def get_variants_by_family_for_gene(mall,
family_list,
inheritance_mode,
gene_id,
variant_filter=None,
quality_filter=None,
user=None):
if variant_filter is None:
variant_filter = VariantFilter()
variant_filter.add_gene(gene_id)
by_family = {}
for family in family_list:
family_t = (family.project_id, family.family_id)
variants = list(
get_variants_with_inheritance_mode(
mall,
family,
inheritance_mode,
variant_filter,
quality_filter,
user=user,
))
by_family[family_t] = variants
return by_family
def get_variants_for_inheritance_for_project(project, inheritance_mode):
"""
Get the variants for this project / inheritance combo
Return dict of family -> list of variants
"""
# create search specification
# this could theoretically differ by project, if there are different reference populations
variant_filter = get_default_variant_filter('moderate_impact')
variant_filter.ref_freqs.append(('1kg_wgs_phase3', g1k_freq_threshold))
variant_filter.ref_freqs.append(('1kg_wgs_phase3_popmax', g1k_popmax_freq_threshold))
variant_filter.ref_freqs.append(('exac_v3', exac_freq_threshold))
variant_filter.ref_freqs.append(('exac_v3_popmax', exac_popmax_threshold))
quality_filter = {
'vcf_filter': 'pass',
'min_gq': GQ_threshold,
'min_ab': AB_threshold,
}
# run MendelianVariantSearch for each family, collect results
families = project.get_families()
for i, family in enumerate(families):
sys.stdout.write("Processing %s - family %s (%d / %d) .." % (inheritance_mode, family.family_id, i+1, len(families)))
variant_list = list(get_variants_with_inheritance_mode(
get_mall(project.project_id),
family.xfamily(),
inheritance_mode,
variant_filter=variant_filter,
quality_filter=quality_filter,
))
yield family, variant_list
print(" got %d variants" % len(variant_list))
def get_variants_for_inheritance_for_project(project, inheritance_mode):
"""
Get the variants for this project / inheritance combo
Return dict of family -> list of variants
"""
# create search specification
# this could theoretically differ by project, if there are different reference populations
variant_filter = get_default_variant_filter("moderate_impact")
variant_filter.ref_freqs.append(("1kg_wgs_phase3", g1k_freq_threshold))
variant_filter.ref_freqs.append(("1kg_wgs_phase3_popmax", g1k_popmax_freq_threshold))
variant_filter.ref_freqs.append(("exac_v3", exac_freq_threshold))
variant_filter.ref_freqs.append(("exac_v3_popmax", exac_popmax_threshold))
quality_filter = {"vcf_filter": "pass", "min_gq": GQ_threshold, "min_ab": AB_threshold}
# run MendelianVariantSearch for each family, collect results
families = project.get_families()
for i, family in enumerate(families):
sys.stdout.write(
"Processing %s - family %s (%d / %d) .." % (inheritance_mode, family.family_id, i + 1, len(families))
)
variant_list = list(
get_variants_with_inheritance_mode(
get_mall(project.project_id),
family.xfamily(),
inheritance_mode,
variant_filter=variant_filter,
quality_filter=quality_filter,
)
)
yield family, variant_list
print(" got %d variants" % len(variant_list))
def calculate_mendelian_variant_search(search_spec, xfamily):
variants = None
if search_spec.search_mode == 'standard_inheritance':
variants = list(get_variants_with_inheritance_mode(
get_mall(),
xfamily,
search_spec.inheritance_mode,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
elif search_spec.search_mode == 'custom_inheritance':
variants = list(get_variants_family(
get_datastore(),
xfamily,
genotype_filter=search_spec.genotype_inheritance_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
