def test_sanity():
"""Test to build variant dicts"""
vcf = get_vcf_file()
head = get_header(vcf)
header_line = head.header
for line in vcf:
if not line.startswith('#'):
variant = get_variant_dict(variant_line=line,
header_line=header_line)
def test_sanity():
"""Test to build variant dicts"""
vcf = get_vcf_file()
head = get_header(vcf)
header_line = head.header
for line in vcf:
if not line.startswith('#'):
variant = get_variant_dict(
variant_line = line,
header_line = header_line
)
def test_wrong_variant():
"""Test to build a variant dict with a wrong formatted variant"""
vcf = get_vcf_file()
head = get_header(vcf)
header_line = head.header
# Missing one gt call
wrong_variant = '3\t973348\t.\tG\tA\t100\tPASS\tMQ=1\tGT:GQ\t0/0:60\t0/0:60\n'
with pytest.raises(SyntaxError):
variant = get_variant_dict(variant_line=wrong_variant,
header_line=header_line)
def test_wrong_variant():
"""Test to build a variant dict with a wrong formatted variant"""
vcf = get_vcf_file()
head = get_header(vcf)
header_line = head.header
# Missing one gt call
wrong_variant = '3\t973348\t.\tG\tA\t100\tPASS\tMQ=1\tGT:GQ\t0/0:60\t0/0:60\n'
with pytest.raises(SyntaxError):
variant = get_variant_dict(
variant_line = wrong_variant,
header_line = header_line
)
def _formated_variants(self, raw_variants, case_obj):
"""Return variant objects
Args:
raw_variants (Iterable): An iterable with variant lines
case_obj (puzzle.nodels.Case): A case object
"""
vcf_file_path = case_obj.variant_source
logger.info("Parsing file {0}".format(vcf_file_path))
head = HeaderParser()
handle = get_vcf_handle(infile=vcf_file_path)
# Parse the header
for line in handle:
line = line.rstrip()
if line.startswith("#"):
if line.startswith("##"):
head.parse_meta_data(line)
else:
head.parse_header_line(line)
else:
break
handle.close()
header_line = head.header
# Get the individual ids for individuals in vcf file
vcf_individuals = set([ind_id for ind_id in head.individuals])
variant_columns = ["CHROM", "POS", "ID", "REF", "ALT", "QUAL", "FILTER"]
vep_header = head.vep_columns
snpeff_header = head.snpeff_columns
index = 0
for variant_line in raw_variants:
if not variant_line.startswith("#"):
index += 1
# Create a variant dict:
variant_dict = get_variant_dict(variant_line=variant_line, header_line=header_line)
variant_dict["CHROM"] = variant_dict["CHROM"].lstrip("chrCHR")
# Crreate a info dict:
info_dict = get_info_dict(info_line=variant_dict["INFO"])
# Check if vep annotation:
vep_string = info_dict.get("CSQ")
# Check if snpeff annotation:
snpeff_string = info_dict.get("ANN")
if vep_string:
# Get the vep annotations
vep_info = get_vep_info(vep_string=vep_string, vep_header=vep_header)
elif snpeff_string:
# Get the vep annotations
snpeff_info = get_snpeff_info(snpeff_string=snpeff_string, snpeff_header=snpeff_header)
variant = Variant(**{column: variant_dict.get(column, ".") for column in variant_columns})
logger.debug("Creating a variant object of variant {0}".format(variant.get("variant_id")))
variant["index"] = index
logger.debug("Updating index to: {0}".format(index))
variant["start"] = int(variant_dict["POS"])
if self.variant_type == "sv":
other_chrom = variant["CHROM"]
# If we have a translocation:
if ":" in variant_dict["ALT"] and not "<" in variant_dict["ALT"]:
other_coordinates = variant_dict["ALT"].strip("ACGTN[]").split(":")
other_chrom = other_coordinates[0].lstrip("chrCHR")
other_position = other_coordinates[1]
variant["stop"] = other_position
# Set 'infinity' to length if translocation
variant["sv_len"] = float("inf")
else:
variant["stop"] = int(info_dict.get("END", variant_dict["POS"]))
variant["sv_len"] = variant["stop"] - variant["start"]
variant["stop_chrom"] = other_chrom
else:
variant["stop"] = int(variant_dict["POS"]) + (len(variant_dict["REF"]) - len(variant_dict["ALT"]))
variant["sv_type"] = info_dict.get("SVTYPE")
variant["cytoband_start"] = get_cytoband_coord(chrom=variant["CHROM"], pos=variant["start"])
if variant.get("stop_chrom"):
variant["cytoband_stop"] = get_cytoband_coord(chrom=variant["stop_chrom"], pos=variant["stop"])
