def get_dbscSNV_ant(chrom, pos, ref, alt):
'''Checks if the variant affects splicing'''
f = '/dbscSNV/dbscSNV1.1.chr' + chrom + '.gz'
fo = Tabix(f)
for rec in fo.query(chrom, pos - 1, pos + 1):
if int(rec[1]) == pos and rec[2] == ref and rec[3] == alt:
ada_score, rf_score = rec[-2], rec[-1].strip()
if rf_score and rf_score != '.':
rf_score = ''
return float(ada_score), rf_score
return '', ''
def _get_coord_depth(invcf, capture_file_path, genelist_config, patient_capture_config):
#global capcoord_dp
# Load the genelist
gl = _load_genelist(genelist_config)
#gl = ['CCDC39', 'CSF2RA', 'CTC1', 'DNM1L', 'FBN1', 'HSD17B4', 'HYDIN', 'TERT']
# Load the patient capture file
pc = _load_patient_capture(patient_capture_config)
# Load the captured coordinates for each sample
CAP = {'v2': '4-Hopkins_clinical_panel_capture_v2paper.bed',
'v1b': '4-Hopkins_clinical_panel_capture_v1b.bed'}
cap_exon_mean = {}
vcfo = Tabix(invcf)
samples = _get_sample_ids(invcf)
for cver, pl in pc.items():
capture_file = os.path.join(capture_file_path, CAP[cver])
cap_gene_coords = _load_capture(capture_file, gl)
for gene, cap_exons in cap_gene_coords.items():
for idx, exon in enumerate(cap_exons):
exno = idx + 1
chrom, sp, ep = exon
key = (chrom, sp, ep, exno)
temp = {}
for rec in vcfo.query(chrom, sp, ep):
gt_info_raw = rec[9:]
for sid, gt_info in zip(samples, gt_info_raw):
if sid not in pl:
continue
if sid not in temp:
temp[sid] = []
if gt_info == './.':
temp[sid].append(0)
else:
gt_d = gt_info.split(':')
if len(gt_d) <= 3:
temp[sid].append(int(gt_d[-1]))
else:
temp[sid].append(int(gt_d[2]))
# Compute the mean
for sid, val in temp.items():
m = np.average(val)
if sid not in cap_exon_mean:
cap_exon_mean[sid] = {}
if gene not in cap_exon_mean[sid]:
cap_exon_mean[sid][gene] = {}
cap_exon_mean[sid][gene][key] = m
return cap_exon_mean
def gerp(vf, af, name="gerp"):
print "inside gerp"
v = BedTool(vf)
t = Tabix(af)
results = {}
for var in v:
try:
result = 0.0
num = 0
for res in t.query(var.chrom, var.start, var.end):
result += float(res[4])
num += 1
if num > 0:
results[var.name] = result/num
except:
pass
print "exit gerp"
return Series(results, name=name)
samples = []
results = {}
for ln in open(sys.argv[2]):
sample, fn = ln.rstrip().split("\t")
samples.append(sample)
results[sample] = {}
# p = subprocess.Popen(['bcftools', 'view', fn], stdout=subprocess.PIPE)
# vcf_reader = vcf.VCFReader(p.stdout, 'rb')
# for record in vcf_reader:
# pos = (record.CHROM, record.POS)
print fn
tab = Tabix(fn)
for pos in positions:
search = "%s:%d-%d;" % (pos[0], pos[1], pos[1])
try:
itr = tab.fetch(search)
rec = itr.next()
except StopIteration:
print "can't find: %s" % (search,)
pass
cols = rec.split("\t")
record = {}
record['REF'] = cols[3]
record['ALT'] = cols[4]
record['QUAL'] = float(cols[5])
def _get_coord_depth(invcf, capture_file_path, patient_capture_config,
sample_ids, qgene):
# Load the patient capture file
pc = _load_patient_capture(patient_capture_config, sample_ids)
print pc
# Load the captured coordinates for each sample
CAP = {
'v2': '4-Hopkins_clinical_panel_capture_v2.bed',