elif search_spec.search_mode == 'gene_burden':
gene_stream = get_genes_family(
get_datastore(),
get_reference(),
xfamily,
burden_filter=search_spec.gene_burden_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
)
variants = list(stream_utils.gene_stream_to_variant_stream(gene_stream, get_reference()))
elif search_spec.search_mode == 'allele_count':
variants = list(get_variants_allele_count(
get_datastore(),
xfamily,
search_spec.allele_count_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
elif search_spec.search_mode == 'all_variants':
variants = list(get_variants_family(
get_datastore(),
xfamily,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
return variants
def calculate_mendelian_variant_search(search_spec, xfamily):
sys.stderr.write(" mendelian_variant_search for %s - search mode: %s %s\n" % (xfamily.project_id, search_spec.search_mode, search_spec.__dict__))
variants = None
if search_spec.search_mode == 'standard_inheritance':
variants = list(get_variants_with_inheritance_mode(
get_mall(xfamily.project_id),
xfamily,
search_spec.inheritance_mode,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
))
elif search_spec.search_mode == 'custom_inheritance':
variants = list(get_variants_family(
get_datastore(xfamily.project_id),
xfamily,
genotype_filter=search_spec.genotype_inheritance_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
))
elif search_spec.search_mode == 'gene_burden':
gene_stream = get_genes_family(
get_datastore(xfamily.project_id),
get_reference(),
xfamily,
burden_filter=search_spec.gene_burden_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
)
variants = list(stream_utils.gene_stream_to_variant_stream(gene_stream, get_reference()))
elif search_spec.search_mode == 'allele_count':
variants = list(get_variants_allele_count(
get_datastore(xfamily.project_id),
xfamily,
search_spec.allele_count_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
))
elif search_spec.search_mode == 'all_variants':
variants = list(get_variants_family(
get_datastore(xfamily.project_id),
xfamily,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
indivs_to_consider=xfamily.indiv_id_list(),
))
return variants
def get_variants_for_inheritance_for_project(project, inheritance_mode):
"""
Get the variants for this project / inheritance combo
Return dict of family -> list of variants
"""
# create search specification
# this could theoretically differ by project, if there are different reference populations
#variant_filter = VariantFilter(so_annotations=SO_SEVERITY_ORDER, ref_freqs=[])
variant_filter = get_default_variant_filter('moderate_impact')
variant_filter.ref_freqs.append(('1kg_wgs_phase3', g1k_freq_threshold))
variant_filter.ref_freqs.append(
('1kg_wgs_phase3_popmax', g1k_popmax_freq_threshold))
variant_filter.ref_freqs.append(('exac_v3', exac_freq_threshold))
variant_filter.ref_freqs.append(('exac_v3_popmax', exac_popmax_threshold))
variant_filter.ref_freqs.append(
('merck-wgs-3793', merck_wgs_3793_threshold))
#variant_filter.ref_freqs.append(('merck-pcr-free-wgs-144', merck_wgs_144_threshold))
quality_filter = {
# 'vcf_filter': 'pass',
'min_gq': GQ_threshold,
'min_ab': AB_threshold,
}
# run MendelianVariantSearch for each family, collect results
families = project.get_families()
for i, family in enumerate(families):
print("Processing %s - family %s (%d / %d)" %
(inheritance_mode, family.family_id, i + 1, len(families)))
try:
if inheritance_mode == "all_variants":
yield family, list(
get_variants(get_datastore(project.project_id),
