def openWriteBeagleFiles(self,
pedigreeFamilyData=None,
outputFnamePrefix=None):
"""
2013.05.02
The non-likelihood (unphased, trios, pairs) Beagle format:
I id sample1 sample1 sample2 sample2
A diabetes 1 1 2 2
M rs12082861 C C C C
M rs4912233 T C C C
M rs12732823 G A A A
M rs17451521 C C C C
M rs12033358 C T T T
The likelihood version is
marker alleleA alleleB 1000_709_1996093_GA_vs_524 1000_709_1996093_GA_vs_524 1000_709_1996093_GA_vs_524 1001_710_1995025_GA_vs_524 1001_710_1995025_GA_vs_524 1001_710_1995025_GA_vs_524 1002_711_2001039_GA_vs_524
Contig791:1086 C A 0.9693 0.0307 0.0000 0.6660 0.3338 0.0003 0.0000
Contig791:1649 G C 0.9406 0.0594 0.0000 0.9693 0.0307 0.0000 0.0000
Contig791:4084 A C 0.9980 0.0020 0.0000 0.9844 0.0156 0.0000 0.0000
The markers file has this format (markerID, position, alleleA, alleleB)
Contig791:1086 1086 C A
"""
sys.stderr.write(
"Opening beagle files (outputFnamePrefix =%s) to write ..." %
(outputFnamePrefix))
familySize2BeagleFileHandler = {}
familySize2SampleIDList = pedigreeFamilyData.familySize2SampleIDList
counter = 0
for familySize, sampleIDList in familySize2SampleIDList.items():
if familySize not in familySize2BeagleFileHandler:
tmpOutputFnamePrefix = '%s_familySize%s' % (outputFnamePrefix,
familySize)
writer = MatrixFile(path='%s.bgl' % (tmpOutputFnamePrefix),
mode='w',
delimiter=' ')
familySize2BeagleFileHandler[familySize] = writer
if familySize == 1:
headerRow = ['marker', 'alleleA', 'alleleB']
else:
headerRow = ['I', 'id']
for sampleID in sampleIDList:
if familySize == 1:
#likelihood format has sample name replicated three times, rather than 2 times
headerRow.extend([sampleID] * 3)
else:
headerRow.extend([sampleID] * 2)
writer.writeHeader(headerRow)
counter += 1
markersFile = MatrixFile(path='%s.markers' % (outputFnamePrefix),
mode='w',
delimiter=' ')
counter += 1
sys.stderr.write("%s files outputted.\n" % (counter))
return PassingData(
familySize2BeagleFileHandler=familySize2BeagleFileHandler,
markersFile=markersFile)
def run(self):
if self.debug:
import pdb
pdb.set_trace()
outputDir = os.path.split(self.outputFname)[0]
if outputDir and not os.path.isdir(outputDir):
os.makedirs(outputDir)
reader = VCFFile(inputFname=self.inputFname)
alignmentFile = pysam.Samfile(self.alignmentFilename, "rb")
writer = VCFFile(outputFname=self.outputFname, mode='w')
writer.metaInfoLs = reader.metaInfoLs
writer.header = reader.header
writer.writeMetaAndHeader()
statWriter = MatrixFile(self.missingStatFname, mode='w', delimiter='\t')
header = ["sampleID", "locusID", 'chromosome', 'start', 'stop', 'occurrence', 'missingReason', \
'fractionOfGoodRead', 'medianMapQ', 'totalNoOfReads']
statWriter.writeHeader(header)
counter = 0
real_counter = 0
minDepth = self.alignmentMedianDepth/self.alignmentDepthFold
maxDepth = self.alignmentMedianDepth*self.alignmentDepthFold
for vcfRecord in reader:
locusID = "%s_%s"%(vcfRecord.chromosome, vcfRecord.position)
alignedReadLs = alignmentFile.fetch(vcfRecord.chromosome, vcfRecord.position-1, vcfRecord.position+1) #start and end in fetch() are 0-based.
