def doTDT(v, family, thresh):
""" Perform the Transmission Disequilibrium Test (TDT).
Parameters
----------
v: line of the VCF
family is a hash table with:
Key: individual id Value: [father id, mother id, sex]
thresh: hash table of thresholds
"""
# If filters on the line failed move on.
if not v:
return None
TU = [0, 0] # Array of [transmissions, untransmissions]
TU_m = [0, 0] # same as TU but for males
TU_f = [0, 0] # same as TU but for females
mErr = 0 # Count mendelian error: Parents: ref, child: het
mErr_o = 0 # Count other mendelian errors
Nproband_alt = 0 # Number of homozygous alt probands that passed all thresholds
AN = 0 # Number of families that passed all thresholds
indivs_T = [v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT']
] # Array of individuals who were transmitted the variant
indivs_U = [v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT']
] # Array of individuals who did not receive the variant
for indiv_id in family.keys(): #loop through all the probands
# indiv_data is their GT:AD:DP:GQ:PL stats
indiv_data = v[indiv_id]
if indiv_data == None:
continue
# Apply quality control filters on proband.
if not filters.passFilters(indiv_data, thresh,
thresh['GQ_Kid_Thresh']):
continue
# Apply PL filter to child.
if not filters.PhredScaleFilter(indiv_data, thresh['PL_Thresh']):
continue
father = v[family[indiv_id][0]]
mother = v[family[indiv_id][1]]
# Check if the parents have the alternate allele
# so they can pass it on AND apply quality control filters.
if filters.TDT_Parent_Filters(father, mother, thresh):
AN += 1 # all individuals in the nuclear family passed the filters
# TDT operates differently in the hemizygous chromosomes
# PAR regions defined from
# http://www.ncbi.nlm.nih.gov/projects/genome/assembly/grc/human/
# in this case we are in the Par region so transmission is normal
if filters.check_Hemizgyous(v['CHROM'], family[indiv_id][2],
filters.inPar(v['POS'])):
TU, TU_m, TU_f, mErr, mErr_o, transFlag = numberTransmissions(
indiv_data['GT'], father['GT'], mother['GT'], TU, TU_m,
TU_f, family[indiv_id][2], False, mErr, mErr_o)
else:
TU, TU_m, TU_f, mErr, mErr_o, transFlag = numberTransmissions(
indiv_data['GT'], father['GT'], mother['GT'], TU, TU_m,
TU_f, family[indiv_id][2], True, mErr, mErr_o)
if indiv_data['GT'] == 'homoAlt':
Nproband_alt += 1
if transFlag == True: # if the variant was transmitted
indivs_T.extend(
(indiv_id, family[indiv_id][2], indiv_data,
family[indiv_id][0], father, family[indiv_id][1], mother))
elif transFlag == False:
indivs_U.extend(
(indiv_id, family[indiv_id][2], indiv_data,
family[indiv_id][0], father, family[indiv_id][1], mother))
# Ignore the cases in which we have 0 transmissions and 0 untransmissions.
if TU[0] + TU[1] == 0:
return None
# Calculate percentage of mendelian errors.
mendErrorPercent = (mErr + mErr_o) / (TU[0] + TU[1] + mErr + mErr_o)
if vepFieldNames:
gene, anno, pph2, sift, lof = vepA.findVariantAnnotation(
v, args, vepFieldNames)
else:
gene, anno, pph2, sift, lof = ('', '', '', '', '')
return [
v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT'], v['FILTER'],
v['VQSLOD'], gene, anno, pph2, sift, lof, v['AF'], v['AC'], AN,
Nproband_alt, TU[0], TU[1], TU_m[0], TU_m[1], TU_f[0], TU_f[1], mErr,
mErr_o, mendErrorPercent
], indivs_T, indivs_U
def doTDT(v, family, thresh):
""" Perform the Transmission Disequilibrium Test (TDT).