family.xfamily(),
variant_filter=variant_filter,
quality_filter=quality_filter,
indivs_to_consider=family.indiv_id_list()))
else:
yield family, list(
get_variants_with_inheritance_mode(
get_mall(project.project_id),
family.xfamily(),
inheritance_mode,
variant_filter=variant_filter,
quality_filter=quality_filter,
))
except ValueError as e:
print("Error: %s. Skipping family %s" % (str(e), str(family)))
def get_variants_for_inheritance_for_project(project, inheritance_mode):
"""
Get the variants for this project / inheritance combo
Return dict of family -> list of variants
"""
# create search specification
# this could theoretically differ by project, if there are different reference populations
#variant_filter = VariantFilter(so_annotations=SO_SEVERITY_ORDER, ref_freqs=[])
variant_filter = get_default_variant_filter('moderate_impact')
variant_filter.ref_freqs.append(('1kg_wgs_phase3', g1k_freq_threshold))
variant_filter.ref_freqs.append(('1kg_wgs_phase3_popmax', g1k_popmax_freq_threshold))
variant_filter.ref_freqs.append(('exac_v3', exac_freq_threshold))
variant_filter.ref_freqs.append(('exac_v3_popmax', exac_popmax_threshold))
variant_filter.ref_freqs.append(('merck-wgs-3793', merck_wgs_3793_threshold))
#variant_filter.ref_freqs.append(('merck-pcr-free-wgs-144', merck_wgs_144_threshold))
quality_filter = {
# 'vcf_filter': 'pass',
'min_gq': GQ_threshold,
'min_ab': AB_threshold,
}
# run MendelianVariantSearch for each family, collect results
families = project.get_families()
for i, family in enumerate(families):
print("Processing %s - family %s (%d / %d)" % (inheritance_mode, family.family_id, i+1, len(families)))
try:
if inheritance_mode == "all_variants":
yield family, list(get_variants(
get_datastore(project.project_id),
family.xfamily(),
variant_filter=variant_filter,
quality_filter=quality_filter,
indivs_to_consider=family.indiv_id_list()
))
else:
yield family, list(get_variants_with_inheritance_mode(
get_mall(project.project_id),
family.xfamily(),
inheritance_mode,
variant_filter=variant_filter,
quality_filter=quality_filter,
))
except ValueError as e:
print("Error: %s. Skipping family %s" % (str(e), str(family)))
def get_families_by_gene(mall, family_group, inheritance_mode, variant_filter=None, quality_filter=None):
families_by_gene = defaultdict(set)
for family in family_group.get_families():
for variant in get_variants_with_inheritance_mode(
mall,
family,
inheritance_mode,
variant_filter,
quality_filter
):
for gene_id in variant.coding_gene_ids:
families_by_gene[gene_id].add((family.project_id, family.family_id))
for gene_id, family_set in families_by_gene.items():
yield gene_id, sorted(list(family_set))
def get_variants_by_family_for_gene(mall, family_list, inheritance_mode, gene_id, variant_filter=None, quality_filter=None):
if variant_filter is None:
variant_filter = VariantFilter()
variant_filter.add_gene(gene_id)
by_family = {}
for family in family_list:
family_t = (family.project_id, family.family_id)
variants = list(get_variants_with_inheritance_mode(
mall,
family,
inheritance_mode,
variant_filter,
quality_filter
))
by_family[family_t] = variants
return by_family
def combine_mendelian_families_spec(request):
project, family_group = utils.get_project_and_family_group_for_user(request.user, request.GET)
if not project.can_view(request.user):
raise PermissionDenied
search_hash = request.GET.get('search_hash')
search_spec, genes = cache_utils.get_cached_results(project.project_id, search_hash)
search_spec_obj = MendelianVariantSearchSpec.fromJSON(search_spec)