# It would be easy to update these keys...
thousand_g = info_dict.get("1000GAF")
if thousand_g:
logger.debug("Updating thousand_g to: {0}".format(thousand_g))
variant["thousand_g"] = float(thousand_g)
variant.add_frequency("1000GAF", variant.get("thousand_g"))
# SV specific tag for number of occurances
occurances = info_dict.get("OCC")
if occurances:
logger.debug("Updating occurances to: {0}".format(occurances))
variant["occurances"] = float(occurances)
variant.add_frequency("OCC", occurances)
cadd_score = info_dict.get("CADD")
if cadd_score:
logger.debug("Updating cadd_score to: {0}".format(cadd_score))
variant["cadd_score"] = float(cadd_score)
rank_score_entry = info_dict.get("RankScore")
if rank_score_entry:
for family_annotation in rank_score_entry.split(","):
rank_score = family_annotation.split(":")[-1]
logger.debug("Updating rank_score to: {0}".format(rank_score))
variant["rank_score"] = float(rank_score)
genetic_models_entry = info_dict.get("GeneticModels")
if genetic_models_entry:
genetic_models = []
for family_annotation in genetic_models_entry.split(","):
for genetic_model in family_annotation.split(":")[-1].split("|"):
genetic_models.append(genetic_model)
logger.debug("Updating rank_score to: {0}".format(rank_score))
variant["genetic_models"] = genetic_models
# Add genotype calls:
for individual in case_obj.individuals:
sample_id = individual.ind_id
if sample_id in vcf_individuals:
raw_call = dict(zip(variant_dict["FORMAT"].split(":"), variant_dict[sample_id].split(":")))
variant.add_individual(
Genotype(
sample_id=sample_id,
genotype=raw_call.get("GT", "./."),
case_id=individual.case_name,
phenotype=individual.phenotype,
ref_depth=raw_call.get("AD", ",").split(",")[0],
alt_depth=raw_call.get("AD", ",").split(",")[1],
genotype_quality=raw_call.get("GQ", "."),
depth=raw_call.get("DP", "."),
supporting_evidence=raw_call.get("SU", "0"),
pe_support=raw_call.get("PE", "0"),
sr_support=raw_call.get("SR", "0"),
)
)
# Add transcript information:
gmaf = None
if vep_string:
for transcript_info in vep_info:
transcript = self._get_vep_transcripts(transcript_info)
gmaf_raw = transcript_info.get("GMAF")
if gmaf_raw:
gmaf = float(gmaf_raw.split(":")[-1])
variant.add_transcript(transcript)
if gmaf:
variant.add_frequency("GMAF", gmaf)
if not variant.thousand_g:
variant.thousand_g = gmaf
elif snpeff_string:
for transcript_info in snpeff_info:
transcript = self._get_snpeff_transcripts(transcript_info)
variant.add_transcript(transcript)
variant["most_severe_consequence"] = get_most_severe_consequence(variant["transcripts"])
for gene in self._get_genes(variant):
variant.add_gene(gene)
self._add_compounds(variant=variant, info_dict=info_dict)
yield variant
def father(ctx):
"""Check number of variants in common"""
logger.info("Running variant_integrity father version {0}".format(
variant_integrity.__version__))
print_columns = ['ind_id', 'fraction_of_common_variants', 'common_variants', 'number_calls']
# Children is a dictionary of children that counts the number of errors
duos = []
children = {}
analysis_individuals = set()
for ind_id in ctx.parent.individuals:
individual_object = ctx.parent.individuals[ind_id]
if individual_object.father != '0':
duo = {
'child': ind_id,
'father': individual_object.father
}
analysis_individuals.add(ind_id)
analysis_individuals.add(individual_object.father)
duos.append(duo)
logger.info("Duo found: {0}".format(
', '.join(list(duo.values()))
))
logger.info("Individuals included in analysis: {0}".format(
','.join(list(analysis_individuals))))
for duo in duos:
children[duo['child']] = dict(zip(