locusLowMapQData = self.returnLocusLowMapQualityIndicator(alignedReadLs=alignedReadLs,\
minMapQGoodRead=self.minMapQGoodRead, minFractionOfGoodRead=self.minFractionOfGoodRead)
locusLowMapQIndicator = locusLowMapQData.locusLowMapQIndicator
depth = locusLowMapQData.totalNoOfReads
if depth>=minDepth and depth <=maxDepth:
locusOutOfDepthIndicator = 0 #good
else:
locusOutOfDepthIndicator = 1
locusLowQualityIndicator = locusOutOfDepthIndicator + locusLowMapQIndicator
data_row = [self.sampleID, locusID, vcfRecord.chromosome, vcfRecord.position, vcfRecord.position,\
1, locusLowQualityIndicator, locusLowMapQData.fractionOfGoodRead, \
locusLowMapQData.medianMapQ, locusLowMapQData.totalNoOfReads]
statWriter.writerow(data_row)
if locusLowQualityIndicator>0:
real_counter += 1
#modify the VCF record
#get sample ID column, then set its genotype missing
vcfRecord.setGenotypeCallForOneSample(sampleID=self.sampleID, genotype="./.", convertGLToPL=True)
#2014.1.4 output VCF record
writer.writeVCFRecord(vcfRecord)
counter += 1
reader.close()
statWriter.close()
writer.close()
sys.stderr.write("%s (out of %s, %s) genotypes marked missing.\n"%(real_counter, counter, \
real_counter/float(counter)))
def run(self):
if self.debug:
import pdb
pdb.set_trace()
outputDir = os.path.split(self.outputFname)[0]
if outputDir and not os.path.isdir(outputDir):
os.makedirs(outputDir)
snp_pos2genotypeVectorLs = self.readInSNPID2GenotypeVectorLs(
self.inputFname).snp_pos2returnData
writer = MatrixFile(self.outputFname, mode='w', delimiter='\t')
header = [
'chromosome', 'position', 'noOfMatches', 'noOfTotal', 'concordance'
]
writer.writeHeader(header)
counter = 0
real_counter = 0
no_of_pairs = 0
snp_pos_ls = sorted(snp_pos2genotypeVectorLs)
for i in range(len(snp_pos_ls)):
counter += 1
key = snp_pos_ls[i]
chromosome, position = snp_pos_ls[i][:2]
genotypeVectorLs = snp_pos2genotypeVectorLs.get(key)
if len(genotypeVectorLs) > 1:
real_counter += 1
for k in range(0, len(genotypeVectorLs) - 1):
for l in range(k + 1, len(genotypeVectorLs)):
no_of_pairs += 1
noOfMatches = 0
noOfTotal = 0
genotypeVector0 = genotypeVectorLs[k]
genotypeVector1 = genotypeVectorLs[l]
for j in range(len(genotypeVector0)):
call1 = genotypeVector0[j]['GT']
call2 = genotypeVector1[j]['GT']
if call1 != 'NA' and call2 != 'NA':
noOfTotal += 1
if SNP.nt2number[call1] == SNP.nt2number[
call2]:
noOfMatches += 1
if noOfTotal > 0:
concordance = float(noOfMatches) / float(noOfTotal)
else:
concordance = -1
data_row = [
chromosome, position, noOfMatches, noOfTotal,
concordance
]
writer.writerow(data_row)
writer.close()
sys.stderr.write("%s (out of %s, %s) snps have >1 same-position entries. %s pairs.\n"%(real_counter, counter, \
real_counter/float(counter), no_of_pairs))
def run(self):
if self.debug:
import pdb
pdb.set_trace()
outputDir = os.path.split(self.outputFname)[0]
if outputDir and not os.path.isdir(outputDir):
os.makedirs(outputDir)
switchPointData = self.readInStats(inputFname=self.inputFname)
sys.stderr.write("Processing data ...")