'v1b': '4-Hopkins_clinical_panel_capture_v1b.bed'
}
vcfo = Tabix(invcf)
samples = _get_sample_ids(invcf)
coord_depth = {}
for cver, pl in pc.items():
if cver == 'v2':
continue
capture_file = os.path.join(capture_file_path, CAP[cver])
cap_exons = _load_capture(capture_file, qgene)
ex_cnt = len(cap_exons)
for idx, exon in enumerate(cap_exons):
exno = idx + 1
chrom, sp, ep = exon
midpos = range(sp, ep)[(ep - sp) / 2]
for rec in vcfo.query(chrom, sp, ep):
pos = int(rec[1])
gt_info_raw = rec[9:]
tsid = []
for sid, gt_info in zip(samples, gt_info_raw):
if sid not in pl:
continue
tsid.append(sid)
if sid not in coord_depth:
coord_depth[sid] = [[], [], []]
if gt_info == './.':
dp = 0
else:
gt_d = gt_info.split(':')
if len(gt_d) <= 3:
dp = int(gt_d[-1])
else:
dp = int(gt_d[2])
coord_depth[sid][0].append(pos)
coord_depth[sid][1].append(dp)
if pos == midpos:
coord_depth[sid][2].append('CR-' + str(exno))
else:
coord_depth[sid][2].append('')
if idx + 1 == ex_cnt:
continue
for sid in tsid:
n = 105
for fp in range(ep + 1, ep + 1 + n):
coord_depth[sid][0].append(fp)
coord_depth[sid][1].append(0)
coord_depth[sid][2].append('')
return coord_depth
def _generate_TSV(outdir, samples, capcoord):
"""Internal Method. Generates two tsv files (one for snv and
other for indel) that contains the variants reported in 1000
Genomes Project Phase 3 dataset for given genomic regions.
Args:
outdir (str): Path where TSV files will be created
samples (list): List of samples for which the variants
will be extracted from 1000 Genomes dataset.
capcoord (list): List of bed format coordinates wrapped as
tuples. For eg.
[(chrom, start_pos, end_pos), ...]
Returns:
None
"""
out_snpfile = os.path.join(outdir, 'snp_control_variants_all.tsv')
outsnp = open(out_snpfile, 'w')
out_indelfile = os.path.join(outdir, 'indel_control_variants_all.tsv')
outindel = open(out_indelfile, 'w')
chroms = [str(e) for e in range(1, 23)] + ['X', 'Y']
hflag = False
for chrom in chroms:
if chrom == 'X':
fp1 = 'ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chrX.phase3_shapeit2_mvncall_integrated_v1b.20130502.genotypes.vcf.gz'
ofp1 = os.path.join(outdir, 'ALL.chrX.phase3_shapeit2_mvncall_integrated_v1b.20130502.genotypes.vcf.gz')
fp2 = 'ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chrX.phase3_shapeit2_mvncall_integrated_v1b.20130502.genotypes.vcf.gz.tbi'
ofp2 = os.path.join(outdir, 'ALL.chrX.phase3_shapeit2_mvncall_integrated_v1b.20130502.genotypes.vcf.gz.tbi')
else:
fp1 = 'ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr%s.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz' % chrom
ofp1 = os.path.join(outdir, 'ALL.chr%s.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz' % chrom)
fp2 = 'ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr%s.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz.tbi' % chrom
ofp2 = os.path.join(outdir, 'ALL.chr%s.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz.tbi' % chrom)
if not os.path.exists(ofp1):
cmd = "wget %s -O %s" % (fp1, ofp1)
print cmd
os.system(cmd)
if not os.path.exists(ofp2):
cmd = "wget %s -O %s" % (fp2, ofp2)
print cmd
os.system(cmd)