print_columns, [duo['child'], 0, 0, 0]))
for line in ctx.parent.variant_file:
variant_dict = get_variant_dict(
variant_line=line,
header_line=ctx.parent.header_line
)
logger.debug("Checking genotype calls for variant {0}".format(
get_variant_id(variant_dict=variant_dict)
))
genotypes = get_genotypes(variant_dict, analysis_individuals)
for duo in duos:
child_id = duo['child']
child_genotype = genotypes[child_id]
father_genotype = genotypes[duo['father']]
duo_genotypes = [
child_genotype,
father_genotype
]
#First check if the child has the variant:
if child_genotype.has_variant:
# If child have high quality we count the variant
if check_high_quality([child_genotype], ctx.parent.gq_treshold):
children[child_id]['number_calls'] += 1
if check_common_variant(duo_genotypes):
children[child_id]['common_variants'] += 1
results = []
for child_id in children:
child_info = children[child_id]
common = child_info['common_variants']
variants = child_info['number_calls']
percentage = common/variants
child_info['fraction_of_common_variants'] = round(percentage, 3)
results.append(child_info)
to_json = ctx.parent.to_json
outfile = ctx.parent.outfile
if to_json:
if outfile:
json.dump(results, outfile)
else:
print(json.dumps(results))
else:
if outfile:
outfile.write("#{0}\n".format('\t'.join(print_columns)))
else:
print("#{0}".format('\t'.join(print_columns)))
for result in results:
print_line = "{0}\t{1}\t{2}\t{3}".format(
result['ind_id'], result['fraction_of_common_variants'],
result['common_variants'], result['number_calls']
)
if outfile:
outfile.write("{0}\n".format(print_line))
else:
print(print_line)
def mendel(ctx):
"""Check mendelian errors in all trios"""
logger.info("Running variant_integrity mendel {0}".format(
variant_integrity.__version__))
print_columns = ['ind_id', 'fraction_of_errors', 'mendelian_errors', 'number_calls']
# Children is a dictionary of children that counts the number of errors
trios = []
children = {}
analysis_individuals = set()
for family in ctx.parent.families:
family_object = ctx.parent.families[family]
for trio in family_object.trios:
trio_individuals = {
'mother':None,
'father':None,
'child':None,
}
for ind_id in trio:
analysis_individuals.add(ind_id)
individual_object = ctx.parent.individuals[ind_id]
if individual_object.mother in trio:
trio_individuals['child'] = ind_id
elif individual_object.sex == 1:
trio_individuals['father'] = ind_id
else:
trio_individuals['mother'] = ind_id
trios.append(trio_individuals)
logger.info("Trio found: {0}".format(
', '.join(list(trio_individuals.values()))
))
logger.info("Individuals included in analysis: {0}".format(
','.join(list(analysis_individuals))))
for trio in trios:
children[trio['child']] = dict(zip(
print_columns, [trio['child'], 0, 0, 0]))
for line in ctx.parent.variant_file:
variant_dict = get_variant_dict(
variant_line=line,
header_line=ctx.parent.header_line
)
logger.debug("Checking genotype calls for variant {0}".format(
get_variant_id(variant_dict=variant_dict)
))
genotypes = get_genotypes(variant_dict, analysis_individuals)
for trio in trios:
child_id = trio['child']
child_genotype = genotypes[child_id]
mother_genotype = genotypes[trio['mother']]
father_genotype = genotypes[trio['father']]
trio_genotypes = [
child_genotype,
mother_genotype,
father_genotype
]
#First check if the child has the variant:
if child_genotype.has_variant:
# If all individuals are high quality we count the variant
if check_high_quality(trio_genotypes, ctx.parent.gq_treshold):
children[child_id]['number_calls'] += 1
if check_mendelian_error(child_genotype, mother_genotype, father_genotype):