writer = MatrixFile(self.outputFname, mode='w')
header = [
"maxSwitchFrequency", "genomeCovered", 'genomeCoveredFraction',
"noOfLoci", 'noOfLociFraction'
]
writer.writeHeader(header)
data_matrix = switchPointData.data_matrix
totalSpan = switchPointData.totalSpan
totalNoOfLoci = switchPointData.totalNoOfLoci
#sort it based on switchFrequency
data_matrix.sort(reverse=True)
maxSwitchFrequencyLs = []
cumulativeRegionSpanLs = []
cumulativeNoOfLociLs = []
for i in range(len(data_matrix)):
switchFrequency, regionSpan, noOfLoci = data_matrix[i]
maxSwitchFrequencyLs.append(switchFrequency)
if i == 0:
cumulativeRegionSpan = totalSpan - regionSpan
cumulativeNoOfLoci = totalNoOfLoci - noOfLoci
else:
cumulativeRegionSpan = cumulativeRegionSpanLs[i -
1] - regionSpan
cumulativeNoOfLoci = cumulativeNoOfLociLs[i - 1] - noOfLoci
cumulativeRegionSpanLs.append(cumulativeRegionSpan)
cumulativeNoOfLociLs.append(cumulativeNoOfLoci)
writer.writerow([switchFrequency, cumulativeRegionSpan, cumulativeRegionSpan/float(totalSpan),\
cumulativeNoOfLoci, cumulativeNoOfLoci/float(totalNoOfLoci)])
writer.close()
sys.stderr.write(".\n")
def run(self):
"""
"""
if self.debug:
import pdb
pdb.set_trace()
reader = MatrixFile(path=self.inputFname)
reader.constructColName2IndexFromHeader()
meanMendelErrorIndex = reader.getColIndexGivenColHeader(
"meanMendelError")
noOfLociIndex = reader.getColIndexGivenColHeader("sampled_base_count")
sumOfMendelErrorIndex = reader.getColIndexGivenColHeader(
"sumOfMendelError")
plinkPedigreeFile = PlinkPedigreeFile(path=self.pedigreeFname)
familyStructureData = plinkPedigreeFile.getFamilyStructurePlinkWay()
twoParentFamilyCountData = self.getNoOfFamiliesAndKidsGivenParentSetSize(noOfParents2FamilyData=familyStructureData.noOfParents2FamilyData, \
parentSetSize=2)
singleParentFamilyCountData = self.getNoOfFamiliesAndKidsGivenParentSetSize(noOfParents2FamilyData=familyStructureData.noOfParents2FamilyData, \
parentSetSize=1)
zeroParentFamilyCountData = self.getNoOfFamiliesAndKidsGivenParentSetSize(noOfParents2FamilyData=familyStructureData.noOfParents2FamilyData, \
parentSetSize=0)
writer = MatrixFile(self.outputFname, mode='w', delimiter='\t')
header = ["ID", "noOfTotalLoci", \
"noOfTwoParentFamilies", "noOfParentsInTwoParentFamilies", "noOfKidsInTwoParentFamilies", "noOfIndividualsInTwoParentFamilies", \
"noOfSingleParentFamilies", "noOfParentsInSingleParentFamilies", "noOfKidsInSingleParentFamilies", "noOfIndividualsInSingleParentFamilies", \
"noOfZeroParentFamilies", "noOfParentsInZeroParentFamilies", "noOfKidsInZeroParentFamilies", "noOfIndividualsInZeroParentFamilies", \
"noOfTotalMendelErrors", \
"noOfMendelErrorsPerLocusPerNuclearFamily", "noOfMendelErrorsPerNuclearFamily"]
writer.writeHeader(header)
for row in reader:
meanMendelError = float(row[meanMendelErrorIndex])
noOfLoci = int(row[noOfLociIndex])
sumOfMendelError = int(row[sumOfMendelErrorIndex])
noOfNuclearFamilies = twoParentFamilyCountData.noOfFamilies
if noOfNuclearFamilies > 0:
noOfMendelErrorsPerLocusPerNuclearFamily = meanMendelError / float(
noOfNuclearFamilies)
noOfMendelErrorsPerNuclearFamily = sumOfMendelError / float(
noOfNuclearFamilies)
else:
noOfMendelErrorsPerLocusPerNuclearFamily = -1
noOfMendelErrorsPerNuclearFamily = -1
data_row = [row[0], noOfLoci, \
noOfNuclearFamilies, twoParentFamilyCountData.noOfParents, twoParentFamilyCountData.noOfKids, \
twoParentFamilyCountData.noOfIndividuals,\
singleParentFamilyCountData.noOfFamilies, singleParentFamilyCountData.noOfParents, singleParentFamilyCountData.noOfKids,\
singleParentFamilyCountData.noOfIndividuals,\
zeroParentFamilyCountData.noOfFamilies, zeroParentFamilyCountData.noOfParents, zeroParentFamilyCountData.noOfKids,\
zeroParentFamilyCountData.noOfIndividuals,\
sumOfMendelError, \
noOfMendelErrorsPerLocusPerNuclearFamily,noOfMendelErrorsPerNuclearFamily ]
writer.writerow(data_row)
plinkPedigreeFile.close()
reader.close()
writer.close()