def main(vcf_file):
"""Driver function."""
tmp_filename = os.path.basename(vcf_file)
h5_file = ''.join([
os.path.dirname(vcf_file), '/',
os.path.splitext(tmp_filename)[0], '.h5'
])
# Convert VCF to hdf5
allel.vcf_to_hdf5(vcf_file, h5_file, fields='*', overwrite=True)
def vcf_to_hdf5(vcf_path, zarr_path, fields=None):
if fields is None:
fields = DEF_VCF_FIELDS
# convert our fields to allele zarr fields
zarr_fields = [VARIATION_ZARR_FIELD_MAPPING[field] for field in fields]
if 'samples' not in zarr_fields:
zarr_fields.append('samples')
allel.vcf_to_hdf5(str(vcf_path), str(zarr_path), fields=zarr_fields)
def create_h5_file(vcf_file):
h5_file_path = vcf_file.split(".vcf")[0] + ".h5"
if not check_file_existence(h5_file_path):
allel.vcf_to_hdf5(
vcf_file,
h5_file_path,
fields="*",
exclude_fields=["calldata/GQ"],
overwrite=True,
)
def sim_load_vcf_to_h5(vcf_path, h5_path):
'''
transform data from vcf to h5
input: paths
output: new file on h5_path
'''
#download ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/supporting/hd_genotype_chip/
#Reference: http://alimanfoo.github.io/2016/06/10/scikit-allel-tour.html
#Example: http://alimanfoo.github.io/2017/06/14/read-vcf.html
#vcf_path = 'C://Users//raque//Documents//GitHub//ParKCa//data_s//ALL.chip.omni_broad_sanger_combined.20140818.snps.genotypes.vcf.gz'
#h5_path = 'data_s//ALL.chip.omni_broad_sanger_combined.20140818.snps.genotypes.h5'
allel.vcf_to_hdf5(vcf_path, h5_path, fields='*', overwrite=True)
def loadvcf(vcfFile):
"""Reads VCF using scikit-allel, object is stored as pandas DF
"""
print("loading vcf file...")
print("using scitkit allele version:", allel.__version__)
h5 = "{}h5".format(vcfFile.strip("vcf"))
if os.path.isfile(h5):
callset = h5py.File(h5, 'r')
else:
callset = allel.read_vcf(vcfFile)
print("creating h5 for faster loading")
allel.vcf_to_hdf5(vcfFile, h5)
return(callset)
def loadvcf(vcfFile):
"""Reads VCF using scikit-allel, object is stored as pandas DF
"""
print("loading vcf file...")
print("using scitkit allele version:", allel.__version__)
h5 = "{}h5".format(vcfFile.strip("vcf"))
if os.path.isfile(h5):
callset = h5py.File(h5, 'r')
else:
callset = allel.read_vcf(vcfFile)
print("creating h5 for faster loading")
allel.vcf_to_hdf5(vcfFile, h5)
return (callset)
def makeh5fromvcf(vcfin, chromlist):
"""
"""
for c in chromlist:
h5out = "{}{}.h5".format(vcfin.rstrip("vcf"), c)
fieldsfromvcf = ['samples', 'calldata/GQ', 'variants/ALT',
'variants/REF', 'variants/QUAL', 'variants/CHROM',
'variants/POS', 'variants/AF', 'variants/AB',
'variants/MQM', 'variants/DP', 'calldata/DP',
'calldata/AD', 'calldata/GT']
allel.vcf_to_hdf5(vcfin, h5out, fields=fieldsfromvcf,
types={'calldata/GQ': 'float32'}, region=c)
return(None)
def convertVCFToH5(vcfFileName):
names = vcfFileName.split('.')
h5FileName = names[0] + '.h5'
h5File = Path(h5FileName)
if h5File.is_file():
print(' ')
else:
print('Convertion')
allel.vcf_to_hdf5(
vcfFileName, h5FileName, fields='*', overwrite=True
) # The saved data can be accessed via the h5py library, e.g.:
def vcf_to_hdf5(vcf_path: str, hdf5_path: str):
""" Conversion from VCF or VCF.gz to HDF5 file format.