children[child_id]['mendelian_errors'] += 1
results = []
for child_id in children:
child_info = children[child_id]
errors = child_info['mendelian_errors']
variants = child_info['number_calls']
percentage = errors/variants
child_info['fraction_of_errors'] = round(percentage, 3)
results.append(child_info)
to_json = ctx.parent.to_json
outfile = ctx.parent.outfile
if to_json:
if outfile:
json.dump(results, outfile)
else:
print(json.dumps(results))
else:
if outfile:
outfile.write("#{0}\n".format('\t'.join(print_columns)))
else:
print("#{0}".format('\t'.join(print_columns)))
for result in results:
print_line = "{0}\t{1}\t{2}\t{3}".format(
result['ind_id'], result['fraction_of_errors'],
result['mendelian_errors'], result['number_calls']
)
if outfile:
outfile.write("{0}\n".format(print_line))
else:
print(print_line)
def _formated_variants(self, raw_variants, case_obj):
"""Return variant objects
Args:
raw_variants (Iterable): An iterable with variant lines
case_obj (puzzle.nodels.Case): A case object
"""
vcf_file_path = case_obj.variant_source
logger.info("Parsing file {0}".format(vcf_file_path))
head = HeaderParser()
handle = get_vcf_handle(infile=vcf_file_path)
# Parse the header
for line in handle:
line = line.rstrip()
if line.startswith('#'):
if line.startswith('##'):
head.parse_meta_data(line)
else:
head.parse_header_line(line)
else:
break
handle.close()
header_line = head.header
# Get the individual ids for individuals in vcf file
vcf_individuals = set([ind_id for ind_id in head.individuals])
variant_columns = ['CHROM', 'POS', 'ID', 'REF', 'ALT', 'QUAL', 'FILTER']
vep_header = head.vep_columns
snpeff_header = head.snpeff_columns
index = 0
for variant_line in raw_variants:
if not variant_line.startswith('#'):
index += 1
#Create a variant dict:
variant_dict = get_variant_dict(
variant_line = variant_line,
header_line = header_line
)
variant_dict['CHROM'] = variant_dict['CHROM'].lstrip('chrCHR')
#Crreate a info dict:
info_dict = get_info_dict(
info_line = variant_dict['INFO']
)
#Check if vep annotation:
vep_string = info_dict.get('CSQ')
#Check if snpeff annotation:
snpeff_string = info_dict.get('ANN')
if vep_string:
#Get the vep annotations
vep_info = get_vep_info(
vep_string = vep_string,
vep_header = vep_header
)
elif snpeff_string:
#Get the vep annotations
snpeff_info = get_snpeff_info(
snpeff_string = snpeff_string,
snpeff_header = snpeff_header
)
variant = Variant(
**{column: variant_dict.get(column, '.')
for column in variant_columns}
)
logger.debug("Creating a variant object of variant {0}".format(
variant.get('variant_id')))
variant['index'] = index
logger.debug("Updating index to: {0}".format(
index))
variant['start'] = int(variant_dict['POS'])
if self.variant_type == 'sv':
other_chrom = variant['CHROM']
# If we have a translocation:
if ':' in variant_dict['ALT']:
other_coordinates = variant_dict['ALT'].strip('ACGTN[]').split(':')
other_chrom = other_coordinates[0].lstrip('chrCHR')
other_position = other_coordinates[1]
variant['stop'] = other_position
#Set 'infinity' to length if translocation
variant['sv_len'] = float('inf')
else:
variant['stop'] = int(info_dict.get('END', variant_dict['POS']))
variant['sv_len'] = variant['stop'] - variant['start']
variant['stop_chrom'] = other_chrom
else:
variant['stop'] = int(variant_dict['POS']) + \
(len(variant_dict['REF']) - len(variant_dict['ALT']))
variant['sv_type'] = info_dict.get('SVTYPE')
variant['cytoband_start'] = get_cytoband_coord(
chrom=variant['CHROM'],
pos=variant['start'])
if variant.get('stop_chrom'):
variant['cytoband_stop'] = get_cytoband_coord(
chrom=variant['stop_chrom'],
pos=variant['stop'])