if request.GET.get('return_type') != 'csv' or not request.GET.get('group_by_variants'):
if genes is None:
genes = api_utils.calculate_combine_mendelian_families(family_group, search_spec)
api_utils.add_extra_info_to_genes(project, get_reference(), genes)
if request.GET.get('return_type') != 'csv':
return JSONResponse({
'is_error': False,
'genes': genes,
'search_spec': search_spec,
})
else:
response = HttpResponse(content_type='text/csv')
response['Content-Disposition'] = 'attachment; filename="family_group_results_{}.csv"'.format(search_hash)
writer = csv.writer(response)
writer.writerow(["gene", "# families", "family list", "chrom", "start", "end"])
for gene in genes:
family_id_list = [family_id for (project_id, family_id) in gene["family_id_list"]]
writer.writerow(map(str, [gene["gene_name"], len(family_id_list), " ".join(family_id_list), gene["chr"], gene["start"], gene["end"], ""]))
return response
else:
# download results grouped by variant
indiv_id_list = []
for family in family_group.get_families():
indiv_id_list.extend(family.indiv_ids_with_variant_data())
response = HttpResponse(content_type='text/csv')
response['Content-Disposition'] = 'attachment; filename="results_{}.csv"'.format(search_hash)
writer = csv.writer(response)
headers = ['genes','chr','pos','ref','alt','worst_annotation' ]
headers.extend(project.get_reference_population_slugs())
headers.extend([ 'polyphen','sift','muttaster','fathmm'])
for indiv_id in indiv_id_list:
headers.append(indiv_id)
headers.append(indiv_id+'_gq')
headers.append(indiv_id+'_dp')
writer.writerow(headers)
mall = get_mall(project.project_id)
variant_key_to_individual_id_to_variant = defaultdict(dict)
variant_key_to_variant = {}
for family in family_group.get_families():
for variant in get_variants_with_inheritance_mode(
mall,
family.xfamily(),
search_spec_obj.inheritance_mode,
search_spec_obj.variant_filter,
search_spec_obj.quality_filter,
):
if len(variant.coding_gene_ids) == 0:
continue
variant_key = (variant.xpos, variant.ref, variant.alt)
variant_key_to_variant[variant_key] = variant
for indiv_id in family.indiv_ids_with_variant_data():
variant_key_to_individual_id_to_variant[variant_key][indiv_id] = variant
for variant_key in sorted(variant_key_to_individual_id_to_variant.keys()):
variant = variant_key_to_variant[variant_key]
individual_id_to_variant = variant_key_to_individual_id_to_variant[variant_key]
genes = [mall.reference.get_gene_symbol(gene_id) for gene_id in variant.coding_gene_ids]
fields = []
fields.append(','.join(genes))
fields.extend([
variant.chr,
str(variant.pos),
variant.ref,
variant.alt,
variant.annotation.get('vep_group', '.'),
])
for ref_population_slug in project.get_reference_population_slugs():
fields.append(variant.annotation['freqs'][ref_population_slug])
for field_key in ['polyphen', 'sift', 'muttaster', 'fathmm']:
fields.append(variant.annotation[field_key])
for indiv_id in indiv_id_list:
variant = individual_id_to_variant.get(indiv_id)
genotype = None
if variant is not None:
genotype = variant.get_genotype(indiv_id)
if genotype is None:
fields.extend(['.', '.', '.'])
else:
if genotype.num_alt == 0:
fields.append("%s/%s" % (variant.ref, variant.ref))
elif genotype.num_alt == 1:
fields.append("%s/%s" % (variant.ref, variant.alt))
elif genotype.num_alt == 2:
fields.append("%s/%s" % (variant.alt, variant.alt))
else:
fields.append("./.")
fields.append(str(genotype.gq) if genotype.gq is not None else '.')
fields.append(genotype.extras['dp'] if genotype.extras.get('dp') is not None else '.')
