def merge_sv(raw, SVs, minOvr):
non, result, merged, premerge = {}, {}, [], []
for locus in SVs:
non[locus] = SVs[locus].med_alt
dels = BedTool([(format_chrom(x[0]), x[1], x[2], x[3]) for x in raw
if 'DEL' in x[3]])
dups = BedTool([(format_chrom(x[0]), x[1], x[2], x[3]) for x in raw
if 'DUP' in x[3]])
dels_nomerge = first_merge(dels, minOvr)
while len(check_merge(dels, minOvr)) > 0:
dels = merging(check_merge(dels, minOvr), non)
for x in dels:
premerge.append(x)
dups_nomerge = first_merge(dups, minOvr)
while len(check_merge(dups, minOvr)) > 0:
dups = merging(check_merge(dups, minOvr), non)
for x in dups:
premerge.append(x)
premerge = set(premerge)
for sv in list(
itertools.chain(dels, dels_nomerge, dups, dups_nomerge, premerge)):
sv = tokenize_sv(tuple(sv))
result[sv] = 1
for sv in raw:
chrom, start, end, svtype = sv
_sv = (format_chrom(chrom), start, end, svtype)
if result.get(_sv) != None: merged.append(sv)
return merged
def output(Structural_Variant, SVs, Ped, ids, gen, ofh, anno_flag, tmp_dir):
Ofh = VCF(ofh)
Ofh.init_header(datetime.date.today(), ids, Structural_Variant, gen)
Ofh.load_genotypes(Structural_Variant, SVs, Ped, ids, gen, anno_flag,
tmp_dir)
if not ofh.endswith('.vcf'): ofh = ofh + '.vcf'
vcf_ofh = open(ofh, 'w')
vcf_ofh.write('\n'.join(Ofh.head) + '\n')
chroms = {}
for entry in Structural_Variant.raw:
chroms[entry[0]] = 1
variant_id = '.'
if Structural_Variant.variant_id.get(entry) != None:
variant_id = Structural_Variant.variant_id[entry]
locus = (format_chrom(entry[0]), int(entry[1]), int(entry[2]),
str(entry[3]))
genotypes = []
if Ofh.genotypes.get(locus) == None:
for sample_id in ids:
gt = './.'
if (locus[0] == 'chrX'
or locus[0] == 'chrY') and Ped.males.get(
sample_id
) != None and Structural_Variant.par[locus] == False:
gt = '.'
genotypes.append(gt)
else:
genotypes = Ofh.genotypes[locus]
genotypes = '\t'.join(genotypes)
out = Ofh.init_row(entry)
vcf_ofh.write('{}\t{}\t{}\t.\t{}\t{}\n'.format(entry[0],
int(entry[1]) + 1,
variant_id, out,
genotypes))
vcf_ofh.close()
def tmp_chrom_file(self, tmp_dir=None, genome=True, chrom=None):
make_dir(tmp_dir)
tmp_genome = tmp_dir + '{}.genome'.format(self.id)
tmpfh = open(tmp_genome, 'w')
if genome == True:
if chrom == None:
for ref in self.refs:
tmpfh.write('{}\t{}\n'.format(format_chrom(ref),
self.refs[ref]))
else:
tmpfh.write('{}\t{}\n'.format(format_chrom(chrom),
self.refs[chrom]))
if genome == False:
for ref in self.refs:
tmpfh.write('{}\t0\t{}\n'.format(format_chrom(ref),
self.refs[ref]))
tmpfh.close()
return tmp_genome
def __init__(self, sv=None, variant_id=None, gen=None):
self.raw = [(str(x[0]), int(x[1]), int(x[2]), str(x[3]))
for x in BedTool(list(set(sv))).sort()]
self.variant_id = variant_id
self.par = {} # [locus]=True/False; true if intersects PAR
for locus in self.raw:
chrom = format_chrom(locus[0])
if chrom != 'chrX' and chrom != 'chrY': continue
key = (chrom, int(locus[1]), int(locus[2]), str(locus[3]))
self.par[key] = check_PAR(
'{} {} {}'.format(chrom, locus[1], locus[2]), gen)
def __init__(self,data=None):
self.locus=(format_chrom(data[0]),int(data[1]),int(data[2]),str(data[3])) # (chrom,start,end,type)
self.ref={} # [classifier_name]=[reference genotype likelihoods]
self.med_ref='NA'
self.alt={} # [classifier_name]=[alternate genotype likelihoods]
self.med_alt='NA'
self.format={} # [locus+id]='GT:CN:PE:SR:SC:SP:AR:HT:SQ:GL'
self.gt={} # gt[locus+id]=genotype
