def add_vcf_file_to_annotator(self, vcf_file_path, force_all=False):
"""
Add the variants in vcf_file_path to annotator
Convenience wrapper around add_variants_to_annotator
"""
if not force_all and self.get_annotator_datastore().vcf_files.find_one(
{'vcf_file_path': vcf_file_path}):
print "VCF already annotated"
return
print "Scanning VCF file first..."
variant_t_list = []
for variant_t in vcf_stuff.iterate_tuples(
compressed_file(vcf_file_path)):
variant_t_list.append(variant_t)
if len(variant_t_list) == 100000:
print "Adding another 100000 variants, through {}".format(
variant_t_list[-1][0])
self.add_variants_to_annotator(variant_t_list, force_all)
variant_t_list = []
self.add_variants_to_annotator(variant_t_list, force_all)
self.get_annotator_datastore().vcf_files.insert({
'vcf_file_path':
vcf_file_path,
'date_added':
datetime.datetime.utcnow()
})
def add_vcf_file_to_annotator(self, vcf_file_path, force_all=False):
"""
Add the variants in vcf_file_path to annotator
Convenience wrapper around add_variants_to_annotator
"""
print "Scanning VCF file first..."
variant_t_list = []
for variant_t in vcf_stuff.iterate_tuples(compressed_file(vcf_file_path)):
variant_t_list.append(variant_t)
if len(variant_t_list) == 100000:
print "Adding another 100000 variants, through {}".format(variant_t_list[-1][0])
self.add_variants_to_annotator(variant_t_list, force_all)
variant_t_list = []
self.add_variants_to_annotator(variant_t_list, force_all)
def add_vcf_file_to_annotator(self, vcf_file_path, force_all=False):
"""
Add the variants in vcf_file_path to annotator
Convenience wrapper around add_variants_to_annotator
"""
if not force_all and self._db.vcf_files.find_one({'vcf_file_path': vcf_file_path}):
print "VCF already annotated"
return
print "Scanning VCF file first..."
variant_t_list = []
for variant_t in vcf_stuff.iterate_tuples(compressed_file(vcf_file_path)):
variant_t_list.append(variant_t)
if len(variant_t_list) == 100000:
print "Adding another 100000 variants, through {}".format(variant_t_list[-1][0])
self.add_variants_to_annotator(variant_t_list, force_all)
variant_t_list = []
self.add_variants_to_annotator(variant_t_list, force_all)
self._db.vcf_files.insert({'vcf_file_path': vcf_file_path, 'date_added': datetime.datetime.utcnow()})
def _add_vcf_file_for_family_set(self, family_info_list, vcf_file_path, reference_populations=None, vcf_id_map=None):
collections = {f['family_id']: self._db[f['coll_name']] for f in family_info_list}
for collection in collections.values():
collection.drop_indexes()
indiv_id_list = [i for f in family_info_list for i in f['individuals']]
vcf_file = compressed_file(vcf_file_path)
size = os.path.getsize(vcf_file_path)
progress = get_progressbar(size, 'Loading VCF: {}'.format(vcf_file_path))
for variant in vcf_stuff.iterate_vcf(vcf_file, genotypes=True, indiv_id_list=indiv_id_list, vcf_id_map=vcf_id_map):
progress.update(vcf_file.tell_progress())
annotation = self._annotator.get_annotation(variant.xpos, variant.ref, variant.alt, populations=reference_populations)
for family in family_info_list:
