def get_variant_summary_probs(self, rsid, threshold):
variant_array = []
msg = ""
docs = self.var_coll.get_variant_data_multi(rsid)
for doc in docs:
# always force chromosome to 2 digits
chromosome = "%.2d" % int(doc["chromosome"])
fpath = self.filepaths_coll.get_filepath(doc["assaytype"],
chromosome)
fullrec = self.var_coll.get_raw_variant_values(
fpath, chromosome, doc['position'])
vcfr = VCFrecord(fullrec)
prfx, sfx = vcfr.get_prfx_sfx()
probidx = vcfr.get_probidx()
(gc_count_dict, sample_count, geno_count, maf, alleleAf, alleleBf,
p_hwe) = self.var_coll.get_genotype_probs(sfx, threshold, probidx)
doc['selected'] = 1
doc['a_af'] = alleleAf
doc['b_af'] = alleleBf
doc['hwe_p'] = p_hwe
doc['Missing'] = 0
if 'Missing' in gc_count_dict:
doc['Missing'] = gc_count_dict['Missing']
variant_array.append(doc)
if len(variant_array) == 0:
msg = "Variant NOT FOUND - %s, " % (rsid)
return (variant_array, msg)
def check_concordancies(self, data_list, assays, chipval):
hwe_values = [0.0] * len(data_list)
maf_values = [0.0] * len(data_list)
obs = [0.0] * 3
exp = [0.0] * 3
allele_ref_1 = ""
allele_alt_1 = ""
allele_ref_2 = ""
allele_alt_2 = ""
#print "CHECK_CONC:", len(data_list)
for i, vcf_record in enumerate(data_list):
if len(vcf_record) > 0:
vcfr = VCFrecord(vcf_record)
probidx = vcfr.get_probidx()
homref_count, het_count, homalt_count, nc_count, miss_count = self.vcfr.get_allele_counts_from_array(
data)
allele_a, allele_b = vcfr.get_alleles()
if allele_ref_1 == "":
allele_ref_1 = allele_a
allele_alt_1 = allele_b
# Add 1 to prevent 0-divide
obs[0] = homref_count + 1
obs[1] = het_count + 1
obs[2] = homalt_count + 1
else:
allele_ref_2 = allele_a
allele_alt_2 = allele_b
exp[0] = homref_count + 1
exp[1] = het_count + 1
exp[2] = homalt_count + 1
if (allele_ref_1 != allele_ref_2) or (allele_alt_1 !=
allele_alt_2):
varid = vcfr.get_varid(data)
posn = vcfr.get_posn(data)
self.allele_discord_count += 1
logging.info(
"Allele discordancy: assay1=%s, assay2=%s, varid=%s, posn=%d, ref1=%s, alt1=%s, ref2=%s, alt2=%s",
assays[0], assays[i], varid, int(posn),
allele_ref_1, allele_alt_1, allele_ref_2,
allele_alt_2)
#print "Allele discord"
return False
chi_stat, chi_p_value = chisquare(obs, f_exp=exp)
varid = vcfr.get_varid(data)
posn = vcfr.get_posn(data)
if chi_p_value < chipval:
self.chisq_count += 1
logging.info(
"CHI SQ test REJECT: assay1=%s, assay2=%s, varid=%s, posn=%d, chistat=%f, p_val=%e, chipval=%e, obs=%s, exp=%s, at %d",
assays[0], assays[i], varid, int(posn), chi_stat,
chi_p_value, chipval, str(obs), str(exp), i)
#print "CHISQ discord"
return False
logging.info(
"CHI SQ test OK: assay1=%s, assay2=%s, varid=%s, posn=%d, chistat=%f, p_val=%e, obs=%s, exp=%s, at %d",
assays[0], assays[i], varid, int(posn), chi_stat,
chi_p_value, str(obs), str(exp), i)
return True
def get_combined_array(self,
buffer_list,
cr_list,
assay_list,
threshold=0.9):
"""
For each list of data, for each element of list of data:
1) Find the col header from the corresonding file_position element
2) Use the col_header to find the combined postion
3) Place the data_element in the combined postion *
TODO - conflict resolution, what to do if a slot is already occupied
TODO - CR check
"""
#print "COMBO", self.combined_positions
#
#print "ASSAY_LIST: %s" % (str(assay_list))
assay_posns = {}
for i, assaytype in enumerate(assay_list):
assay_posns[i] = assaytype
#print "ASSAY_POSNS: %s" % (str(assay_posns))
combo_array = ["."] * len(self.combined_positions)
#print "BUFFL", len(buffer_list)
for i, vcf_record in enumerate(buffer_list):
if len(vcf_record) > 0:
#print "asstp: %d, %s" % (i, assay_list[i])
vcfr = VCFrecord(vcf_record)
prfx, data_list = vcfr.get_prfx_sfx()
probidx = vcfr.get_probidx()
rsid = vcfr.get_varid()
hasAT = vcfr.has_fmt("AT")
