def calculate_FETpvalues_readcounts(vcffile,poolstatustable,poolsize):
maxMAF = 0.01; # maximum minor allele frequency, variants above this are not used
poolDX = determine_poolphenotype(vcffile,poolstatustable); # phenotype status for each pool
variants =0;
File = open(vcffile);
for line in File:
if line[0] == '#': continue;
variants +=1;
variant = line.strip().split('\t');
samples = len(variant)-9;
chrom = variant[0]; position = int(variant[1]); refallele = variant[3]; varalleles = variant[4].split(',');
if len(varalleles) ==2: triallelic=1;
else: triallelic=0;
if len(varalleles) > 2:
print >>sys.stderr, '##triallelic',chrom,position,refallele,varalleles;
continue; # ignore quad or greater allelic variants for now
#if triallelic ==0 or position != 64527465: continue;
H0 = 0.0; H1 = 0.0; H2=0.0; D0 = 0.0; D1 = 0.0; D2=0.0;
casepools=0; controlpools=0;
for i in xrange(samples):
if poolDX[i] == -1: continue; # ignore pool for case control analysis
try:
counts = variant[i+9].split(':');
readsf = counts[2].split(','); readsr = counts[3].split(',');
total = int(readsf[0]) + int(readsf[1]) + int(readsr[0]) + int(readsr[1]);
if triallelic==1: total += int(readsf[2])+int(readsr[2]); # tri-allelic
alt = (float(readsf[1]) + float(readsr[1]))*poolsize; alt /= (total+0.0001);
if triallelic==1: alt2 = (float(readsf[2]) + float(readsr[2]))*poolsize; alt2 /= (total+0.0001);
else: alt2=0.0;
sf = float(2*total)/(2*total + poolsize);
if poolDX[i][0] == 0:
#H0 += poolsize; H1 = alt;
H0 += sf*float(poolsize); H1 += sf*alt; H2 += sf*alt2; controlpools +=1;
# 0/1 bit of allelecounts stores case-control status...
elif poolDX[i][0] >= 1:
D0 += sf*float(poolsize); D1 += sf*alt; D2 += sf*alt2; casepools +=1;
#D0 += poolsize; D1 += alt;
except IndexError: print 'Exception',i,samples,variant;
#if H1 + D1 < 10 or (H0-H1 + D0-D1 <10): lowcountvariants +=1; continue;
pvalue = [0,0,0]; pvalue2=[0,0,0];
if H1 + D1 >= 4: pvalue = fet(int(round(H1,0)),int(round(H0,0)),int(round(D1,0)),int(round(D0,0)));
print '%0.2f %.2f:%.2f %.2f:%.2f' %(pvalue[0],H0,H1,D0,D1),#'Control',float(H1)/(H0+0.001),'Case',float(D1)/(D0+0.001),
if triallelic ==1 and H2+D2 >=4:
pvalue2 = fet(int(round(H2,0)),int(round(H0,0)),int(round(D2,0)),int(round(D0,0)));
print '2nd-allele %0.2f %.2f:%.2f %.2f:%.2f' %(pvalue2[0],H0,H2,D0,D2),
print '%3s %9s %10s %10s %20s ' %(variant[0],variant[1],variant[3],variant[4],variant[5]+':'+variant[6]),
print variant[7];
def calculate_FETpvalues_allelecounts(vcffile, poolstatustable, poolsize):
poolDX = determine_poolphenotype(vcffile, poolstatustable)
# phenotype status for each pool
poolDX_random = []
missing = []
for i in xrange(len(poolDX)):
if poolDX[i][0] == -1: missing.append(i)
poolDX_random.append(poolDX[i])
"""
"""
random.shuffle(poolDX_random)
#print missing,len(poolDX_random);
j = 0
for i in xrange(len(poolDX_random)):
if poolDX_random[i][0] == -1 and j < len(missing):
#print j,poolDX_random[i],i;
poolDX_random[i] = poolDX_random[missing[j]]
poolDX_random[missing[j]] = [-1]
j += 1
for i in xrange(len(poolDX)):
poolDX[i] = poolDX_random[i]
print >> sys.stderr, poolDX_random
variants = 0
trivariants = 0
File = open(vcffile)
for line in File:
if line[0] == '#': continue
variant = line.strip().split('\t')
samples = len(variant) - 9
