def __init__(self,
vgz,
chm=None,
bp0=None,
bp1=None,
wnd=1024,
dsg='012',
dtp=int):
"""
vgz: name of the bgzipped VCF file (*.vcf.gz), if None,
only the reference genome will be returned.
chm: chromosome to be sampled, zero based.
bp0: initial position in the chromosome, zero based.
wnd: sample window size, (defaut=1024).
dsg: encoding for alternative allele count 0, 1, and 2; use '012' for
additive, '022' for dominative, and '002' for recessive encoding. The
default scheme is addtive, '012'.
"""
self.VGZ = vgz
# find the first locus in the VCF file
gv1 = next(vcfR(filename=vgz))
# VGZ reader point to the *.vcf.gz
vgz = vcfR(filename=vgz)
# the chromosome and its length
chm = gv1.CHROM if chm is None else CKY[chm]
cln = vgz.contigs[chm].length
self.CHR = int(chm)
# starting position
if not bp0:
bp0 = gv1.start
elif bp0 < 0:
bp0 = gv1.start + bp0
self.BP0 = bp0
# ending position
if not bp1:
bp1 = cln
elif bp1 < 0:
bp1 = cln + bp1
self.BP1 = bp1
# private members
self.__vgz__ = vgz
self.__gvr__ = gv1
self.__chm__ = chm
self.__bp0__ = bp0
self.__bp1__ = bp1
self.__wnd__ = wnd
self.__dsg__ = dsg
self.__dtp__ = dtp
self.__mem__ = []
self.__pos__ = []
self.__vid__ = []
def __init__(self, vgz, chm=None, bp0=None, bp1=None, sbj=None):
"""
vgz: bgzipped VCF file, with tabix index.
fsq: FASTQ file storing the reference genome
chm: the chromosome
bp0: the starting basepair, 0 based, inclusive
bp1: the ending basepair, 0 based, exclusive
"""
# record the parameters
self.VGZ = vgz
# find the first locus in the VCF file
gv1 = next(vcfR(filename=vgz))
# VGZ reader point to the *.vcf.gz
vgz = vcfR(filename=vgz)
# the chromosome and its length
chm = gv1.CHROM if chm is None else CKY[chm]
cln = vgz.contigs[chm].length
# starting position
if not bp0:
bp0 = gv1.start
elif bp0 < 0:
bp0 = gv1.start + bp0
# ending position
if not bp1:
bp1 = cln
elif bp1 < 0:
bp1 = cln + bp1
# restrict VCF range
vgz = vgz.fetch(chm, bp0, bp1)
# find the first locus in the region
gvr = None
while gvr is None:
try:
gvr = next(vgz)
except StopIteration:
break
if gvr.start < bp0:
gvr = None
# private members
self.__pos__ = bp0 # the pointer
self.__gvr__ = gvr # the variant
self.__vgz__ = vgz # vcf reader
self.__bp0__ = bp0 # starting position
self.__bp1__ = bp1 # ending position
def loadVCF(vcf):
""" Read genomic VCF file by name. Fix MAF and allele orders.
return genomic matrix and subject IDs. The function loads the
entire VCF into memory, so it is not recommended for huge VCF.
"""
# Python VCF reader: pip install pyvcf
from vcf import Reader as vcfR
# the two homogenuous chromosomes
A, B = [], []
reader = vcfR(filename=vcf)
sbj = reader.samples # subject IDs
for v in reader:
# copy #1 and #2
a = [int(g.gt_alleles[0] > '0') for g in v.samples]
b = [int(g.gt_alleles[1] > '0') for g in v.samples]
A.append(a)
B.append(b)
# compile genomic matrix
gmx = np.array([A, B], dtype='uint8')
# MAF fixing
i = np.where(gmx.sum((0, 2)) > gmx.shape[2])[0]
gmx[:, i, :] = 1 - gmx[:, i, :]
# Allele order fix, make sure copy(a) >= copy(b)
i = np.where(gmx[0, :, :] < gmx[1, :, :])
gmx[0, :, :][i] = 1
gmx[1, :, :][i] = 0
# dim_0: sample index; dim_1: copy number; dim_2: variant index
gmx = gmx.transpose(2, 0, 1)
return gmx, sbj
def loadVCF(vcf):
""" Read genomic VCF file by name.
Fix MAF and allele orders.
return genomic matrix and subject IDs.
"""
# Python VCF reader: pip install pyvcf
from vcf import Reader as vcfR
# the two homogenuous chromosomes
A, B = [], []
reader = vcfR(filename=vcf)
sbj = reader.samples # subject IDs
for v in reader:
# copy #1 and #2
a = [int(g.gt_alleles[0] > '0') for g in v.samples]
b = [int(g.gt_alleles[1] > '0') for g in v.samples]
A.append(a)
B.append(b)
# compile genomic matrix
gmx = np.array([A, B], dtype='uint8')
# MAF fixing
i = np.where(gmx.sum((0, 2)) > gmx.shape[2])[0]
gmx[:, i, :] = 1 - gmx[:, i, :]
# Allele order fix, make sure copy(a) >= copy(b)
i = np.where(gmx[0, :, :] < gmx[1, :, :])
gmx[0, :, :][i] = 1
gmx[1, :, :][i] = 0
# dim_0: sample index; dim_1: copy number; dim_2: variant index
gmx = gmx.transpose(2, 0, 1)
return gmx, sbj
def __init__(self, vgz, chm=None, bp0=0, bp1=None, wnd=1024, dsg='012'):
"""
vgz: bgzipped VCF file, with tabix index.
chm: the chromosome
bp0: the starting basepair, 0 based, inclusive
bp1: the ending basepair, 0 based, exclusive
"""
# record the parameters
self.VGZ = vgz
# find the first locus in the VCF file
gv1 = next(vcfR(filename=vgz))
# VGZ reader point to the *.vcf.gz
vgz = vcfR(filename=vgz)
# the chromosome and its length
chm = gv1.CHROM if chm is None else CKY[chm]
cln = vgz.contigs[chm].length
self.CHR = int(chm)
# starting position
bp0 = bp0 % cln
# ending position
if not bp1:
bp1 = cln
if bp1 < 0:
bp1 = cln + bp1
# restrict VCF range
vgz = vgz.fetch(chm, bp0, bp1)
# private members
self.__chm__ = chm
self.__pos__ = bp0 # the pointer
self.__vgz__ = vgz # vcf reader
self.__bp0__ = bp0 # starting position
self.__bp1__ = bp1 # ending position
self.__wnd__ = wnd
self.__dsg__ = dsg
self.__pos__ = []
self.__vid__ = []