def load_json(json) -> GraphContainer:
"""
Construct graph object from JSON representation
:param json: Dictionary of JSON file contents
"""
graph = GraphContainer()
for node in json["nodes"]:
seqs = node.get("sequences", ())
if "reference" in node:
chrom, start, end = parse_region(node["reference"])
graph.add_refNode(chrom, start, end, seqs, node["name"])
elif "position" in node:
chrom, start, end = parse_region(node["position"])
graph.add_altNode(chrom, start, end, node["sequence"], seqs,
node["name"])
else:
graph.nodes[node["name"]] = node
for edge in json["edges"]:
seqs = edge.get("sequences", ())
graph.add_edge(graph.nodes[edge["from"]], graph.nodes[edge["to"]],
seqs)
graph.name = json["model_name"]
graph.paths = json.get("paths", [])
graph.target_regions = json.get("target_regions", [])
graph.check()
return graph
def add_reference_information(paragraph_dict, reference_fasta):
""" Adds reference sequence information to reference nodes """
fasta = pysam.FastaFile(reference_fasta)
for n in paragraph_dict["nodes"]:
if "reference" in n:
chrom, start, end = parse_region(n["reference"])
n["reference_sequence"] = fasta.fetch(chrom, start - 1, end).upper()
def _parse_breakend(self, alt):
""" Parse remote breakend info from BND VCF record
:param record: VCF record
:param alt: Alt allele from VCF record (BND)
:return Inserted sequence and position of first base after remote breakend
"""
# We only support forward strand breakends.
be_match = re.match(r'([ACGTNXacgtnx]+)([\[\]])([^\[\]]+)([\[\]])',
alt)
if not be_match:
raise Exception("Unsupported breakend ALT: %s" % alt)
ins_sequence = be_match.group(1)
be_direction1 = be_match.group(2)
be_pos = be_match.group(3)
be_direction2 = be_match.group(4)
be_chrom, be_start, be_end = parse_region(be_pos)
if be_direction1 != "[" or be_direction2 != "[":
raise Exception("Reverse-comp breakends are not supported.")
assert not be_end
if be_chrom != self.chrom:
raise Exception("Breakends across chromosomes are not supported.")
return ins_sequence, be_start
def run_vcf2paragraph(event_and_args):
"""
run vcf2paragraph for one single variant
"""
event = event_and_args[0]
params = event_and_args[1]
tempfiles = []
result = {}
try:
logging.debug("Converting: %s", str(event))
result["graph"] = convert_vcf(
event,
params["reference"],
None,
ref_node_padding=params["read_length"],
ref_node_max_length=params["max_ref_node_length"],
allele_graph=params["graph_type"] == "alleles",
alt_splitting=params["alt_splitting"],
alt_paths=params["alt_paths"])
chrom = None
start = None
end = None
if "vcf_records" in result["graph"]:
for r in result["graph"]["vcf_records"]:
if chrom is None:
chrom = r["chrom"]
else:
assert chrom == r["chrom"]
if start is None:
start = r["pos"]
else:
start = min(start, r["pos"])
if end is None:
end = r["end"]
else:
end = max(end, r["end"])
else:
for tr in result["graph"]["target_regions"]:
c, s, e = parse_region(tr)
if chrom is None:
chrom = c
else:
assert chrom == c
if start is None:
start = s
else:
start = min(start, s)
if end is None:
end = e
else:
end = max(end, e)
assert chrom is not None
assert start is not None
assert end is not None
result["chrom"] = chrom
result["start"] = start
result["end"] = end
except Exception: # pylint: disable=broad-except
logging.error("Exception when running vcf2paragraph on %s", str(event))
traceback.print_exc(file=LoggingWriter(logging.ERROR))
raise
finally:
for x in tempfiles:
try:
os.remove(x)
except: # pylint: disable=bare-except
pass
if params["retrieve_reference_sequence"]:
add_reference_information(result["graph"], params["reference"])
return result