chrom = variant[0]
position = int(variant[1])
refallele = variant[3]
varalleles = variant[4].split(',')
if len(varalleles) == 2:
triallelic = 1
trivariants += 1
else:
triallelic = 0
variants += 1
if len(varalleles) >= 3 or len(varalleles) == 4:
#print >>sys.stderr, '##triallelic',chrom,position,refallele,varalleles,variant[2];
continue
# ignore multi-allelic variants for now
## healthy (controls), D (disease), E (early onset), DC (diabetic complications)
H0 = 0.00001
H1 = 0.0
H2 = 0.0
D0 = 0.00001
D1 = 0.0
D2 = 0.0
E0 = 0.00001
E1 = 0.0
E2 = 0
DC0 = 0.00001
DC1 = 0
DC2 = 0
casepools = 0
controlpools = 0
for i in xrange(samples):
if poolDX_random[i][0] == -1:
continue
# ignore pool for case control analysis
try:
genotypes = variant[i + 9].split(':')
if triallelic == 0:
MLAC = int(genotypes[0])
#meanAC = float(genotypes[2]);
QAC = float(genotypes[1])
MLAC2 = 0
else:
MLAC = int(genotypes[0].split(',')[0])
#meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
MLAC2 = int(genotypes[0].split(',')[1])
#if genotypes[3] == '-inf': varAF = -0.1*QAC;
#else: varAF = float(genotypes[3]);
if poolDX_random[i][0] == 0:
H0 += poolsize
H1 += MLAC
H2 += MLAC2
elif poolDX_random[i][0] >= 1:
D0 += poolsize
D1 += MLAC
D2 += MLAC2
if poolDX_random[i][0] == 2:
E0 += poolsize
E1 += MLAC
E2 += MLAC2
if len(poolDX_random[i]) >= 2 and poolDX_random[i][1] == 3:
DC0 += poolsize
DC1 += MLAC
DC2 += MLAC2
except IndexError:
print 'Exception', i, samples, variant
nopvalue_calc = 0
if (H1 + D1 >= 5):
pvalue = fet(int(round(H1, 0)), int(round(H0, 0)),
int(round(D1, 0)), int(round(D0, 0)))
else:
pvalue = [0, 0, 0]
nopvalue_calc = 1
## calculate p-values between early onset and controls
if (H1 + E1 >= 5):
pvalue1 = fet(int(round(H1, 0)), int(round(H0, 0)),
int(round(E1, 0)), int(round(E0, 0)))
else:
pvalue1 = [0, 0, 0]
## calculate p-value between early onset and late-onset
if (D1 + E1 >= 5):
pvalue2 = fet(int(round(D1, 0)), int(round(D0, 0)),
int(round(E1, 0)), int(round(E0, 0)))
else:
pvalue2 = [0, 0, 0]
if nopvalue_calc == 1: continue
print '%0.2f %.1f/%.1f %.1f/%.1f %.1f/%.1f %.1f/%.1f' % (
pvalue[0], H1, H0, D1, D0, E1, E0, DC1,
DC0), #'Control',float(H1)/(H0+0.001),'Case',float(D1)/(D0+0.001),
print '%0.2f %0.2f' % (pvalue1[0], pvalue2[0]),
print '%0.4f %0.4f %0.4f %0.4f' % (H1 / H0, D1 / D0, E1 / E0,
DC1 / DC0),
if triallelic == 1:
if H2 + D2 >= 4:
pvalue_tri = fet(int(round(H2, 0)), int(round(H0, 0)),
int(round(D2, 0)), int(round(D0, 0)))
else:
pvalue_tri = [0, 0, 0]
print 'TRIALLELIC:%0.2f:%.1f:%.1f' % (pvalue_tri[0], H2, D2),
else:
print '-',
print '%3s %9s %s %10s %10s %8s %10s' % (
variant[0], variant[1], variant[2], variant[3], variant[4],
variant[5], variant[6]),
print variant[7],
if PRINTGENOTYPES == 1:
for i in xrange(samples):
if poolDX_random[i][0] == -1:
continue
# ignore pool for case control analysis
genotypes = variant[i + 9].split(':')
print variant[i + 9],
#MLAC = int(genotypes[0]); meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
#if genotypes[3] == '-inf': varAF = -0.1*QAC;
#else: varAF = float(genotypes[3]);
#print '%2d:%0.2f:%0.2f' %(MLAC,meanAC,math.sqrt(pow(10,varAF))),
print
print >> sys.stderr, "variants evaluated", variants, "triallelic or more", trivariants
def calculate_FETpvalues_allelecounts(vcffile,poolstatustable):