writer.writerow(fields)
return response
def calculate_mendelian_variant_search(search_spec, family, user=None):
xfamily = family.xfamily()
project = family.project
variants = None
if search_spec.search_mode == 'standard_inheritance':
variants = list(get_variants_with_inheritance_mode(
get_mall(project),
xfamily,
search_spec.inheritance_mode,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
))
elif search_spec.search_mode == 'custom_inheritance':
variants = list(get_variants_family(
get_datastore(project),
xfamily,
genotype_filter=search_spec.genotype_inheritance_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
))
elif search_spec.search_mode == 'gene_burden':
gene_stream = get_genes_family(
get_datastore(project),
get_reference(),
xfamily,
burden_filter=search_spec.gene_burden_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
)
variants = list(stream_utils.gene_stream_to_variant_stream(gene_stream, get_reference()))
elif search_spec.search_mode == 'allele_count':
variants = list(get_variants_allele_count(
get_datastore(project),
xfamily,
search_spec.allele_count_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
))
elif search_spec.search_mode == 'all_variants':
variants = list(get_variants_family(
get_datastore(project),
xfamily,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
indivs_to_consider=xfamily.indiv_id_list(),
user=user,
))
for variant in variants:
variant.set_extra('family_id', family.family_id)
return variants
def calculate_mendelian_variant_search(search_spec, xfamily):
sys.stderr.write((
"mendelian_variant_search for %s - search mode: %s \n"
"variant_filter: %s \ninheritance_mode: %s \nallele_count_filter: %s \nquality_filter: %s \ngenotype_inheritance_filter: %s \n"
) % (xfamily.project_id, search_spec.search_mode,
search_spec.variant_filter.toJSON() if search_spec.variant_filter else
'', search_spec.inheritance_mode, search_spec.allele_count_filter,
search_spec.quality_filter, search_spec.genotype_inheritance_filter))
variants = None
if search_spec.search_mode == 'standard_inheritance':
variants = list(
get_variants_with_inheritance_mode(
get_mall(xfamily.project_id),
xfamily,
search_spec.inheritance_mode,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
))
elif search_spec.search_mode == 'custom_inheritance':
variants = list(
get_variants_family(
get_datastore(xfamily.project_id),
xfamily,
genotype_filter=search_spec.genotype_inheritance_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
))
elif search_spec.search_mode == 'gene_burden':
gene_stream = get_genes_family(
get_datastore(xfamily.project_id),
get_reference(),
xfamily,
burden_filter=search_spec.gene_burden_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
)
variants = list(
stream_utils.gene_stream_to_variant_stream(gene_stream,
get_reference()))
elif search_spec.search_mode == 'allele_count':
variants = list(
get_variants_allele_count(
get_datastore(xfamily.project_id),
xfamily,
search_spec.allele_count_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
))
elif search_spec.search_mode == 'all_variants':
variants = list(
get_variants_family(
get_datastore(xfamily.project_id),
xfamily,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
indivs_to_consider=xfamily.indiv_id_list(),
))
return variants
def calculate_mendelian_variant_search(search_spec, xfamily):
sys.stderr.write(" cohort_variant_search - inheritance_mode: %s" %
search_spec.inheritance_mode)
variants = None
if search_spec.search_mode == 'standard_inheritance':
variants = list(
get_variants_with_inheritance_mode(
get_mall(xfamily.project_id),
xfamily,
search_spec.inheritance_mode,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
elif search_spec.search_mode == 'custom_inheritance':
variants = list(
get_variants_family(
get_datastore(xfamily.project_id),
xfamily,
genotype_filter=search_spec.genotype_inheritance_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
elif search_spec.search_mode == 'gene_burden':
