def dumpReadInfo(referenceFile,
bam,
minQual,
outputCSV,
snpsFile,
log,
printOnly=False,
verbose=True,
force=False):
if (not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing T->C per reads position for file " + bam,
file=log)
else:
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
outputFile = SlamSeqWriter(outputCSV)
#Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome)
for read in readIterator:
outputFile.write(read)
outputFile.close()
def statsComputeOverallRates(referenceFile,
bam,
minBaseQual,
outputCSV,
outputPDF,
log,
printOnly=False,
verbose=True,
force=False):
if (not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing overall rates for file " + bam, file=log)
else:
# Init
totalRatesFwd = [0] * 25
totalRatesRev = [0] * 25
tcCount = [0] * 100
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minBaseQual)
for read in readIterator:
# Compute rates for current read
rates = read.conversionRates
# Get T -> C conversions for current read
tc = read.tcCount
tcCount[tc] += 1
# Add rates from read to total rates
if (read.direction == ReadDirection.Reverse):
totalRatesRev = sumLists(totalRatesRev, rates)
else:
totalRatesFwd = sumLists(totalRatesFwd, rates)
# Print rates in correct format for plotting
fo = open(outputCSV, "w")
print("# slamdunk rates v" + __version__, file=fo)
printRates(totalRatesFwd, totalRatesRev, fo)
fo.close()
if (not checkStep([bam, referenceFile], [outputPDF], force)):
print("Skipped computing overall rate pdfs for file " + bam, file=log)
else:
#f = tempfile.NamedTemporaryFile(delete=False)
#print(removeExtension(basename(bam)), outputCSV, sep='\t', file=f)
#f.close()
callR(getPlotter("compute_overall_rates") + " -f " + outputCSV +
" -n " + removeExtension(os.path.basename(bam)) + " -O " +
outputPDF,
log,
dry=printOnly,
verbose=verbose)
def statsComputeOverallRates(referenceFile, bam, minBaseQual, outputCSV, outputPDF, log, printOnly=False, verbose=True, force=False):
if(not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing overall rates for file " + bam, file=log)
else:
# Init
totalRatesFwd = [0] * 25
totalRatesRev = [0] * 25
tcCount = [0] * 100
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minBaseQual)
for read in readIterator:
# Compute rates for current read
rates = read.conversionRates
# Get T -> C conversions for current read
tc = read.tcCount
tcCount[tc] += 1
# Add rates from read to total rates
if(read.direction == ReadDirection.Reverse):
totalRatesRev = sumLists(totalRatesRev, rates)
else:
totalRatesFwd = sumLists(totalRatesFwd, rates)
# Print rates in correct format for plotting
fo = open(outputCSV, "w")
print("# slamdunk rates v" + __version__, file=fo)
printRates(totalRatesFwd, totalRatesRev, fo)
fo.close()
if(not checkStep([bam, referenceFile], [outputPDF], force)):
print("Skipped computing overall rate pdfs for file " + bam, file=log)
else:
#f = tempfile.NamedTemporaryFile(delete=False)
#print(removeExtension(basename(bam)), outputCSV, sep='\t', file=f)
#f.close()
callR(getPlotter("compute_overall_rates") + " -f " + outputCSV + " -n " + removeExtension(os.path.basename(bam)) + " -O " + outputPDF, log, dry=printOnly, verbose=verbose)
def evaluateReads(bam, referenceFile, bed, outputFile, mainOutput):
print("Run " + bam)
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
chromosomes = testFile.getChromosomes()
bedTree = bedToIntervallTree(bed)
#evalHist = [0] *
outFile = open(outputFile, "w")