Parameters
----------
v: line of the VCF
family is a hash table with:
Key: individual id Value: [father id, mother id, sex]
thresh: hash table of thresholds
"""
# If filters on the line failed move on.
if not v:
return None
TU = [0, 0] # Array of [transmissions, untransmissions]
TU_m = [0, 0] # same as TU but for males
TU_f = [0, 0] # same as TU but for females
mErr = 0 # Count mendelian error: Parents: ref, child: het
mErr_o = 0 # Count other mendelian errors
N_het = 0 # Number of heterozygous individuals that passed all thresholds
Nproband_alt = 0 # Number of homozygous alt probands that passed all thresholds
AN = 0 # Number of families that passed all thresholds
DP_het = [] # Array pf depth of all het individuals who passed filters
DP = [] # Array of depth of all non-het individuals who passed filters
AB = [] # Array of the allelic balance
indivs_T = [v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT']] # Array of individuals who were transmitted the variant
indivs_U = [v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT']] # Array of individuals who did not receive the variant
for indiv_id in family.keys(): #loop through all the probands
# indiv_data is their GT:AD:DP:GQ:PL stats
indiv_data = v[indiv_id]
if indiv_data == None:
continue
# Apply quality control filters on proband.
if not filters.passFilters(indiv_data, thresh, thresh['GQ_Kid_Thresh']):
continue
# Apply PL filter to child.
if not filters.PhredScaleFilter(indiv_data, thresh['PL_Thresh']):
continue
father = v[family[indiv_id][0]]
mother = v[family[indiv_id][1]]
# Check if the parents have the alternate allele
# so they can pass it on AND apply quality control filters.
if filters.TDT_Parent_Filters(father, mother, thresh):
AN += 1 # all individuals in the nuclear family passed the filters
# TDT operates differently in the hemizygous chromosomes
# PAR regions defined from
# http://www.ncbi.nlm.nih.gov/projects/genome/assembly/grc/human/
# in this case we are in the Par region so transmission is normal
if filters.check_Hemizgyous(v['CHROM'], family[indiv_id][2], filters.inPar(v['POS'])):
TU, TU_m, TU_f, mErr, mErr_o, transFlag = numberTransmissions(indiv_data['GT'], father['GT'], mother['GT'], TU, TU_m, TU_f, family[indiv_id][2], False, mErr, mErr_o)
else:
TU, TU_m, TU_f, mErr, mErr_o, transFlag = numberTransmissions(indiv_data['GT'], father['GT'], mother['GT'], TU, TU_m, TU_f, family[indiv_id][2], True, mErr, mErr_o)
# Update totals
AB, N_het, Nproband_alt, DP, DP_het = updateTotals(AB, N_het, Nproband_alt, DP, DP_het, indiv_data, father, mother)
if transFlag == True: # if the variant was transmitted
indivs_T.extend((indiv_id, indiv_data, family[indiv_id][0], father, family[indiv_id][1], mother))
elif transFlag == False:
indivs_U.extend((indiv_id, indiv_data, family[indiv_id][0], father, family[indiv_id][1], mother))
# Ignore the cases in which we have 0 transmissions and 0 untransmissions.
if TU[0] + TU[1] == 0:
return None
# Calculate percentage of mendelian errors.
mendErrorPercent = (mErr + mErr_o) / (TU[0] + TU[1] + mErr + mErr_o)
# Calculate averages for allelic balance (AB), depth (DP), and depth of hets (DP_het).
AB = np.average(np.array(AB))
DP = np.average(np.concatenate((np.array(DP),np.array(DP_het))))
DP_het = np.average(np.array(DP_het))
gene, anno, pph2, sift, lof = vepA.findVariantAnnotation(v, args, vepFieldNames)
return [v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT'], gene, anno, pph2,
sift, lof, v['AF'], v['AC'], AN, AB, DP, DP_het, Nproband_alt,
TU[0], TU[1], TU_m[0], TU_m[1], TU_f[0], TU_f[1],
mErr, mErr_o, mendErrorPercent], indivs_T, indivs_U
def doCaseControl(v, cases, controls, thresh):
""" Analyzes case/control data, counting the
number of reference and alternate alleles.
Parameters
----------
v: line of the VCF
case and control are hash tables with:
Key: individual id Value: gender
thresh is a hash table with:
Key: name of threshold Value: threshold
"""
# If filters on the line failed move on.
if not v:
return None
# Count the number of ref and alt alleles in cases and controls.
caseRefs_m = 0
caseAlts_m = 0
controlRefs_m = 0
controlAlts_m = 0
caseRefs_f = 0
caseAlts_f = 0
controlRefs_f = 0
controlAlts_f = 0
# Loop through all the individuals in the case hash table.
for indiv_id in cases:
# indiv_data is their GT:AD:DP:GQ:PL stats
indiv_data = v[indiv_id]
if indiv_data == None:
continue
# Apply filters and update counts Note: cases[indiv_id] is gender
if ProcessCC(indiv_data, thresh):
parFlag = filters.check_Hemizgyous(v['CHROM'], cases[indiv_id],
filters.inPar(v['POS']))
if cases[indiv_id] == 'male':
caseRefs_m, caseAlts_m = Counts(indiv_data, caseRefs_m,
caseAlts_m, parFlag)
elif cases[indiv_id] == 'female':
caseRefs_f, caseAlts_f = Counts(indiv_data, caseRefs_f,
caseAlts_f, parFlag)
# Loop through all indivs in the control hash table.
for indiv_id in controls:
# indiv_data is their GT:AD:DP:GQ:PL stats
indiv_data = v.get(indiv_id)
if indiv_data == None:
continue
# Apply filters and update counts Note: controls[indiv_id] is gender
if ProcessCC(indiv_data, thresh):
parFlag = filters.check_Hemizgyous(v['CHROM'], controls[indiv_id],
filters.inPar(v['POS']))
if controls[indiv_id] == 'male':
controlRefs_m, controlAlts_m = Counts(indiv_data,
controlRefs_m,
controlAlts_m, parFlag)
elif controls[indiv_id] == 'female':
controlRefs_f, controlAlts_f = Counts(indiv_data,
controlRefs_f,
controlAlts_f, parFlag)
caseAlts = caseAlts_f + caseAlts_m
caseRefs = caseRefs_f + caseRefs_m
controlAlts = controlAlts_f + controlAlts_m
controlRefs = controlRefs_f + controlRefs_m
if (caseAlts + controlAlts == 0) | (caseRefs + controlRefs == 0):
return None
AC = caseAlts + controlAlts
AN = AC + caseRefs + controlRefs
AF = AC / AN
if vepFieldNames:
gene, anno, pph2, sift, lof = vepA.findVariantAnnotation(
v, args, vepFieldNames)
else:
gene, anno, pph2, sift, lof = ('', '', '', '', '')
return [
v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT'], v['FILTER'],
v['VQSLOD'], gene, anno, pph2, sift, lof, AF, AC, AN, caseRefs,
caseAlts, controlRefs, controlAlts, caseRefs_m, caseAlts_m,
controlRefs_m, controlAlts_m, caseRefs_f, caseAlts_f, controlRefs_f,
controlAlts_f
]
def doCaseControl(v, cases, controls, thresh):
""" Analyzes case/control data, counting the
number of reference and alternate alleles.
Parameters
----------
v: line of the VCF
case and control are hash tables with:
Key: individual id Value: gender
thresh is a hash table with:
Key: name of threshold Value: threshold
"""
# If filters on the line failed move on.
if not v:
return None
# Count the number of ref and alt alleles in cases and controls.
caseRefs = 0
caseAlts = 0
controlRefs = 0
controlAlts = 0
# Loop through all the individuals in the case hash table.
for indiv_id in cases:
# indiv_data is their GT:AD:DP:GQ:PL stats
indiv_data = v[indiv_id]
if indiv_data == None:
continue
# Apply filters and update counts Note: cases[indiv_id] is gender
if ProcessCC(indiv_data, thresh):
parFlag = filters.check_Hemizgyous(v['CHROM'], cases[indiv_id], filters.inPar(v['POS']) )
caseRefs, caseAlts = Counts(indiv_data, caseRefs, caseAlts, parFlag)
# Loop through all indivs in the control hash table.
for indiv_id in controls:
# indiv_data is their GT:AD:DP:GQ:PL stats
indiv_data = v.get(indiv_id)
if indiv_data == None:
continue
# Apply filters and update counts Note: controls[indiv_id] is gender
if ProcessCC(indiv_data, thresh):
parFlag = filters.check_Hemizgyous(v['CHROM'], controls[indiv_id], filters.inPar(v['POS']) )
controlRefs, controlAlts = Counts(indiv_data, controlRefs, controlAlts, parFlag)
if (caseAlts + controlAlts == 0) | (caseRefs + controlRefs == 0):
return None
AC = caseAlts + controlAlts
AN = AC + caseRefs + controlRefs
AF = AC / AN
gene, anno, pph2, sift, lof = vepA.findVariantAnnotation(v, args, vepFieldNames)
return [v['CHROM'], v['POS'], v['ID'], v['REF'], v['ALT'], gene, anno, pph2, sift,
lof, AF, AC, AN, caseRefs, caseAlts, controlRefs, controlAlts]