# TODO: can we move this inside the if relevant clause below?
family_variant = variant.make_copy(restrict_to_genotypes=family['individuals'])
family_variant_dict = family_variant.toJSON()
_add_index_fields_to_variant(family_variant_dict, annotation)
if xbrowse_utils.is_variant_relevant_for_individuals(family_variant, family['individuals']):
collection = collections[family['family_id']]
collection.insert(family_variant_dict)
def _add_vcf_file_for_family_set(self,
family_info_list,
vcf_file_path,
reference_populations=None,
vcf_id_map=None,
start_from_chrom=None,
end_with_chrom=None):
collections = {
f['family_id']: self._db[f['coll_name']]
for f in family_info_list
}
#for collection in collections.values():
# collection.drop_indexes()
indiv_id_list = [i for f in family_info_list for i in f['individuals']]
number_of_families = len(family_info_list)
sys.stderr.write(
"Loading variants for %(number_of_families)d families %(family_info_list)s from %(vcf_file_path)s\n"
% locals())
for family in family_info_list:
print("Indexing family: " + str(family))
collection = collections[family['family_id']]
collection.ensure_index([('xpos', 1), ('ref', 1), ('alt', 1)])
# check whether some of the variants for this chromosome has been loaded already
# if yes, start from the last loaded variant, and not from the beginning
if "_chr" in vcf_file_path or ".chr" in vcf_file_path:
# if the VCF files are split by chromosome (eg. for WGS projects), check within the chromosome
vcf_file = compressed_file(vcf_file_path)
variant = next(
vcf_stuff.iterate_vcf(vcf_file,
genotypes=False,
indiv_id_list=indiv_id_list,
vcf_id_map=vcf_id_map))
print(vcf_file_path + " - chromsome: " + str(variant.chr))
vcf_file.close()
position_per_chrom = {}
for chrom in range(1, 24):
position_per_chrom[chrom] = defaultdict(int)
for family in family_info_list: #variants = collections[family['family_id']].find().sort([('xpos',-1)]).limit(1)
variants = list(collections[family['family_id']].find({
'$and': [{
'xpos': {
'$gte': chrom * 1e9
}
}, {
'xpos': {
'$lt': (chrom + 1) * 1e9
}
}]
}).sort([('xpos', -1)]).limit(1))
if len(variants) > 0:
position_per_chrom[chrom][family[
'family_id']] = variants[0]['xpos'] - chrom * 1e9
else:
position_per_chrom[chrom][family['family_id']] = 0
for chrom in range(1, 24):
position_per_chrom[chrom] = min(
position_per_chrom[chrom].values()
) # get the smallest last-loaded variant position for this chromosome across all families
chr_idx = int(variant.xpos / 1e9)
start_from_pos = int(position_per_chrom[chr_idx])
print("Start from: %s - %s (%0.1f%% done)" %
(chr_idx, start_from_pos, 100. * start_from_pos /
CHROMOSOME_SIZES[variant.chr.replace("chr", "")]))
tabix_file = pysam.TabixFile(vcf_file_path)
vcf_iter = itertools.chain(
tabix_file.header,
tabix_file.fetch(variant.chr.replace("chr", ""),
start_from_pos, int(2.5e8)))
elif start_from_chrom or end_with_chrom:
if start_from_chrom:
print("Start chrom: chr%s" % start_from_chrom)
if end_with_chrom:
print("End chrom: chr%s" % end_with_chrom)
chrom_list = list(map(str, range(1, 23))) + ['X', 'Y']
chrom_list_start_index = 0
if start_from_chrom:
chrom_list_start_index = chrom_list.index(
start_from_chrom.replace("chr", "").upper())
chrom_list_end_index = len(chrom_list)
if end_with_chrom:
chrom_list_end_index = chrom_list.index(
end_with_chrom.replace("chr", "").upper())
tabix_file = pysam.TabixFile(vcf_file_path)
vcf_iter = tabix_file.header
for chrom in chrom_list[
chrom_list_start_index:chrom_list_end_index + 1]:
print("Will load chrom: " + chrom)
try:
vcf_iter = itertools.chain(vcf_iter,
tabix_file.fetch(chrom))
except ValueError as e:
print("WARNING: " + str(e))
else:
vcf_iter = vcf_file = compressed_file(vcf_file_path)
# TODO handle case where it's one vcf file, not split by chromosome
size = os.path.getsize(vcf_file_path)
#progress = get_progressbar(size, 'Loading VCF: {}'.format(vcf_file_path))
def insert_all_variants_in_buffer(buff, collections_dict):
for family_id in buff:
if len(buff[family_id]) == 0: # defensive programming
raise ValueError(
"%s has zero variants to insert. Should not be in buff."
% family_id)
while len(buff) > 0:
# choose a random family for which to insert a variant from among families that still have variants to insert
family_id = random.choice(buff.keys())
# pop a variant off the list for this family, and insert it
family_variant_dict_to_insert = buff[family_id].pop()
c = collections_dict[family_id]
c.insert(family_variant_dict_to_insert)
if len(buff[family_id]) == 0:
del buff[
family_id] # if no more variants for this family, delete it
vcf_rows_counter = 0
variants_buffered_counter = 0
family_id_to_variant_list = defaultdict(
list) # will accumulate variants to be inserted all at once
for variant in vcf_stuff.iterate_vcf(vcf_iter,
genotypes=True,
indiv_id_list=indiv_id_list,
vcf_id_map=vcf_id_map):
if variant.alt == "*":
#print("Skipping GATK 3.4 * alt allele: " + str(variant.unique_tuple()))
continue
try:
annotation = self._annotator.get_annotation(
variant.xpos,
variant.ref,
variant.alt,
populations=reference_populations)
except ValueError, e:
sys.stderr.write("WARNING: " + str(e) + "\n")
continue
vcf_rows_counter += 1
for family in family_info_list:
# TODO: can we move this inside the if relevant clause below?
try:
family_variant = variant.make_copy(
restrict_to_genotypes=family['individuals'])
family_variant_dict = family_variant.toJSON()
_add_index_fields_to_variant(family_variant_dict,
annotation)
if xbrowse_utils.is_variant_relevant_for_individuals(
family_variant, family['individuals']):
collection = collections[family['family_id']]
if not collection.find_one({
'xpos': family_variant.xpos,
'ref': family_variant.ref,
'alt': family_variant.alt
}):
family_id_to_variant_list[family[
'family_id']].append(family_variant_dict)
variants_buffered_counter += 1
except Exception, e:
sys.stderr.write(
"ERROR: on variant %s, family: %s - %s\n" %
(variant.toJSON(), family, e))
import sys
from xbrowse.parsers.vcf_stuff import iterate_vcf
from xbrowse.utils import get_aaf, compressed_file
if __name__ == '__main__':
vcf_file = compressed_file(sys.argv[1])
for variant in iterate_vcf(vcf_file, genotypes=True):
print '\t'.join([
str(variant.xpos),
variant.ref,
variant.alt,
str(get_aaf(variant)),
])
def _add_vcf_file_for_family_set(self, family_info_list, vcf_file_path, reference_populations=None, vcf_id_map=None):
collections = {f['family_id']: self._db[f['coll_name']] for f in family_info_list}
#for collection in collections.values():
# collection.drop_indexes()
indiv_id_list = [i for f in family_info_list for i in f['individuals']]
number_of_families = len(family_info_list)
sys.stderr.write("Loading variants for %(number_of_families)d families %(family_info_list)s from %(vcf_file_path)s\n" % locals())
for family in family_info_list:
print("Indexing family: " + str(family))
collection = collections[family['family_id']]
collection.ensure_index([('xpos', 1), ('ref', 1), ('alt', 1)])
# check whether some of the variants for this chromosome has been loaded already
# if yes, start from the last loaded variant, and not from the beginning
if "_chr" in vcf_file_path or ".chr" in vcf_file_path:
# if the VCF files are split by chromosome (eg. for WGS projects), check within the chromosome
vcf_file = compressed_file(vcf_file_path)
variant = next(vcf_stuff.iterate_vcf(vcf_file, genotypes=False, indiv_id_list=indiv_id_list, vcf_id_map=vcf_id_map))
print(vcf_file_path + " - chromsome: " + str(variant.chr))
vcf_file.close()
position_per_chrom = {}
for chrom in range(1,24):
position_per_chrom[chrom] = defaultdict(int)
for family in family_info_list: #variants = collections[family['family_id']].find().sort([('xpos',-1)]).limit(1)
variants = list(collections[family['family_id']].find({'$and': [{'xpos': { '$gte': chrom*1e9 }}, {'xpos': { '$lt': (chrom+1)*1e9}}] }).sort([('xpos',-1)]).limit(1))
if len(variants) > 0:
position_per_chrom[chrom][family['family_id']] = variants[0]['xpos'] - chrom*1e9
else:
position_per_chrom[chrom][family['family_id']] = 0
for chrom in range(1,24):
position_per_chrom[chrom] = min(position_per_chrom[chrom].values()) # get the smallest last-loaded variant position for this chromosome across all families
chr_idx = int(variant.xpos/1e9)
start_from_pos = int(position_per_chrom[chr_idx])
print("Start from: %s - %s (%0.1f%% done)" % (chr_idx, start_from_pos, 100.*start_from_pos/CHROMOSOME_SIZES[variant.chr.replace("chr", "")]))
tabix_file = pysam.TabixFile(vcf_file_path)
vcf_iter = itertools.chain(tabix_file.header, tabix_file.fetch(variant.chr.replace("chr", ""), start_from_pos, int(2.5e8)))
else:
vcf_iter = vcf_file = compressed_file(vcf_file_path)
# TODO handle case where it's one vcf file, not split by chromosome
size = os.path.getsize(vcf_file_path)
progress = get_progressbar(size, 'Loading VCF: {}'.format(vcf_file_path))
def insert_all_variants_in_buffer(buff, collections_dict):
for family_id in buff:
if len(buff[family_id]) == 0: # defensive programming
raise ValueError("%s has zero variants to insert. Should not be in buff." % family_id)
while len(buff) > 0:
# choose a random family for which to insert a variant from among families that still have variants to insert
family_id = random.choice(buff.keys())
# pop a variant off the list for this family, and insert it
family_variant_dict_to_insert = buff[family_id].pop()
c = collections_dict[family_id]
c.insert(family_variant_dict_to_insert)
if len(buff[family_id]) == 0:
del buff[family_id] # if no more variants for this family, delete it
vcf_rows_counter = 0
variants_buffered_counter = 0
family_id_to_variant_list = defaultdict(list) # will accumulate variants to be inserted all at once
for variant in vcf_stuff.iterate_vcf(vcf_iter, genotypes=True, indiv_id_list=indiv_id_list, vcf_id_map=vcf_id_map):
if variant.alt == "*":
#print("Skipping GATK 3.4 * alt allele: " + str(variant.unique_tuple()))
continue
try:
annotation = self._annotator.get_annotation(variant.xpos, variant.ref, variant.alt, populations=reference_populations)
except ValueError, e:
sys.stderr.write("WARNING: " + str(e) + "\n")
continue
vcf_rows_counter += 1
for family in family_info_list:
# TODO: can we move this inside the if relevant clause below?
try:
family_variant = variant.make_copy(restrict_to_genotypes=family['individuals'])
family_variant_dict = family_variant.toJSON()
_add_index_fields_to_variant(family_variant_dict, annotation)
if xbrowse_utils.is_variant_relevant_for_individuals(family_variant, family['individuals']):
collection = collections[family['family_id']]
if not collection.find_one({'xpos': family_variant.xpos, 'ref': family_variant.ref, 'alt': family_variant.alt}):
family_id_to_variant_list[family['family_id']].append(family_variant_dict)
variants_buffered_counter += 1
except Exception, e:
sys.stderr.write("ERROR: on variant %s, family: %s - %s\n" % (variant.toJSON(), family, e))
def _add_vcf_file_for_family_set(self, family_info_list, vcf_file_path, reference_populations=None, vcf_id_map=None):
collections = {f['family_id']: self._db[f['coll_name']] for f in family_info_list}
#for collection in collections.values():
# collection.drop_indexes()
indiv_id_list = [i for f in family_info_list for i in f['individuals']]
number_of_families = len(family_info_list)
sys.stderr.write("Loading variants for %(number_of_families)d families %(family_info_list)s from %(vcf_file_path)s\n" % locals())
#for family in family_info_list:
# print("Indexing family: " + str(family))
# collection = collections[family['family_id']]
# collection.ensure_index([('xpos', 1), ('ref', 1), ('alt', 1)])
vcf_file = compressed_file(vcf_file_path)
size = os.path.getsize(vcf_file_path)
progress = get_progressbar(size, 'Loading VCF: {}'.format(vcf_file_path))
def insert_all_variants_in_buffer(buff, collections_dict):
for family_id in buff:
if len(buff[family_id]) == 0: # defensive programming
raise ValueError("%s has zero variants to insert. Should not be in buff." % family_id)
while len(buff) > 0:
# choose a random family for which to insert a variant from among families that still have variants to insert
family_id = random.choice(buff.keys())
# pop a variant off the list for this family, and insert it
family_variant_dict_to_insert = buff[family_id].pop()
c = collections_dict[family_id]
c.insert(family_variant_dict_to_insert)
if len(buff[family_id]) == 0:
del buff[family_id] # if no more variants for this family, delete it
vcf_rows_counter = 0
variants_buffered_counter = 0
family_id_to_variant_list = defaultdict(list) # will accumulate variants to be inserted all at once
for variant in vcf_stuff.iterate_vcf(vcf_file, genotypes=True, indiv_id_list=indiv_id_list, vcf_id_map=vcf_id_map):
progress.update(vcf_file.tell_progress())
try:
annotation = self._annotator.get_annotation(variant.xpos, variant.ref, variant.alt, populations=reference_populations)
except ValueError, e:
sys.stderr.write("WARNING: " + str(e) + "\n")
continue
vcf_rows_counter += 1
for family in family_info_list:
# TODO: can we move this inside the if relevant clause below?
family_variant = variant.make_copy(restrict_to_genotypes=family['individuals'])
family_variant_dict = family_variant.toJSON()
_add_index_fields_to_variant(family_variant_dict, annotation)
if xbrowse_utils.is_variant_relevant_for_individuals(family_variant, family['individuals']):
collection = collections[family['family_id']]
if not collection.find_one({'xpos': family_variant.xpos, 'ref': family_variant.ref, 'alt': family_variant.alt}):
family_id_to_variant_list[family['family_id']].append(family_variant_dict)
variants_buffered_counter += 1
if variants_buffered_counter > 10000:
print(date.strftime(datetime.now(), "%m/%d/%Y %H:%M:%S") + "-- inserting %d family-variants from %d vcf rows into %s families" % (variants_buffered_counter, vcf_rows_counter, len(family_id_to_variant_list)))
insert_all_variants_in_buffer(family_id_to_variant_list, collections)
assert len(family_id_to_variant_list) == 0
vcf_rows_counter = 0
variants_buffered_counter = 0
import sys
from xbrowse.parsers.vcf_stuff import iterate_vcf
from xbrowse.utils import get_aaf, compressed_file
if __name__ == '__main__':
vcf_file = compressed_file(sys.argv[1])
for variant in iterate_vcf(vcf_file, genotypes=True):
print '\t'.join([
str(variant.xpos),
variant.ref,
variant.alt,
str(get_aaf(variant)),
])