poolDX = determine_poolphenotype(vcffile,poolstatustable); # phenotype status for each pool
variants =0; trivariants =0;
File = open(vcffile);
for line in File:
if line[0] == '#': continue;
variant = line.strip().split('\t');
samples = len(variant)-9;
chrom = variant[0]; position = int(variant[1]); refallele = variant[3]; varalleles = variant[4].split(',');
if len(varalleles) ==2: triallelic=1; trivariants +=1;
else: triallelic=0; variants +=1;
if len(varalleles) >= 3 or len(varalleles) ==4:
#print >>sys.stderr, '##triallelic',chrom,position,refallele,varalleles,variant[2];
continue; # ignore multi-allelic variants for now
## healthy (controls), D (disease), E (early onset), DC (diabetic complications)
H0 = 0.0; H1 = 0.0; H2 = 0.0; D0 = 0.0; D1 = 0.0; D2 = 0.0; E0 = 0.0; E1= 0.0; E2 = 0;
casepools=0; controlpools=0;
for i in xrange(samples):
if poolDX[i][0] == -1 or variant[i+9].split(':')[0] == '.': continue; # ignore pool for case control analysis
try:
poolsize = poolDX[i][1]
genotypes = variant[i+9].split(':');
if triallelic==0:
MLAC = int(genotypes[0].split(',')[1]); #meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
QAC = float(genotypes[1]);
MLAC2 = 0;
else:
MLAC = int(genotypes[0].split(',')[1]); #meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
QAC = float(genotypes[1]);
MLAC2 = int(genotypes[0].split(',')[2]); #meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
#else: varAF = float(genotypes[3]);
if poolDX[i][0] == 0: H0 += poolsize; H1 += MLAC; H2 += MLAC2;
elif poolDX[i][0] >= 1: D0 +=poolsize; D1 += MLAC; D2 += MLAC2;
if poolDX[i][0] == 2: E0 += poolsize; E1 += MLAC; E2 += MLAC2;
except IndexError: print 'Exception',i,samples,genotypes;
if (H1 + D1 >= 4): pvalue = fet(int(round(H1,0)),int(round(H0,0)),int(round(D1,0)),int(round(D0,0)));
else: pvalue = [0,0,0];
## calculate p-values between early onset and controls
if (H1+E1 >=4): pvalue1 = fet(int(round(H1,0)),int(round(H0,0)),int(round(E1,0)),int(round(E0,0)));
else: pvalue1 = [0,0,0];
## calculate p-value between early onset and late-onset
if (D1+E1 >=4): pvalue2 = fet(int(round(D1,0)),int(round(D0,0)),int(round(E1,0)),int(round(E0,0)));
else: pvalue2 = [0,0,0];
print '%0.2f %.1f/%.1f %.1f/%.1f %.1f/%.1f %.1f/%.1f' %(pvalue[0],H1,H0,D1,D0,E1,E0,DC1,DC0),#'Control',float(H1)/(H0+0.001),'Case',float(D1)/(D0+0.001),
print '%0.2f %0.2f' %(pvalue1[0],pvalue2[0]),
print '%0.4f %0.4f %0.4f %0.4f' %(H1/(H0+epsilon),D1/(D0+epsilon),E1/(E0+epsilon),DC1/(DC0+epsilon)),
#sys.exit()
if triallelic ==1:
if H2 + D2 >=4: pvalue_tri = fet(int(round(H2,0)),int(round(H0,0)),int(round(D2,0)),int(round(D0,0)));
else: pvalue_tri = [0,0,0];
print 'TRIALLELIC:%0.2f:%.1f:%.1f' %(pvalue_tri[0],H2,D2),
else: print '-',
print '%3s %9s %s %10s %10s %8s %10s' %(variant[0],variant[1],variant[2],variant[3],variant[4],variant[5],variant[6]),
print variant[7],
print;
print >>sys.stderr, "variants evaluated",variants,"triallelic or more",trivariants;
def calculate_FETpvalues_new(vcffile, poolstatustable, poolsize):
maxMAF = 0.01
# maximum minor allele frequency, variants above this are not used
poolDX = []
# phenotype status for each pool
Variants = []
File = open(vcffile)
for line in File:
if line[0] == '#' and line[1] == '#': continue
variant = line.strip().split('\t')