gene_stream = get_genes_family(
get_datastore(xfamily.project_id),
get_reference(),
xfamily,
burden_filter=search_spec.gene_burden_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
)
variants = list(
stream_utils.gene_stream_to_variant_stream(gene_stream,
get_reference()))
elif search_spec.search_mode == 'allele_count':
variants = list(
get_variants_allele_count(
get_datastore(xfamily.project_id),
xfamily,
search_spec.allele_count_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
elif search_spec.search_mode == 'all_variants':
variants = list(
get_variants_family(
get_datastore(xfamily.project_id),
xfamily,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.genotype_quality_filter,
))
return variants
def calculate_mendelian_variant_search(search_spec, family, user=None):
xfamily = family.xfamily()
project = family.project
variants = None
if search_spec.search_mode == 'standard_inheritance':
variants = list(
get_variants_with_inheritance_mode(
get_mall(project),
xfamily,
search_spec.inheritance_mode,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
))
elif search_spec.search_mode == 'custom_inheritance':
variants = list(
get_variants_family(
get_datastore(project),
xfamily,
genotype_filter=search_spec.genotype_inheritance_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
))
elif search_spec.search_mode == 'gene_burden':
gene_stream = get_genes_family(
get_datastore(project),
get_reference(),
xfamily,
burden_filter=search_spec.gene_burden_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
)
variants = list(
stream_utils.gene_stream_to_variant_stream(gene_stream,
get_reference()))
elif search_spec.search_mode == 'allele_count':
variants = list(
get_variants_allele_count(
get_datastore(project),
xfamily,
search_spec.allele_count_filter,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
user=user,
))
elif search_spec.search_mode == 'all_variants':
variants = list(
get_variants_family(
get_datastore(project),
xfamily,
variant_filter=search_spec.variant_filter,
quality_filter=search_spec.quality_filter,
indivs_to_consider=xfamily.indiv_id_list(),
user=user,
))
for variant in variants:
variant.set_extra('family_id', family.family_id)
return variants
def handle(self, *args, **options):
project_id = args[0]
inheritance_mode = args[1]
fam_list_file_path = args[2]
project = Project.objects.get(project_id=project_id)
families = []
for line in open(fam_list_file_path):
family_id = line.strip('\n')
families.append(Family.objects.get(project=project, family_id=family_id))
# create search spec
variant_filter = next(f for f in project.get_default_variant_filters() if f['slug'] == 'moderate_impact')['variant_filter']
quality_filter = {
'min_gq': 20,
'min_ab': 25,
}
# run MendelianVariantSearch for each family, collect results
family_results = {}
for family in families:
family_results[family] = list(get_variants_with_inheritance_mode(
get_mall(project_id),
family.xfamily(),
inheritance_mode,
variant_filter=variant_filter,
quality_filter=quality_filter,
))
# create family_variants.tsv
f = open('family_variants.tsv', 'w')
writer = csv.writer(f, dialect='excel', delimiter='\t')
writer.writerow([
'#family_id',
'gene',
'chrom',
'ref',
'alt',
'rsid',
'annotation',
])
for family in families:
for variant in family_results[family]:
writer.writerow([
family.family_id,
get_gene_symbol(variant),
variant.chr,
variant.ref,
variant.alt,
variant.vcf_id,
variant.annotation['vep_group'],
])
f.close()
# create variants.tsv
by_variant = {}
variant_info = {}
for family in families:
for variant in family_results[family]:
if variant.unique_tuple() not in by_variant:
by_variant[variant.unique_tuple()] = set()
variant_info[variant.unique_tuple()] = variant
by_variant[variant.unique_tuple()].add(family.family_id)
f = open('variants.tsv', 'w')
writer = csv.writer(f, dialect='excel', delimiter='\t')
headers = [
'#chrom',
'ref',
'alt',
'rsid',
'gene'
'annotation',
'num_families',
]
headers.extend([fam.family_id for fam in families])