# It would be easy to update these keys...
thousand_g = info_dict.get('1000GAF')
if thousand_g:
logger.debug("Updating thousand_g to: {0}".format(
thousand_g))
variant['thousand_g'] = float(thousand_g)
variant.add_frequency('1000GAF', variant.get('thousand_g'))
#SV specific tag for number of occurances
occurances = info_dict.get('OCC')
if occurances:
logger.debug("Updating occurances to: {0}".format(
occurances))
variant['occurances'] = float(occurances)
variant.add_frequency('OCC', occurances)
cadd_score = info_dict.get('CADD')
if cadd_score:
logger.debug("Updating cadd_score to: {0}".format(
cadd_score))
variant['cadd_score'] = float(cadd_score)
rank_score_entry = info_dict.get('RankScore')
if rank_score_entry:
for family_annotation in rank_score_entry.split(','):
rank_score = family_annotation.split(':')[-1]
logger.debug("Updating rank_score to: {0}".format(
rank_score))
variant['rank_score'] = float(rank_score)
genetic_models_entry = info_dict.get('GeneticModels')
if genetic_models_entry:
genetic_models = []
for family_annotation in genetic_models_entry.split(','):
for genetic_model in family_annotation.split(':')[-1].split('|'):
genetic_models.append(genetic_model)
logger.debug("Updating rank_score to: {0}".format(
rank_score))
variant['genetic_models'] = genetic_models
#Add genotype calls:
for individual in case_obj.individuals:
sample_id = individual.ind_id
if sample_id in vcf_individuals:
raw_call = dict(zip(
variant_dict['FORMAT'].split(':'),
variant_dict[sample_id].split(':'))
)
variant.add_individual(Genotype(
sample_id = sample_id,
genotype = raw_call.get('GT', './.'),
case_id = individual.case_name,
phenotype = individual.phenotype,
ref_depth = raw_call.get('AD', ',').split(',')[0],
alt_depth = raw_call.get('AD', ',').split(',')[1],
genotype_quality = raw_call.get('GQ', '.'),
depth = raw_call.get('DP', '.'),
supporting_evidence = raw_call.get('SU', '0'),
pe_support = raw_call.get('PE', '0'),
sr_support = raw_call.get('SR', '0'),
))
# Add transcript information:
if vep_string:
for transcript in self._get_vep_transcripts(variant, vep_info):
variant.add_transcript(transcript)
elif snpeff_string:
for transcript in self._get_snpeff_transcripts(variant, snpeff_info):
variant.add_transcript(transcript)
variant['most_severe_consequence'] = get_most_severe_consequence(
variant['transcripts']
)
for gene in self._get_genes(variant):
variant.add_gene(gene)
self._add_compounds(variant=variant, info_dict=info_dict)
yield variant
def _format_variant(self, variant_line, index, case_obj, head):
"""Return variant objects
Args:
raw_variants (Iterable): An iterable with variant lines
case_obj (puzzle.nodels.Case): A case object
"""
header_line = head.header
# Get the individual ids for individuals in vcf file
vcf_individuals = set([ind_id for ind_id in head.individuals])
variant_columns = ['CHROM', 'POS', 'ID', 'REF', 'ALT', 'QUAL', 'FILTER']
vep_header = head.vep_columns
snpeff_header = head.snpeff_columns
#Create a variant dict:
variant_dict = get_variant_dict(
variant_line = variant_line,
header_line = header_line
)
variant_dict['CHROM'] = variant_dict['CHROM'].lstrip('chrCHR')
#Crreate a info dict:
info_dict = get_info_dict(
info_line = variant_dict['INFO']
)
#Check if vep annotation:
vep_string = info_dict.get('CSQ')
#Check if snpeff annotation:
snpeff_string = info_dict.get('ANN')
if vep_string:
#Get the vep annotations
vep_info = get_vep_info(
vep_string = vep_string,
vep_header = vep_header
)
elif snpeff_string:
#Get the vep annotations
snpeff_info = get_snpeff_info(
snpeff_string = snpeff_string,
snpeff_header = snpeff_header
)