if variant[0] == '#CHROM':
for i in xrange(9, len(variant)):
sampleid = variant[i].split('/')[-1].split('.')[0]
#print sampleid;
try:
status = poolstatustable[sampleid]
except KeyError:
status = -1
poolDX.append(status)
print >> sys.stderr, status,
#samples = len(variant)-9; print 'samples',samples,len(variant);
print >> sys.stderr
continue
samples = len(variant) - 9
chrom = variant[0]
position = int(variant[1])
refallele = variant[3]
varalleles = variant[4].split(',')
if len(varalleles) == 2: triallelic = 1
else: triallelic = 0
if len(varalleles) > 2:
continue
# ignore multi-allelic variants for now
#if triallelic ==0 or position != 64527465: continue;
H0 = 0.0
H1 = 0.0
H2 = 0.0
D0 = 0.0
D1 = 0.0
D2 = 0.0
casepools = 0
controlpools = 0
for i in xrange(samples):
if poolDX[i] == -1:
continue
# ignore pool for case control analysis
try:
counts = variant[i + 9].split(':')
readsf = counts[2].split(',')
readsr = counts[3].split(',')
total = int(readsf[0]) + int(readsf[1]) + int(readsr[0]) + int(
readsr[1])
if triallelic == 1:
total += int(readsf[2]) + int(readsr[2])
# tri-allelic
alt = (float(readsf[1]) + float(readsr[1])) * poolsize
alt /= (total + 0.0001)
if triallelic == 1:
alt2 = (float(readsf[2]) + float(readsr[2])) * poolsize
alt2 /= (total + 0.0001)
else:
alt2 = 0.0
sf = float(2 * total) / (2 * total + poolsize)
if poolDX[i] == 0:
#H0 += poolsize; H1 = alt;
H0 += sf * float(poolsize)
H1 += sf * alt
H2 += sf * alt2
controlpools += 1
# 0/1 bit of allelecounts stores case-control status...
elif poolDX[i] >= 1:
D0 += sf * float(poolsize)
D1 += sf * alt
D2 += sf * alt2
casepools += 1
#D0 += poolsize; D1 += alt;
except IndexError:
print 'Exception', i, samples, variant
#if H1 + D1 < 10 or (H0-H1 + D0-D1 <10): lowcountvariants +=1; continue;
pvalue = [0, 0, 0]
pvalue2 = [0, 0, 0]
if H1 + D1 >= 4:
pvalue = fet(int(round(H1, 0)), int(round(H0, 0)),
int(round(D1, 0)), int(round(D0, 0)))
print '%0.2f %.2f:%.2f %.2f:%.2f' % (
pvalue[0], H0, H1, D0,
D1), #'Control',float(H1)/(H0+0.001),'Case',float(D1)/(D0+0.001),
if triallelic == 1 and H2 + D2 >= 4:
pvalue2 = fet(int(round(H2, 0)), int(round(H0, 0)),
int(round(D2, 0)), int(round(D0, 0)))
print '2nd-allele %0.2f %.2f:%.2f %.2f:%.2f' % (pvalue2[0], H0, H2,
D0, D2),
print '%3s %9s %10s %10s %20s ' % (variant[0], variant[1], variant[3],
variant[4],
variant[5] + ':' + variant[6]),
print variant[7]
def calculate_FETpvalues(vcffile, poolDX, poolsize):
File = open(vcffile)
for line in File:
if line[0] == '#' and line[1] == '#': continue
variant = line.strip().split()
if variant[0] == '#CHROM':
samples = len(variant) - 9
#samplelist = variant[9+offset:]; samples = len(samplelist);
#print samplelist,samples;
continue
if 'SNP' in variant[2] or 'INDEL' in variant[2]:
H0 = 0
H1 = float(0)
D0 = 0
D1 = float(0)
allelecounts = []
allelecountsD = []
poolsH = 0
poolsD = 0
for i in xrange(samples):
if poolDX[i] == -1:
continue
# ignore pool for case control analysis
counts = variant[i + 9].split(':')
total = int(counts[0]) + int(counts[1])
#if total < 240: continue;
alt = float(counts[1]) * poolsize
alt /= (total + 0.01)
#if alt < 0.5: alt = 0;
#if alt > 0.5 and alt < 1: alt = 1;
if poolDX[i] == 0:
H0 += poolsize
H1 += alt
allelecounts.append([0, 0,
int(counts[0]),
int(counts[1])])