writer.writerow(headers)
for variant_t in sorted(variant_info.keys()):
variant = variant_info[variant_t]
fields = [
variant.chr,
variant.ref,
variant.alt,
variant.vcf_id,
get_gene_symbol(variant_info[variant_t]),
variant.annotation['vep_group'],
str(len(by_variant[variant_t])),
]
for family in families:
fields.append('1' if family.family_id in by_variant[variant_t] else '0')
writer.writerow(fields)
f.close()
# create genes.tsv
by_gene = {}
for family in families:
for variant in family_results[family]:
gene_symbol = get_gene_symbol(variant)
if gene_symbol not in by_gene:
by_gene[gene_symbol] = set()
by_gene[gene_symbol].add(family.family_id)
f = open('genes.tsv', 'w')
writer = csv.writer(f, dialect='excel', delimiter='\t')
headers = [
'#gene',
'num_families',
]
headers.extend([fam.family_id for fam in families])
writer.writerow(headers)
for gene_symbol in sorted(by_gene.keys()):
fields = [
gene_symbol,
str(len(by_gene[gene_symbol])),
]
for family in families:
fields.append('1' if family.family_id in by_gene[gene_symbol] else '0')
writer.writerow(fields)
f.close()
def handle(self, *args, **options):
project_id = args[0]
inheritance_mode = args[1]
fam_list_file_path = args[2]
project = Project.objects.get(project_id=project_id)
families = []
for line in open(fam_list_file_path):
family_id = line.strip('\n')
families.append(
Family.objects.get(project=project, family_id=family_id))
# create search spec
variant_filter = next(
f for f in project.get_default_variant_filters()
if f['slug'] == 'moderate_impact')['variant_filter']
quality_filter = {
'min_gq': 30,
'min_ab': 25,
}
# run MendelianVariantSearch for each family, collect results
family_results = {}
for family in families:
family_results[family] = list(
get_variants_with_inheritance_mode(
get_mall(project_id),
family.xfamily(),
inheritance_mode,
variant_filter=variant_filter,
quality_filter=quality_filter,
))
# create family_variants.tsv
f = open('family_variants.tsv', 'w')
writer = csv.writer(f, dialect='excel', delimiter='\t')
writer.writerow([
'#family_id',
'gene',
'chrom',
'ref',
'alt',
'rsid',
'annotation',
])
for family in families:
for variant in family_results[family]:
writer.writerow([
family.family_id,
get_gene_symbol(variant),
variant.chr,
variant.ref,
variant.alt,
variant.vcf_id,
variant.annotation['vep_group'],
])
f.close()
# create variants.tsv
by_variant = {}
variant_info = {}
for family in families:
for variant in family_results[family]:
if variant.unique_tuple() not in by_variant:
by_variant[variant.unique_tuple()] = set()
variant_info[variant.unique_tuple()] = variant
by_variant[variant.unique_tuple()].add(family.family_id)
f = open('variants.tsv', 'w')
writer = csv.writer(f, dialect='excel', delimiter='\t')
headers = [
'#chrom',
'ref',
'alt',
'rsid',
'gene'
'annotation',
'num_families',
]
headers.extend([fam.family_id for fam in families])
writer.writerow(headers)
for variant_t in sorted(variant_info.keys()):
variant = variant_info[variant_t]
fields = [
variant.chr,
variant.ref,
variant.alt,
variant.vcf_id,
get_gene_symbol(variant_info[variant_t]),
variant.annotation['vep_group'],
str(len(by_variant[variant_t])),
]
for family in families:
fields.append('1' if family.family_id in
by_variant[variant_t] else '0')
writer.writerow(fields)
f.close()
# create genes.tsv
by_gene = {}
for family in families:
for variant in family_results[family]:
gene_symbol = get_gene_symbol(variant)
if gene_symbol not in by_gene:
by_gene[gene_symbol] = set()
by_gene[gene_symbol].add(family.family_id)
f = open('genes.tsv', 'w')
writer = csv.writer(f, dialect='excel', delimiter='\t')
headers = [
'#gene',
'num_families',
]
headers.extend([fam.family_id for fam in families])
writer.writerow(headers)
for gene_symbol in sorted(by_gene.keys()):
fields = [
gene_symbol,
str(len(by_gene[gene_symbol])),
]
for family in families:
fields.append('1' if family.family_id in
by_gene[gene_symbol] else '0')
writer.writerow(fields)
f.close()