self.gq={} # gq[locus+id]=(ref,het,hom) likelihoods
self.breakpoint_feats=0 # breakpoint features: discordant paired-ends + split-reads
self.clf=[] # classifiers
self.standard_filter='PASS'
self.denovo_filter='PASS'
def load_1kgp(self,raw=None,svtype=None,gen=None,tmp_bed=None):
sv = BedTool([(format_chrom(x[0]),x[1],x[2],x[3]) for x in raw if svtype in str(x[3])]).sort()
sv.intersect('{}annotation_files/{}_1000Genomes_{}.bed'.format(Config().resource_path(),gen,svtype), f=0.8, F=0.8, wao=True,output=tmp_bed)
with open(tmp_bed,'r') as f:
for l in f:
x = tuple(l.rstrip().split('\t'))
locus = tokenize_sv(x)+(str(x[3]),)
ovr = int(x[-1])
if ovr==0: continue
ovr = format(float(x[len(x)-1])/(int(x[2])-int(x[1])),'.2f')
if self._1kgp.get(locus)==None:
self._1kgp[locus]=(x[len(x)-2],ovr)
elif self._1kgp.get(locus)!=None and float(ovr) > float(self._1kgp[locus][1]):
self._1kgp[locus]=(x[len(x)-2],ovr)
else: continue
os.remove(tmp_bed)
def load_genotypes(self,
Structural_Variant=None,
SVs=None,
Ped=None,
ids=None,
gen=None,
no_anno=None,
tmp_dir=None):
svs = BedTool(
list(
set([(format_chrom(x[0]), x[1], x[2])
for x in Structural_Variant.raw]))).sort()
if no_anno == False:
Annot = Annotation()
Annot.check_overlap(svs, Structural_Variant.raw, gen, tmp_dir)
self.Annotations = Annot
for locus in SVs:
Variant = SVs[locus]
self.quals[locus] = Variant.med_ref, Variant.med_alt
self.filters[
locus] = Variant.standard_filter, Variant.denovo_filter
for sample_id in ids:
gt = './.'
if (locus[0] == 'chrX'
or locus[0] == 'chrY') and Ped.males.get(
sample_id
) != None and Structural_Variant.par[locus] == False:
gt = '.'
if Variant.gt.get(locus + (sample_id, )) != None:
gt = Variant.gt[locus + (sample_id, )]
for allele in gt.split(':').pop(0).split('/'):
if allele == '.': continue
if self.allele_freq.get(locus) == None:
self.allele_freq[locus] = [int(allele), 1]
else:
self.allele_freq[locus] = [
self.allele_freq[locus][0] + int(allele),
self.allele_freq[locus][1] + 1
]
if self.genotypes.get(locus) == None:
self.genotypes[locus] = [gt]
else:
self.genotypes[locus].append(gt)
def init_header(self,
date=None,
ids=None,
Structural_Variant=None,
gen=None):
chroms, refs, contigs = {}, OrderedDict(), []
from sv2Config import Config
with open('{}{}.genome'.format(Config().resource_path(), gen),
'r') as f:
for l in f:
chrom, leng = l.rstrip().split('\t')
chroms[format_chrom(chrom)] = leng
for x in Structural_Variant.raw:
if chroms.get(format_chrom(x[0])) != None:
refs[x[0]] = chroms[format_chrom(x[0])]
for chrom in refs:
contigs.append('##contig=<ID={},length={}>'.format(
chrom, refs[chrom]))
self.head = [
'##fileformat=VCFv4.1',
'##fileDate={}'.format(date),
'##SV2_CMD="{}"'.format(' '.join(map(str, sys.argv[:]))),
'##INFO=<ID=END,Number=1,Type=Integer,Description="End position of structural variant">',
'##INFO=<ID=SVTYPE,Number=1,Type=String,Description="Type of structural variant">',
'##INFO=<ID=SVLEN,Number=1,Type=Integer,Description="Difference in length between REF and ALT alleles">',
'##INFO=<ID=DENOVO_FILTER,Number=1,Type=String,Description="Stringent filter status, recommended for de novo mutation discovery">',
'##INFO=<ID=REF_GTL,Number=1,Type=Float,Description="Median Phred-adjusted REF genotype likelihood">',
'##INFO=<ID=AF,Number=1,Type=Float,Description="Alternate allele frequency,in the range (0,1)">',
'##INFO=<ID=CYTOBAND,Number=.,Type=String,Description="Cytoband(s) overlapping the variant">',