print("read.name", "read.chromosome", "read.startRefPos", "sim.utr", "read.utr", "sim.tcCount", "read.tcCount", sep = "\t", file=outFile)
total = 0
correct = 0
correcPosWrongTC = 0
wrongPos = 0
minBaseQual = 0
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minBaseQual)
for read in readIterator:
total += 1
simInfo = read.name.split("_")
utrSim = simInfo[0]
tcCountSim = int(simInfo[2])
utrFound = None
if read.chromosome in bedTree:
overlaps = list(bedTree[read.chromosome][read.startRefPos:read.endRefPos])
if len(overlaps) > 0:
utrFound = overlaps[0].data
if utrFound == utrSim:
if tcCountSim == read.tcCount:
correct += 1
else:
correcPosWrongTC += 1
else:
wrongPos += 1
print(read.name, read.chromosome, read.startRefPos, utrSim, utrFound, tcCountSim, read.tcCount, sep = "\t", file=outFile)
print(correct * 100.0 / total, correcPosWrongTC * 100.0 / total, wrongPos * 100.0 / total, total)
def genomewideReadSeparation(referenceFile, snpsFile, bam, minBaseQual,
outputBAMPrefix, conversionThreshold, log):
ref = pysam.FastaFile(referenceFile)
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
samFile = pysam.AlignmentFile(bam, "rb")
chromosomes = testFile.getChromosomes()
backgroundReadFileName = outputBAMPrefix + "_backgroundReads.bam"
tcReadFileName = outputBAMPrefix + "_TCReads.bam"
backgroundReadFile = pysam.AlignmentFile(backgroundReadFileName,
"wb",
template=samFile)
tcReadFile = pysam.AlignmentFile(tcReadFileName, "wb", template=samFile)
tcReadDict = dict()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minBaseQual,
conversionThreshold)
for read in readIterator:
if (read.isTcRead):
tcReadDict[read.name] = 0
for read in samFile.fetch():
if read.query_name in tcReadDict:
tcReadFile.write(read)
else:
backgroundReadFile.write(read)
backgroundReadFile.close()
tcReadFile.close()
pysamIndex(backgroundReadFileName)
pysamIndex(tcReadFileName)
def dumpReadInfo(referenceFile, bam, minQual, outputCSV, snpsFile, log, printOnly=False, verbose=True, force=False):
if(not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing T->C per reads position for file " + bam, file=log)
else:
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
outputFile = SlamSeqWriter(outputCSV)
#Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome)
for read in readIterator:
outputFile.write(read)
outputFile.close()
def genomewideConversionRates(referenceFile, snpsFile, bam, minBaseQual,
outputBedGraphPrefix, conversionThreshold,
coverageCutoff, log):
ref = pysam.FastaFile(referenceFile)
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
chromosomes = testFile.getChromosomes()
bedGraphInfo = re.sub("_slamdunk_mapped.*", "",
basename(outputBedGraphPrefix))
print(bedGraphInfo)
fileBedGraphRatesPlus = open(
outputBedGraphPrefix + "_TC_rates_genomewide.bedGraph", 'w')
fileBedGraphRatesMinus = open(
outputBedGraphPrefix + "_AG_rates_genomewide.bedGraph", 'w')
fileBedGraphCoveragePlus = open(
outputBedGraphPrefix + "_coverage_plus_genomewide.bedGraph", 'w')
fileBedGraphCoverageMinus = open(
outputBedGraphPrefix + "_coverage_minus_genomewide.bedGraph", 'w')
fileBedGraphTCConversions = open(
outputBedGraphPrefix + "_TC_conversions_genomewide.bedGraph", 'w')
fileBedGraphAGConversions = open(
outputBedGraphPrefix + "_AG_conversions_genomewide.bedGraph", 'w')
fileBedGraphT = open(
outputBedGraphPrefix + "_coverage_T_genomewide.bedGraph", 'w')
fileBedGraphA = open(