def combine_mendelian_families_spec(request):
project, family_group = utils.get_project_and_family_group_for_user(request.user, request.GET)
if not project.can_view(request.user):
raise PermissionDenied
search_hash = request.GET.get('search_hash')
search_spec, genes = cache_utils.get_cached_results(project.project_id, search_hash)
search_spec_obj = MendelianVariantSearchSpec.fromJSON(search_spec)
if request.GET.get('return_type') != 'csv' or not request.GET.get('group_by_variants'):
if genes is None:
genes = api_utils.calculate_combine_mendelian_families(family_group, search_spec)
api_utils.add_extra_info_to_genes(project, get_reference(), genes)
if request.GET.get('return_type') != 'csv':
return JSONResponse({
'is_error': False,
'genes': genes,
'search_spec': search_spec,
})
else:
response = HttpResponse(content_type='text/csv')
response['Content-Disposition'] = 'attachment; filename="family_group_results_{}.csv"'.format(search_hash)
writer = csv.writer(response)
writer.writerow(["gene", "# families", "family list", "chrom", "start", "end"])
for gene in genes:
family_id_list = [family_id for (project_id, family_id) in gene["family_id_list"]]
writer.writerow(map(str, [gene["gene_name"], len(family_id_list), " ".join(family_id_list), gene["chr"], gene["start"], gene["end"], ""]))
return response
else:
# download results grouped by variant
indiv_id_list = []
for family in family_group.get_families():
indiv_id_list.extend(family.indiv_ids_with_variant_data())
response = HttpResponse(content_type='text/csv')
response['Content-Disposition'] = 'attachment; filename="results_{}.csv"'.format(search_hash)
writer = csv.writer(response)
headers = ['genes','chr','pos','ref','alt','worst_annotation' ]
headers.extend(project.get_reference_population_slugs())
headers.extend([ 'polyphen','sift','muttaster','fathmm'])
for indiv_id in indiv_id_list:
headers.append(indiv_id)
headers.append(indiv_id+'_gq')
headers.append(indiv_id+'_dp')
writer.writerow(headers)
mall = get_mall(project.project_id)
variant_key_to_individual_id_to_variant = defaultdict(dict)
variant_key_to_variant = {}
for family in family_group.get_families():
for variant in get_variants_with_inheritance_mode(
mall,
family.xfamily(),
search_spec_obj.inheritance_mode,
search_spec_obj.variant_filter,
search_spec_obj.quality_filter,
):
if len(variant.coding_gene_ids) == 0:
continue
variant_key = (variant.xpos, variant.ref, variant.alt)
variant_key_to_variant[variant_key] = variant
for indiv_id in family.indiv_ids_with_variant_data():
variant_key_to_individual_id_to_variant[variant_key][indiv_id] = variant
for variant_key in sorted(variant_key_to_individual_id_to_variant.keys()):
variant = variant_key_to_variant[variant_key]
individual_id_to_variant = variant_key_to_individual_id_to_variant[variant_key]
genes = [mall.reference.get_gene_symbol(gene_id) for gene_id in variant.coding_gene_ids]
fields = []
fields.append(','.join(genes))
fields.extend([
variant.chr,
str(variant.pos),
variant.ref,
variant.alt,
variant.annotation.get('vep_group', '.'),
])
for ref_population_slug in project.get_reference_population_slugs():
fields.append(variant.annotation['freqs'][ref_population_slug])
for field_key in ['polyphen', 'sift', 'muttaster', 'fathmm']:
fields.append(variant.annotation[field_key])
for indiv_id in indiv_id_list:
variant = individual_id_to_variant.get(indiv_id)
genotype = None
if variant is not None:
genotype = variant.get_genotype(indiv_id)
if genotype is None:
fields.extend(['.', '.', '.'])
else:
if genotype.num_alt == 0:
fields.append("%s/%s" % (variant.ref, variant.ref))
elif genotype.num_alt == 1:
fields.append("%s/%s" % (variant.ref, variant.alt))
elif genotype.num_alt == 2:
fields.append("%s/%s" % (variant.alt, variant.alt))
else:
fields.append("./.")
fields.append(str(genotype.gq) if genotype.gq is not None else '.')
fields.append(genotype.extras['dp'] if genotype.extras.get('dp') is not None else '.')
writer.writerow(fields)
return response