variant = Variant(
**{column: variant_dict.get(column, '.')
for column in variant_columns}
)
logger.debug("Creating a variant object of variant {0}".format(
variant.get('variant_id')))
variant['index'] = index
logger.debug("Updating index to: {0}".format(
index))
variant['start'] = int(variant_dict['POS'])
if self.variant_type == 'sv':
other_chrom = variant['CHROM']
# If we have a translocation:
if ':' in variant_dict['ALT'] and not '<' in variant_dict['ALT']:
other_coordinates = variant_dict['ALT'].strip('ACGTN[]').split(':')
other_chrom = other_coordinates[0].lstrip('chrCHR')
other_position = other_coordinates[1]
variant['stop'] = other_position
#Set 'infinity' to length if translocation
variant['sv_len'] = float('inf')
else:
variant['stop'] = int(info_dict.get('END', variant_dict['POS']))
variant['sv_len'] = variant['stop'] - variant['start']
variant['stop_chrom'] = other_chrom
else:
variant['stop'] = int(variant_dict['POS']) + \
(len(variant_dict['REF']) - len(variant_dict['ALT']))
variant['sv_type'] = info_dict.get('SVTYPE')
variant['cytoband_start'] = get_cytoband_coord(
chrom=variant['CHROM'],
pos=variant['start'])
if variant.get('stop_chrom'):
variant['cytoband_stop'] = get_cytoband_coord(
chrom=variant['stop_chrom'],
pos=variant['stop'])
# It would be easy to update these keys...
thousand_g = info_dict.get('1000GAF')
if thousand_g:
logger.debug("Updating thousand_g to: {0}".format(
thousand_g))
variant['thousand_g'] = float(thousand_g)
variant.add_frequency('1000GAF', variant.get('thousand_g'))
#SV specific tag for number of occurances
occurances = info_dict.get('OCC')
if occurances:
logger.debug("Updating occurances to: {0}".format(
occurances))
variant['occurances'] = float(occurances)
variant.add_frequency('OCC', occurances)
cadd_score = info_dict.get('CADD')
if cadd_score:
logger.debug("Updating cadd_score to: {0}".format(
cadd_score))
variant['cadd_score'] = float(cadd_score)
rank_score_entry = info_dict.get('RankScore')
if rank_score_entry:
for family_annotation in rank_score_entry.split(','):
rank_score = family_annotation.split(':')[-1]
logger.debug("Updating rank_score to: {0}".format(
rank_score))
variant['rank_score'] = float(rank_score)
genetic_models_entry = info_dict.get('GeneticModels')
if genetic_models_entry:
genetic_models = []
for family_annotation in genetic_models_entry.split(','):
for genetic_model in family_annotation.split(':')[-1].split('|'):
genetic_models.append(genetic_model)
logger.debug("Updating rank_score to: {0}".format(
rank_score))
variant['genetic_models'] = genetic_models
#Add genotype calls:
for individual in case_obj.individuals:
sample_id = individual.ind_id
if sample_id in vcf_individuals:
raw_call = dict(zip(
variant_dict['FORMAT'].split(':'),
variant_dict[sample_id].split(':'))
)
genotype = Genotype(**raw_call)
variant.add_individual(puzzle_genotype(
sample_id = sample_id,
genotype = genotype.genotype,
case_id = individual.case_name,
phenotype = individual.phenotype,
ref_depth = genotype.ref_depth,
alt_depth = genotype.alt_depth,
genotype_quality = genotype.genotype_quality,
depth = genotype.depth_of_coverage,
supporting_evidence = genotype.supporting_evidence,
pe_support = genotype.pe_support,
sr_support = genotype.sr_support,
))
# Add transcript information:
gmaf = None
if vep_string:
for transcript_info in vep_info:
transcript = self._get_vep_transcripts(transcript_info)
gmaf_raw = transcript_info.get('GMAF')
if gmaf_raw:
gmaf = float(gmaf_raw.split(':')[-1])
variant.add_transcript(transcript)
if gmaf:
variant.add_frequency('GMAF', gmaf)
if not variant.thousand_g:
variant.thousand_g = gmaf
elif snpeff_string:
for transcript_info in snpeff_info:
transcript = self._get_snpeff_transcripts(transcript_info)
variant.add_transcript(transcript)
most_severe_consequence = get_most_severe_consequence(
variant['transcripts']
)
if most_severe_consequence:
variant['most_severe_consequence'] = most_severe_consequence
variant['impact_severity'] = IMPACT_SEVERITIES.get(most_severe_consequence)
for gene in self._get_genes(variant):
variant.add_gene(gene)
self._add_compounds(variant=variant, info_dict=info_dict)
return variant