outputBedGraphPrefix + "_coverage_A_genomewide.bedGraph", 'w')
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" tc-conversions\" description=\"# T->C conversions / # reads on T per position genome-wide\"",
file=fileBedGraphRatesPlus)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" ag-conversions\" description=\"# A->G conversions / # reads on A per position genome-wide\"",
file=fileBedGraphRatesMinus)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" plus-strand coverage\" description=\"# Reads on plus strand genome-wide\"",
file=fileBedGraphCoveragePlus)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" minus-strand coverage\" description=\"# Reads on minus strand genome-wide\"",
file=fileBedGraphCoverageMinus)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" T->C conversions\" description=\"# T->C conversions on plus strand genome-wide\"",
file=fileBedGraphTCConversions)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" A->G conversions\" description=\"# A->G conversions on minus strand genome-wide\"",
file=fileBedGraphAGConversions)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" T-coverage\" description=\"# Plus-strand reads on Ts genome-wide\"",
file=fileBedGraphT)
print(
"track type=bedGraph name=\"" + bedGraphInfo +
" A-coverage\" description=\"# Minus-strand reads on As genome-wide\"",
file=fileBedGraphA)
for chromosome in chromosomes:
chrLength = testFile.getChromosomeLength(chromosome)
tcCount = [0] * chrLength
agCount = [0] * chrLength
coveragePlus = [0] * chrLength
coverageMinus = [0] * chrLength
tCoverage = [0] * chrLength
aCoverage = [0] * chrLength
readIterator = testFile.readsInChromosome(chromosome, minBaseQual,
conversionThreshold)
for read in readIterator:
if (not read.isTcRead):
read.tcCount = 0
read.mismatches = []
read.conversionRates = 0.0
read.tcRate = 0.0
for mismatch in read.mismatches:
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)
and mismatch.referencePosition >= 0
and mismatch.referencePosition < chrLength):
if read.direction == ReadDirection.Reverse:
agCount[mismatch.referencePosition] += 1
else:
tcCount[mismatch.referencePosition] += 1
for i in xrange(read.startRefPos, read.endRefPos):
if (i >= 0 and i < chrLength):
if read.direction == ReadDirection.Reverse:
coverageMinus[i] += 1
else:
coveragePlus[i] += 1
prevCoveragePlus = 0
prevCoveragePlusPos = 0
prevCoverageMinus = 0
prevCoverageMinusPos = 0
prevTCConversionRate = 0
prevTCConversionRatePos = 0
prevAGConversionRate = 0
prevAGConversionRatePos = 0
prevTCConversions = 0
prevTCConversionPos = 0
prevAGConversions = 0
prevAGConversionPos = 0
prevTCoverage = 0
prevTCoveragePos = 0
prevACoverage = 0
prevACoveragePos = 0
for pos in xrange(0, chrLength):
if prevCoveragePlus != coveragePlus[pos]:
print(chromosome + "\t" + str(prevCoveragePlusPos + 1) + "\t" +
str(pos + 1) + "\t" + str(prevCoveragePlus),
file=fileBedGraphCoveragePlus)
prevCoveragePlus = coveragePlus[pos]
prevCoveragePlusPos = pos
if prevCoverageMinus != coverageMinus[pos]:
print(chromosome + "\t" + str(prevCoverageMinusPos + 1) +
"\t" + str(pos + 1) + "\t" + str(prevCoverageMinus),
file=fileBedGraphCoverageMinus)
prevCoverageMinus = coverageMinus[pos]
prevCoverageMinusPos = pos
tCoverage = 0
if coveragePlus[pos] > 0:
base = ref.fetch(reference=chromosome,
start=pos + 1,
end=pos + 2)
if base.upper() == "T":
tCoverage = coveragePlus[pos]
aCoverage = 0
if coverageMinus[pos] > 0:
base = ref.fetch(reference=chromosome,
start=pos + 1,
end=pos + 2)
if base.upper() == "A":
aCoverage = coverageMinus[pos]
if prevTCoverage != tCoverage:
print(chromosome + "\t" + str(prevTCoveragePos + 1) + "\t" +
str(pos + 1) + "\t" + str(prevTCoverage),
file=fileBedGraphT)
prevTCoverage = tCoverage
prevTCoveragePos = pos
if prevACoverage != aCoverage:
print(chromosome + "\t" + str(prevACoveragePos + 1) + "\t" +
str(pos + 1) + "\t" + str(prevACoverage),
file=fileBedGraphA)
prevACoverage = aCoverage
prevACoveragePos = pos
if prevTCConversions != tcCount[pos]:
print(chromosome + "\t" + str(prevTCConversionPos + 1) + "\t" +
str(pos + 1) + "\t" + str(prevTCConversions),
file=fileBedGraphTCConversions)
prevTCConversions = tcCount[pos]
prevTCConversionPos = pos
if prevAGConversions != agCount[pos]:
print(chromosome + "\t" + str(prevAGConversionPos + 1) + "\t" +
str(pos + 1) + "\t" + str(prevAGConversions),
file=fileBedGraphAGConversions)
prevAGConversions = agCount[pos]
prevAGConversionPos = pos
TCconversionRate = 0
if coveragePlus[pos] > 0 and coveragePlus[pos] >= coverageCutoff:
TCconversionRate = float(tcCount[pos]) / float(
coveragePlus[pos])
AGconversionRate = 0
if coverageMinus[pos] > 0 and coverageMinus[pos] >= coverageCutoff:
AGconversionRate = float(agCount[pos]) / float(
coverageMinus[pos])
if prevTCConversionRate != TCconversionRate:
print(chromosome + "\t" + str(prevTCConversionRatePos + 1) +
"\t" + str(pos + 1) + "\t" + str(prevTCConversionRate),
file=fileBedGraphRatesPlus)
prevTCConversionRate = TCconversionRate
prevTCConversionRatePos = pos
if prevAGConversionRate != AGconversionRate:
print(chromosome + "\t" + str(prevAGConversionRatePos + 1) +
"\t" + str(pos + 1) + "\t" + str(prevAGConversionRate),
file=fileBedGraphRatesMinus)
prevAGConversionRate = AGconversionRate
prevAGConversionRatePos = pos
fileBedGraphRatesPlus.close()
fileBedGraphRatesMinus.close()
fileBedGraphCoveragePlus.close()
fileBedGraphCoverageMinus.close()
fileBedGraphTCConversions.close()
fileBedGraphAGConversions.close()
fileBedGraphT.close()
fileBedGraphA.close()
def tcPerReadPos(referenceFile,
bam,
minQual,
maxReadLength,
outputCSV,
outputPDF,
snpsFile,
log,
printOnly=False,
verbose=True,
force=False):
if (not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing T->C per reads position for file " + bam,
file=log)
else:
totalReadCountFwd = [0] * maxReadLength
totalReadCountRev = [0] * maxReadLength
tcPerPosRev = [0] * maxReadLength
tcPerPosFwd = [0] * maxReadLength
allPerPosRev = [0] * maxReadLength
allPerPosFwd = [0] * maxReadLength
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minQual)
for read in readIterator:
tcCounts = [0] * maxReadLength
mutCounts = [0] * maxReadLength
for mismatch in read.mismatches:
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)):
tcCounts[mismatch.readPosition] += 1
else:
mutCounts[mismatch.readPosition] += 1
query_length = len(read.sequence)
if (read.direction == ReadDirection.Reverse):
tcPerPosRev = sumLists(tcPerPosRev, tcCounts)
allPerPosRev = sumLists(allPerPosRev, mutCounts)
for i in range(0, query_length):
totalReadCountRev[i] += 1
else:
tcPerPosFwd = sumLists(tcPerPosFwd, tcCounts)
allPerPosFwd = sumLists(allPerPosFwd, mutCounts)
for i in range(0, query_length):
totalReadCountFwd[i] += 1
foTC = open(outputCSV, "w")
print("# slamdunk tcperreadpos v" + __version__, file=foTC)
for i in range(0, maxReadLength):
print(allPerPosFwd[i],