poolsH += 1
# 0/1 bit of allelecounts stores case-control status...
elif poolDX[i] == 1:
D0 += poolsize
D1 += alt
allelecounts.append([1, 1,
int(counts[0]),
int(counts[1])])
poolsD += 1
# print counts,total,alt;
if H1 + D1 >= 5:
#print allelecounts;
allelecounts.sort()
HAF = estimate_AF(allelecounts, poolsH, poolsize / 2, 0.001, 0)
DAF = estimate_AF(allelecounts, poolsD, poolsize / 2, 0.001,
poolsH)
"""
for p in xrange(HAF[2],HAF[3]): print '%3d %2.2f ' %(p,HAF[0][p]),
print 'Binomial maxll',HAF[1],HAF[0][HAF[1]];
for p in xrange(DAF[2],DAF[3]): print '%3d %2.2f ' %(p,DAF[0][p]),
print 'Binomial maxll',DAF[1],DAF[0][DAF[1]];
"""
pvalue = fet(int(round(H1, 0)), int(H0), int(round(D1, 0)),
int(D0))
if pvalue[0] < -2:
pvalueperm = probabilisticFET(allelecounts, poolsH, poolsD,
poolsize)
# output from allele frequency estimation
print 'PERM',
else:
pvalueperm = 1
print 'FET',
print math.log(
pvalueperm, 10
), pvalue[0], variant[0], variant[1], variant[2], variant[
3], variant[4], variant[5], variant[6], variant[7],
print '%2.1f %2.1f %2.1f %2.1f' % (
H0, H1, D0, D1
), #'Control',float(H1)/(H0+0.001),'Case',float(D1)/(D0+0.001),
if pvalue < 0.001: print 'LOW'
else: print
def calculate_FETpvalues_allelecounts(vcffile,poolstatustable,poolsize,VCFfile2):
[SAMPLES,VARLIST,SAMPLELIST] = read_phase3_VCF(VCFfile2);
poolDX = determine_poolphenotype(vcffile,poolstatustable); # phenotype status for each pool
variants =0; trivariants =0;
File = open(vcffile);
for line in File:
if line[0] == '#': continue;
variant = line.strip().split('\t');
samples = len(variant)-9;
chrom = variant[0]; position = int(variant[1]); refallele = variant[3]; varalleles = variant[4].split(',');
if len(varalleles) ==2: triallelic=1; trivariants +=1;
else: triallelic=0; variants +=1;
if len(varalleles) >= 3 or len(varalleles) ==4:
#print >>sys.stderr, '##triallelic',chrom,position,refallele,varalleles,variant[2];
continue; # ignore multi-allelic variants for now
## healthy (controls), D (disease), E (early onset), DC (diabetic complications)
H0 = 0.0; H1 = 0.0; H2 = 0.0; D0 = 0.0; D1 = 0.0; D2 = 0.0; E0 = 0.0; E1= 0.0; E2 = 0; DC0 = 0; DC1 = 0; DC2 = 0;
casepools=0; controlpools=0;
for i in xrange(samples):
if poolDX[i][0] == -1: continue; # ignore pool for case control analysis
try:
genotypes = variant[i+9].split(':');
if triallelic==0:
MLAC = int(genotypes[0]); #meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
MLAC2 = 0;
else:
MLAC = int(genotypes[0].split(',')[0]); #meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
MLAC2 = int(genotypes[0].split(',')[1]);
#if genotypes[3] == '-inf': varAF = -0.1*QAC;
#else: varAF = float(genotypes[3]);
if poolDX[i][0] == 0: H0 += poolsize; H1 += MLAC; H2 += MLAC2;
elif poolDX[i][0] >= 1: D0 +=poolsize; D1 += MLAC; D2 += MLAC2;
if poolDX[i][0] == 2: E0 += poolsize; E1 += MLAC; E2 += MLAC2;
if len(poolDX[i]) >=2 and poolDX[i][1] == 3: DC0 += poolsize; DC1 += MLAC; DC2 += MLAC2;