Parameters:
vcf_path (str): Input path to the VCF file.
hdf5_path (str): Output path to the HDF5 file.
"""
# sample_list = get_samples(vcf_path)
logger.debug("Converting VCF file '{vcf_path}'".format(vcf_path=vcf_path))
if os.path.exists(hdf5_path):
logger.debug("File '{hdf5_path}' already exists. I will remove it.".format(hdf5_path=hdf5_path))
os.remove(hdf5_path)
allel.vcf_to_hdf5(vcf_path, hdf5_path, fields='*', overwrite=False)
logger.debug("HDF5 file stored in '{hdf5_path}'".format(hdf5_path=hdf5_path))
def makeh5fromvcf(vcfin, altnum, hf5):
"""
"""
h5out = "{}.h5".format(vcfin)
if hf5:
pass
else:
fieldsfromvcf = ['samples', 'calldata/GQ', 'variants/ALT',
'variants/REF', 'variants/QUAL', 'variants/CHROM',
'variants/POS', 'variants/AF', 'variants/AB',
'variants/MQM', 'variants/DP', 'calldata/DP',
'calldata/AD', 'calldata/GT']
allel.vcf_to_hdf5(vcfin, h5out, fields=fieldsfromvcf,
types={'calldata/GQ': 'float32'}, alt_number=2)
# callset = h5py.File(h5out, mode='r')
return(None)
def convert_vcf(vcf_filename, h5_filename):
"""Convert vcf_filename"""
# here we save only CHROM, GT (genotypes) and POS (SNP positions)
# see: https://scikit-allel.readthedocs.io/en/stable/io.html
allel.vcf_to_hdf5(vcf_filename, h5_filename, fields=['CHROM', 'GT', 'POS'])
def create_hdf5_file(vcf_path, hdf5_file_name):
print("converting vcf into hdf5 file anc_panel.h5")
allel.vcf_to_hdf5(vcf_path,
hdf5_file_name,
fields=[
'samples', 'calldata/GT', 'calldata/GQ',
'calldata/PL', 'calldata/DP', 'calldata/RR',
'calldata/VR', 'variants/POS', 'variants/CHROM'
],
overwrite=True,
log=sys.stdout)
print(
"performing formatting tests of hdf5 file, is all the data required there?"
)
local_callset = h5py.File(hdf5_file_name, mode='r+')
get_other_dp = 0
list_of_attributes = list(local_callset['calldata'])
if 'DP' not in list_of_attributes:
get_other_dp = 1
if get_other_dp == 0:
dp = local_callset['calldata/DP']
if np.amax(dp[:]) == -1:
get_other_dp = 1
if get_other_dp == 1:
print("Found no DP entries, reconstructing DP from RR and VR\n")
rr = local_callset['calldata/RR']
rr2 = rr[:]
local_rr = rr2[0]
if not np.any(local_rr):
print("It appears that DP information can not be reconstructed\n")
print("Discarding the option to filter for sufficient DP\n")
else:
vr = local_callset['calldata/VR']
new_dp = rr[:] + vr[:]
if np.amax(new_dp[:]) == -1:
print("Failed to reconstruct DP entries from RR and VR\n")
print("Please reformat your VCF to include DP\n")
data = local_callset['calldata/DP']
data[...] = new_dp[:]
gq = local_callset['calldata/GQ']
temp = gq[:]
if np.amax(gq[:] == -1):
print("Found no GQ entries, reconstructing GQ from PL\n")
pl = local_callset['calldata/PL']
new_gq = calc_gq(pl[:])
if np.amax(new_gq[:]) == -1:
print("Failed to reconstruct GQ entries from PL\n")
print("Please reformat your VCF to include GQ")
print("exiting\n")
exit(1)
data = local_callset['calldata/GQ']
data[...] = new_gq[:]
local_callset.close()
print("Data is stored in hdf5 container and verified")
def vcf_to_hdf5(in_path, out_path, path_vcf100g=""):
"""Transform Full VCF to full HDF5"""
allel.vcf_to_hdf5(input=in_path, output=out_path, compression="gzip")
print("Finished Transforming VCF to HDF5.")