allPerPosRev[i],
tcPerPosFwd[i],
tcPerPosRev[i],
totalReadCountFwd[i],
totalReadCountRev[i],
sep='\t',
file=foTC)
foTC.close()
if (not checkStep([outputCSV], [outputPDF], force)):
print("Skipped computing T->C per reads position plot for file " + bam,
file=log)
else:
callR(getPlotter("conversion_per_read_position") + " -i " + outputCSV +
" -o " + outputPDF,
log,
dry=printOnly,
verbose=verbose)
def evaluateReads(bam, referenceFile, bed, outputFile, mainOutput):
print("Run " + bam)
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
chromosomes = testFile.getChromosomes()
bedTree = bedToIntervallTree(bed)
#evalHist = [0] *
outFile = open(outputFile, "w")
print("read.name",
"read.chromosome",
"read.startRefPos",
"sim.utr",
"read.utr",
"sim.tcCount",
"read.tcCount",
sep="\t",
file=outFile)
total = 0
correct = 0
correcPosWrongTC = 0
wrongPos = 0
minBaseQual = 0
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minBaseQual)
for read in readIterator:
total += 1
simInfo = read.name.split("_")
utrSim = simInfo[0]
tcCountSim = int(simInfo[2])
utrFound = None
if read.chromosome in bedTree:
overlaps = list(
bedTree[read.chromosome][read.startRefPos:read.endRefPos])
if len(overlaps) > 0:
utrFound = overlaps[0].data
if utrFound == utrSim:
if tcCountSim == read.tcCount:
correct += 1
else:
correcPosWrongTC += 1
else:
wrongPos += 1
print(read.name,
read.chromosome,
read.startRefPos,
utrSim,
utrFound,
tcCountSim,
read.tcCount,
sep="\t",
file=outFile)
#print(correct * 100.0 / total, correcPosWrongTC * 100.0 / total, wrongPos * 100.0 / total, total)
print(correct, correcPosWrongTC, wrongPos, total)
def tcPerReadPos(referenceFile, bam, minQual, maxReadLength, outputCSV, outputPDF, snpsFile, log, printOnly=False, verbose=True, force=False):
if(not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing T->C per reads position for file " + bam, file=log)
else:
totalReadCountFwd = [0] * maxReadLength
totalReadCountRev = [0] * maxReadLength
tcPerPosRev = [0] * maxReadLength
tcPerPosFwd = [0] * maxReadLength
allPerPosRev = [0] * maxReadLength
allPerPosFwd = [0] * maxReadLength
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
readIterator = testFile.readsInChromosome(chromosome, minQual)
for read in readIterator:
tcCounts = [0] * maxReadLength
mutCounts = [0] * maxReadLength
for mismatch in read.mismatches:
if(mismatch.isTCMismatch(read.direction == ReadDirection.Reverse)):
tcCounts[mismatch.readPosition] += 1
else :
mutCounts[mismatch.readPosition] += 1
query_length = len(read.sequence)
if(read.direction == ReadDirection.Reverse):
tcPerPosRev = sumLists(tcPerPosRev, tcCounts)
allPerPosRev = sumLists(allPerPosRev, mutCounts)
for i in range(0, query_length):
totalReadCountRev[i] += 1
else:
tcPerPosFwd = sumLists(tcPerPosFwd, tcCounts)
allPerPosFwd = sumLists(allPerPosFwd, mutCounts)
for i in range(0, query_length):
totalReadCountFwd[i] += 1
foTC = open(outputCSV, "w")
print("# slamdunk tcperreadpos v" + __version__, file=foTC)
for i in range(0, maxReadLength):
print(allPerPosFwd[i], allPerPosRev[i], tcPerPosFwd[i], tcPerPosRev[i], totalReadCountFwd[i], totalReadCountRev[i], sep='\t', file=foTC)
foTC.close()
if(not checkStep([outputCSV], [outputPDF], force)):
print("Skipped computing T->C per reads position plot for file " + bam, file=log)
else:
callR(getPlotter("conversion_per_read_position") + " -i " + outputCSV + " -o " + outputPDF, log, dry=printOnly, verbose=verbose)