1')
if panel_type in ["variant", "presence"]:
sequence_probe_coverage = SequenceProbeCoverage(
name=name,
probe_coverage=probe_coverage,
version=version,
length=params.get("length"))
try:
self.covgs[panel_type][name][version] = sequence_probe_coverage
except KeyError:
self.covgs[panel_type][name] = {}
self.covgs[panel_type][name][version] = sequence_probe_coverage
else:
# Species panels are treated differently
l = int(params.get("length", -1))
try:
self.covgs[panel_type][name]["total_bases"] += l
if percent_coverage > 75 and median_depth > 0:
self.covgs[panel_type][name]["percent_coverage"].append(
percent_coverage)
self.covgs[panel_type][name]["length"].append(l)
self.covgs[panel_type][name]["median"].append(median_depth)
except KeyError:
if panel_type not in self.covgs:
self.covgs[panel_type] = {}
self.covgs[panel_type][name] = {}
self.covgs[panel_type][name]["total_bases"] = l
if percent_coverage > 75 and median_depth > 0:
self.covgs[panel_type][name]["percent_coverage"] = [
percent_coverage
]
self.covgs[panel_type][name]["length"] = [l]
self.covgs[panel_type][name]["median"] = [median_depth]
else:
self.covgs[panel_type][name]["percent_coverage"] = []
self.covgs[panel_type][name]["length"] = []
self.covgs[panel_type][name]["median"] = []
def _parse_variant_panel(self, row):
allele, median_depth, min_depth, percent_coverage, k_count = self._parse_summary_covgs_row(
row)
params = get_params(allele)
if 'var_name' in params:
var_name = params.get('var_name')
else:
var_name = allele.split('?')[0].split('-')[1]
num_alts = int(params.get("num_alts", 0))
reference_coverages = [
ProbeCoverage(percent_coverage=percent_coverage,
median_depth=median_depth,
min_depth=min_depth,
k_count=k_count)
]
alt_or_ref = 'ref'
alternate_coverages = []
for i in range(num_alts - 1):
row = next(self.reader)
ref_allele, median_depth, min_depth, percent_coverage, k_count = self._parse_summary_covgs_row(
row)
if ref_allele.split('-')[0] != 'ref':
logger.warning("Fewer ref alleles than alt alleles for %s" %
ref_allele)
alternate_coverages.append(
ProbeCoverage(min_depth=min_depth,
k_count=k_count,
percent_coverage=percent_coverage,
median_depth=median_depth))
num_alts -= 1
break
assert ref_allele.split('-')[0] == 'ref'
reference_coverages.append(
ProbeCoverage(percent_coverage=percent_coverage,
median_depth=median_depth,
min_depth=min_depth,
k_count=k_count))
for i in range(num_alts):
row = next(self.reader)
alt_allele, median_depth, min_depth, percent_coverage, k_count = self._parse_summary_covgs_row(
row)
assert alt_allele.split('-')[0] == 'alt'
alternate_coverages.append(
ProbeCoverage(min_depth=min_depth,
k_count=k_count,
percent_coverage=percent_coverage,
median_depth=median_depth))
variant_probe_coverage = VariantProbeCoverage(
reference_coverages=reference_coverages,
alternate_coverages=alternate_coverages,
var_name=var_name,
params=params)
try:
self.variant_covgs[allele].append(variant_probe_coverage)
except KeyError:
self.variant_covgs[allele] = [variant_probe_coverage]