except IndexError: print 'Exception',i,samples,variant;
## calculate p-value between cases and controls
if (H1 + D1 >= 4): pvalue = fet(int(round(H1,0)),int(round(H0,0)),int(round(D1,0)),int(round(D0,0)));
else: pvalue = [0,0,0];
## calculate p-values between early onset and controls
if (H1+E1 >=4): pvalue1 = fet(int(round(H1,0)),int(round(H0,0)),int(round(E1,0)),int(round(E0,0)));
else: pvalue1 = [0,0,0];
## calculate p-value between early onset and late-onset
if (D1+E1 >=4): pvalue2 = fet(int(round(D1,0)),int(round(D0,0)),int(round(E1,0)),int(round(E0,0)));
else: pvalue2 = [0,0,0];
if (chrom,position,refallele,varalleles[0]) in VARLIST:
VAR2 = VARLIST[(chrom,position,refallele,varalleles[0])];
print 'foundvar',VAR2,
E11 = E1 - VAR2[1];
E01 = E0 - VAR2[1] - VAR2[0];
if E11 < 0: E11 = 0;
D11 = D1 - VAR2[1]-VAR2[3];
if D11 < 0: D11 = 0;
D01 = D0 - VAR2[1] - VAR2[0] - VAR2[2] -VAR2[3];
H01=H0; H11 = H1;
if pvalue[0] < -0.5: ## low p-values
new_pvalue = fet(int(round(H1,0)),int(round(H0,0)),int(round(D11,0)),int(round(D01,0)));
else: new_pvalue = pvalue;
if pvalue[0] < -0.5: ## low p-values
new_pvalue1 = fet(int(round(H1,0)),int(round(H0,0)),int(round(E11,0)),int(round(E01,0)));
else: new_pvalue1 = pvalue1;
print ' %0.2f %0.2f %d/%d %d/%d %d/%d ' %(new_pvalue[0],new_pvalue1[0],int(H1),int(H0),int(D11),int(D01),int(E11),int(E01));
print 'corr-new %0.2f %0.2f %0.4f %0.4f %0.4f ' %(new_pvalue[0],new_pvalue1[0],H11/H01,D11/D01,E11/E01),
else:
#print 'missing'; ## 88 for early onset, 218 for late onset
#E11 = E1; E01 = E0-88; D11 = D1; D01 = D0-218-88;
#print 'corrected -\t-\t
H11 = H1; H01 = H0; D11 = D1; D01 = D0; E11 = E1; E01 = E0;
print 'corr-orig %0.2f %0.2f %0.4f %0.4f %0.4f ' %(pvalue[0],pvalue1[0],H11/H01,D11/D01,E11/E01),
#print 'corrected -\t-\t-\t-',
print '%0.2f %0.2f %.1f/%.1f %.1f/%.1f %.1f/%.1f' %(pvalue[0],pvalue1[0],H1,H0,D1,D0,E1,E0),#'Control',float(H1)/(H0+0.001),'Case',float(D1)/(D0+0.001),
#print '%0.2f %0.2f' %(pvalue1[0],pvalue2[0]),
#print '%0.4f %0.4f %0.4f %0.4f' %(H1/H0,D1/D0,E1/E0,DC1/DC0),
print '%0.4f %0.4f %0.4f' %(H1/H0,D1/D0,E1/E0),
if triallelic ==1:
if H2 + D2 >=4: pvalue_tri = fet(int(round(H2,0)),int(round(H0,0)),int(round(D2,0)),int(round(D0,0)));
else: pvalue_tri = [0,0,0];
print 'TRIALLELIC:%0.2f:%.1f:%.1f' %(pvalue_tri[0],H2,D2),
else: print '-',
print '%3s %9s %s %10s %10s %8s %10s' %(variant[0],variant[1],variant[2],variant[3],variant[4],variant[5],variant[6]),
print variant[7],
if PRINTGENOTYPES ==1:
for i in xrange(samples):
if poolDX[i][0] == -1: continue; # ignore pool for case control analysis
genotypes = variant[i+9].split(':');
print variant[i+9],
#MLAC = int(genotypes[0]); meanAC = float(genotypes[2]); QAC = float(genotypes[1]);
#if genotypes[3] == '-inf': varAF = -0.1*QAC;
#else: varAF = float(genotypes[3]);
#print '%2d:%0.2f:%0.2f' %(MLAC,meanAC,math.sqrt(pow(10,varAF))),
print;
print >>sys.stderr, "variants evaluated",variants,"triallelic or more",trivariants;