def performGroupCalculations(variantObject, groupDict, basisGroup, mode=1, vcfOverride=False):
if variantObject == None or variantObject.poisoned:
return
# First find all the target alleles
targetAlleles = {}
if vcfOverride:
alleles = variantObject.alleles
else:
# First see if we can find a major allele with the people in basisGroup
alleleCounts = countingDict()
for i in groupDict[basisGroup]:
if variantObject.genotypes.has_key(i):
allele1 = variantObject.genotypes[i].allele1
allele2 = variantObject.genotypes[i].allele2
if allele1.text != None:
alleleCounts[allele1] += 1
if allele2.text != None:
alleleCounts[allele2] += 1
alleles = [x[0] for x in sorted(alleleCounts.iteritems(), key=lambda x: x[1])]
if mode > 0:
mode -= 1 # we're in 0-based coordinates, but -1 is still the same
if mode >= len(alleles) or mode < -len(alleles):
targetAllele = None
else:
targetAllele = variantObject.alleles[mode]
for groupID,samples in groupDict.iteritems():
if targetAllele == None:
variantObject.setAttribute(groupID + " AF","Masked") # the original group didn't have the allele, so we're masked
continue
allCount = 0
targetCount = 0
for i in samples:
if variantObject.genotypes.has_key(i):
allele1 = variantObject.genotypes[i].allele1
allele2 = variantObject.genotypes[i].allele2
if allele1 != None:
allCount += 1
if allele1 == targetAllele:
targetCount += 1
if allele2 != None:
allCount += 1
if allele2 == targetAllele:
targetCount += 1
if allCount == 0:
variantObject.setAttribute(groupID,None) # We had no data for this variant, so this thing is undefined
else:
variantObject.setAttribute(groupID,float(targetCount)/allCount)
def performGroupCalculations(self, groupDict, statisticDict, callback, tickInterval):
from dataModels.setupData import statistic
currentLine = 0
nextTick = tickInterval
targetAlleleGroups = {}
for s in statisticDict.itervalues():
if s.statisticType == statistic.ALLELE_FREQUENCY:
if s.parameters.has_key('alleleGroup'):
index = s.parameters['alleleMode']
if index >= 1:
index -= 1 # they'll specify 1 as the most frequent, but we're in 0-based computer land; -1 is still the same though
targetAlleleGroups[s.parameters['alleleGroup']] = s.parameters['alleleMode']
else:
targetAlleleGroups['vcf override'] = s.parameters['alleleMode']
if len(targetAlleleGroups) == 0: # nothing to calculate
return
for key in self.data.iterkeys():
if key == 'variant keys':
continue
variantObject = self.data[key]
currentLine += 1
if currentLine >= nextTick:
nextTick += tickInterval
self.dataConnection.commit()
if callback(): # abort?
return "ABORTED"
if variantObject == None or variantObject.poisoned:
continue
# First find all the target alleles
targetAlleles = {}
for group,mode in targetAlleleGroups.iteritems():
if group == 'vcf override':
alleles = variantObject.alleles
else:
# First see if we can find a major allele with the people in basisGroup
alleleCounts = countingDict()
for i in groupDict[group].samples:
if variantObject.genotypes.has_key(i):
allele1 = variantObject.genotypes[i].allele1
allele2 = variantObject.genotypes[i].allele2
if allele1.text != None:
alleleCounts[allele1] += 1
if allele2.text != None:
alleleCounts[allele2] += 1
alleles = [x[0] for x in sorted(alleleCounts.iteritems(), key=lambda x: x[1])]
if mode >= len(alleles) or mode < -len(alleles):
targetAlleles[group] = None
else:
targetAlleles[group] = variantObject.alleles[mode]
for statisticID,s in statisticDict.iteritems():
targetAllele = targetAlleles[s.parameters.get('alleleGroup','vcf override')]
if s.statisticType == statistic.ALLELE_FREQUENCY:
if targetAllele == None:
variantObject.setAttribute(statisticID,"Masked") # the original group didn't have the allele, so we're masked
continue
allCount = 0
targetCount = 0
for i in groupDict[s.parameters['group']].samples:
if variantObject.genotypes.has_key(i):
allele1 = variantObject.genotypes[i].allele1
allele2 = variantObject.genotypes[i].allele2
if allele1 != None:
allCount += 1
if allele1 == targetAllele:
targetCount += 1
if allele2 != None:
allCount += 1
if allele2 == targetAllele:
targetCount += 1
if allCount == 0:
variantObject.setAttribute(statisticID,None) # We had no data for this variant, so this thing is undefined
else:
variantObject.setAttribute(statisticID,float(targetCount)/allCount)
self.data[variantObject.name].attributes = variantObject.attributes # a way to force durus to acknowledge the change
self.data._p_note_change()
self.dataConnection.commit()
self.dataConnection.commit()