'##INFO=<ID=REPEATMASKER,Number=2,Type=String,Description="Name and reciprocal overlap of RepeatMasker variant">',
'##INFO=<ID=1000G_ID,Number=1,Type=String,Description="1000 Genomes Phase 3 integrated SV callset variant identifier">',
'##INFO=<ID=1000G_OVERLAP,Number=1,Type=Float,Description="Overlap to 1000 Genomes Phase 3 variant, in the range (0,1)">',
'##INFO=<ID=DESCRIPTION,Number=1,Type=String,Description="Verbose description of SV, 1-based coordinates"',
'##INFO=<ID=GENES,Number=1,Type=String,Description="Genes overlapping the variant, pipe-separated by transcripts>"',
'##INFO=<ID=ABPARTS,Number=1,Type=Float,Description="Overlap to antibody parts, in the range (0,1)">',
'##INFO=<ID=CENTROMERE,Number=1,Type=Float,Description="Centromere overlap, in the range (0,1)">',
'##INFO=<ID=GAP,Number=1,Type=Float,Description="Overlap to gaps in the reference, in the range (0,1)">',
'##INFO=<ID=SEGDUP,Number=1,Type=Float,Description="Segmental duplication overlap, in the range (0,1)">',
'##INFO=<ID=STR,Number=1,Type=Float,Description="Short tandem repeat overlap, in the range (0,1)">',
'##INFO=<ID=UNMAPPABLE,Number=1,Type=Float,Description="Overlap to DAC Blacklisted Regions, in the range (0,1)">',
'##FILTER=<ID=ABPARTS,Description="Variant overlaps to antibody parts >50%">',
'##FILTER=<ID=CENTROMERE,Description="Variant overlaps to centromere >50%">',
'##FILTER=<ID=GAP>,Description="Variant overlaps to reference gaps >50%">',
'##FILTER=<ID=GENOTYPEFAIL,Description="Variant was unable to be genotyped">',
'##FILTER=<ID=NOALT,Description="No alternate allele detected">',
'##FILTER=<ID=SEGDUP,Description="Variant overlaps to segmental duplications >50%">',
'##FILTER=<ID=STR,Description="Variant overlaps to short tandem repeats >50%">',
'##FILTER=<ID=UNMAPPABLE,Description="Variant overlaps to DAC Blacklisted Regions regions >50%">',
'##FILTER=<ID=FAIL,Description="Variant failed standard filters">',
'##FILTER=<ID=PASS,Description="Variant passed standard filters">',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">',
'##FORMAT=<ID=CN,Number=1,Type=Float,Description="Copy number estimate">',
'##FORMAT=<ID=PE,Number=1,Type=Float,Description="Normalized discordant paired-end count">',
'##FORMAT=<ID=SR,Number=1,Type=Float,Description="Normalized split-read count">',
'##FORMAT=<ID=SC,Number=1,Type=Float,Description="SNV normalized coverage">',
'##FORMAT=<ID=NS,Number=1,Type=Integer,Description="Number of SNVs within locus">',
'##FORMAT=<ID=HA,Number=1,Type=Float,Description="Heterozygous allele ratio">',
'##FORMAT=<ID=NH,Number=1,Type=Integer,Description="Number of heterozygous SNVs">',
'##FORMAT=<ID=SQ,Number=1,Type=Float,Description="Phred-scaled genotype likelihood">',
'##FORMAT=<ID=GL,Number=G,Type=Float,Description="Phred-scaled genotype likelihoods in the order, REF:(0/0), HET:(0/1), HOM:(1/1)">',
'##ALT=<ID=DEL,Description="Deletion, if 80% reciprocal overlap with RepeatMasker element, the class, name, and family are given separated by colons">',
'##ALT=<ID=DUP,Description="Duplication, if 80% reciprocal overlap with RepeatMasker element, the class, name, and family are given separated by colons">',
'{}'.format('\n'.join(contigs)),
'#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\t{}'.format(
'\t'.join(ids)),
]
def tokenize_sv(x): return (format_chrom(str(x[0])),int(x[1]),int(x[2]))