for line in os.popen("zcat %s | head -260 | grep '#CHROM'" % ofp1):
line = line.strip()
h = line.split('\t')[:9]
kgsamples = line.split('\t')[9:]
if hflag == False:
h += samples
outsnp.write('\t'.join(h) + '\n')
outindel.write('\t'.join(h) + '\n')
hflag = True
vcfo = Tabix(ofp1)
for coord in capcoord:
if coord[0] != chrom:
continue
for rec in vcfo.query(*coord):
if '<' in rec[4] or ',' in rec[4]:
continue
if len(rec[3]) == len(rec[4]): # to make Indel control set
out = outsnp
if len(rec[3]) != len(rec[4]): # to make SNV control set
out = outindel
sv = []
d = dict(zip(kgsamples, rec[9:]))
temp = set([])
for sid in samples:
sv.append(d[sid])
temp.add(d[sid])
if len(temp) == 1 and list(temp)[0] in ['0|0', '0/0']:
continue
out.write('\t'.join(rec[:9] + sv) + '\n')
def get_varcnt_gp(invcf, antvcf, capcoord, samples, varcnt, vao):
'''Returns a dictionary where the keys are sampleid and values are a list
of length 11 whose the elements are the count of -
* Nocall
* VarLQ
* GTLQ
* REF allele
* HetALT
* HomALT
* Ts
* Tv
* Common Variants (Based on 1000 Genomes)
* Rare Variants (Based on 1000 Genomes)
* Novel Variants (Based on 1000 Genomes)
'''
vcfsamples = get_vcfsamples(invcf)
for sid in samples:
varcnt[sid] = [
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
] # NoCall, VarLQ, GTLQ, REF, HetALT, HomALT, Ts, Tv, Common, Rare, Novel
vcfo = Tabix(invcf)
antvcfo = Tabix(antvcf)
for coord in capcoord:
try:
qdata = vcfo.query(*coord)
except:
qdata = None
if qdata is None:
continue
for rec in qdata:
filter = rec[6]
gts = rec[9:]
trans = vao.retrieve_genedef(rec[0], int(rec[1]), int(rec[1]))
if not trans: # indicating its an intergenic region
continue
for sid, gtinfo in zip(vcfsamples, gts):
if sid in samples:
if gtinfo == '.' or './.' in gtinfo: # No Call
varcnt[sid][0] += 1
elif filter != 'PASS':
varcnt[sid][1] += 1 # VarLQ
elif '0/0' in gtinfo:
varcnt[sid][3] += 1 # REF
else:
gt, gq = gtinfo.split(':')[:2]
gq = int(gq)
a1, a2 = gt.split('/')
a1, a2 = int(a1), int(a2)
ref, alt = rec[3], rec[4].split(',')[a2 - 1]
ct = get_change_type(ref, alt)
arec = antvcfo.query(rec[0],
int(rec[1]) - 1, int(rec[1]))
com, rare, novel = False, False, False
iflag = False
for e in arec:
if int(rec[1]) == int(e[1]):
kgaf = 0.0
if 'KGDB' in e[7]:
kgaf = float(e[7].split('KGAF=')[1].split(
';')[0].split(',')[a2 - 1])
exacaf = 0.0
if 'EXACDB' in e[7]:
exacaf = float(
e[7].split('EXACAF=')[1].split(
';')[0].split(',')[a2 - 1])
if kgaf == 0.0 and exacaf == 0.0:
novel = True
elif kgaf < 0.05 and exacaf < 0.05:
rare = True
else:
com = True
break
if gq < 30: # GTLQ
varcnt[sid][2] += 1
elif a1 != a2: # HetALT
varcnt[sid][4] += 1
if ct == 'ts':
varcnt[sid][6] += 1
elif ct == 'tv':
varcnt[sid][7] += 1
if com:
varcnt[sid][8] += 1
elif rare:
varcnt[sid][9] += 1
elif novel:
varcnt[sid][10] += 1
print sid, rec[0], rec[1], rec[3], rec[4]
elif a1 == a2: # HomALT
varcnt[sid][5] += 1
if ct == 'ts':
varcnt[sid][6] += 1
elif ct == 'tv':
varcnt[sid][7] += 1
if com:
varcnt[sid][8] += 1
elif rare:
varcnt[sid][9] += 1
elif novel:
varcnt[sid][10] += 1
print sid, rec[0], rec[1], rec[3], rec[4]
return varcnt