for j, dataelem in enumerate(data_list):
if data_list[j] != ".":
cpos = self.combined_positions[self.file_positions[i]
[j]]
geno = self.call_geno_for_threshold(
data_list[j], probidx, threshold)
if (hasAT == False):
geno = geno + ":" + self.assay_abbrev[
assay_list[i]]
if combo_array[cpos] != ".":
self.geno_overlap_count += 1
#print "OVERLAP %s:%s - %s vs %s" % (rsid, self.file_positions[i][j], combo_array[cpos], geno)
geno = self.call_genotype(combo_array[cpos], geno,
probidx)
combo_array[cpos] = geno
return combo_array
def get_rslist_data(self, input_rslist, threshold, download_list):
msg = None
snpdata = "SNPId,AssayType,chr,pos,REF,REF_fr,ALT,ALT_fr,MAF,Imputed,CallRate,HWE_pval,Info\n"
data = []
assaytypelist = []
probidxlist = []
rslist = []
assaytypes = {}
Afreq = {}
Bfreq = {}
data_count = 0
impDict = {}
for rsid in input_rslist:
docs = self.var_coll.get_variant_data_multi(rsid)
if len(docs) > 0:
rslist.append(rsid)
# handling SNPs on multiple platforms
for doc in docs:
# always force chromosome to 2 digits
chromosome = "%.2d" % int(doc["chromosome"])
# first get filepath
fpath = self.filepaths_coll.get_filepath(
doc["assaytype"], chromosome)
# get raw variant data
fullrec = self.var_coll.get_raw_variant_values(
fpath, chromosome, doc['position'])
geno_count = 0
sample_count = 0
hwep = 0.0
vcfr = VCFrecord(fullrec)
prfx, sfx = vcfr.get_prfx_sfx()
probidx = vcfr.get_probidx()
(gc_count_dict, sample_count, geno_count, maf, alleleAf,
alleleBf, p_hwe) = self.var_coll.get_genotype_probs(
sfx, threshold, probidx)
Afreq[doc["rsid"] + "_" + doc["assaytype"]] = alleleAf
Bfreq[doc["rsid"] + "_" + doc["assaytype"]] = alleleBf
data_count += 1
hwep = float(p_hwe)
assaytypelist.append(doc["assaytype"])
data.append(vcfr)
if doc["assaytype"] not in assaytypes:
assaytypes[doc["assaytype"]] = 1
imputed = 0
if "imputed" in doc:
imputed = 1
if "info" in doc:
if doc["info"] != 1.0:
imputed = 1
impDict[doc["rsid"] + "_" + doc["assaytype"]] = imputed
snpdata += "%s,%s,%s,%d,%s,%s,%s,%s,%s,%d,%.5f,%.5f,%.5f\n" % (
doc["rsid"], doc["assaytype"], doc["chromosome"],
doc["position"], doc["alleleA"],
alleleAf, doc["alleleB"], alleleBf, maf, imputed,
float(geno_count) / sample_count, hwep, doc["info"])
pdata = self.get_sample_values(assaytypelist, data, data_count, rslist,
impDict, assaytypes, Afreq, Bfreq,
threshold)
return (pdata, snpdata, msg)
def main(options):
hdrData = ["id"]
sampleDict = {}
colPosns = {}
RefAlleleDict = {}
AltAlleleDict = {}
count = 0
mafh = Mafhelper()
for line in sys.stdin:
line = line.strip()
if (line.startswith('##')):
pass
else:
vcfr = VCFrecord(line)
prfx, sfx = vcfr.get_prfx_sfx()
#print prfx
if (line.startswith('#')):
# Parse out the header record.
for i, col_hdr in enumerate(sfx):
colPosns[i] = col_hdr
sampleDict[col_hdr] = []
else:
flip = False
varid = vcfr.get_varid_ukb()
#logging.info("varid=%s", varid)
ref, alt = vcfr.get_alleles()
probidx = vcfr.get_probidx()
hdr_allele = alt
homref_count, het_count, homalt_count, nc_count, miss_count = vcfr.get_allele_counts(
)
call_count = homref_count + het_count + homalt_count
maf, ma = mafh.maf(het_count, homref_count, ref, homalt_count,
alt, nc_count)
RefAlleleDict[varid] = ref
AltAlleleDict[varid] = alt
#if ma == ref:
# flip = True
# hdr_allele = ref
# logging.info("FLIP for %s, %s, %s", varid, ref, alt)
hdrData.append(varid)
for i, str_geno in enumerate(sfx):
if str_geno != ".":
geno = str_geno.split(":")
max_prob, max_idx = get_max_prob(geno, probidx)
i_call = icalls[geno[0]]
if flip == True:
if i_call == "0":
i_call == "2"
elif i_call == "2":
i_call = "0"
sampleDict[colPosns[i]].append(str(i_call))
else:
sampleDict[colPosns[i]].append("")
print ",".join(hdrData)
for samp in sampleDict:
count += 1
print ",".join([samp] + sampleDict[samp])
return count