def getConversionRateFromBam(bam, ref, chromosome, start, end, strand):
testFile = SlamSeqBamFile(bam, ref, SNPtools.SNPDictionary(None))
sumConversionRate = 0
readCount = 0
#for chromosome in testFile.getChromosomes():
# readIterator = testFile.readsInChromosome(chromosome)
#readIterator = testFile.readInRegion("chr7", 3217778, 3221036, "+", 55)
readIterator = testFile.readInRegion(chromosome, start, end, strand, 100)
for read in readIterator:
conversionRate = 0
if(read.tCount > 0):
conversionRate = read.tcCount * 1.0 / read.tCount
#if(read.tcCount > 0):
sumConversionRate += conversionRate
readCount += 1
#if(readCount % 1000 == 0 and readCount > 0):
# print(str(readCount) + ": " + str(sumConversionRate) + " / " + str(readCount) + " = " + str(sumConversionRate / readCount))
#if(readCount >= 10000):
# break
print("Read count: " + str(readCount))
print("Avg. conversion rate: " + str(sumConversionRate / readCount))
def getConversionRateFromBam(bam, ref, chromosome, start, end, strand):
testFile = SlamSeqBamFile(bam, ref, SNPtools.SNPDictionary(None))
sumConversionRate = 0
readCount = 0
#for chromosome in testFile.getChromosomes():
# readIterator = testFile.readsInChromosome(chromosome)
#readIterator = testFile.readInRegion("chr7", 3217778, 3221036, "+", 55)
readIterator = testFile.readInRegion(chromosome, start, end, strand, 100)
for read in readIterator:
conversionRate = 0
if (read.tCount > 0):
conversionRate = read.tcCount * 1.0 / read.tCount
#if(read.tcCount > 0):
sumConversionRate += conversionRate
readCount += 1
#if(readCount % 1000 == 0 and readCount > 0):
# print(str(readCount) + ": " + str(sumConversionRate) + " / " + str(readCount) + " = " + str(sumConversionRate / readCount))
#if(readCount >= 10000):
# break
print("Read count: " + str(readCount))
print("Avg. conversion rate: " + str(sumConversionRate / readCount))
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 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 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 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 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 computeTconversions(ref,
bed,
snpsFile,
bam,
maxReadLength,
minQual,
outputCSV,
outputBedgraphPlus,
outputBedgraphMinus,
conversionThreshold,
log,
mle=False):
referenceFile = pysam.FastaFile(ref)
sampleInfo = getSampleInfo(bam)
slamseqInfo = SlamSeqInfo(bam)
#readNumber = slamseqInfo.MappedReads
readNumber = slamseqInfo.FilteredReads
bedMD5 = md5(bed)
if (mle):
fileNameTest = replaceExtension(outputCSV, ".tsv", "_perread")
fileTest = open(fileNameTest, 'w')
print("#slamdunk v" + __version__,
__count_version__,
"sample info:",
sampleInfo.Name,
sampleInfo.ID,
sampleInfo.Type,
sampleInfo.Time,
sep="\t",
file=fileTest)
print("#annotation:",
os.path.basename(bed),
bedMD5,
sep="\t",
file=fileTest)
#print("utr", "n", "k", file=fileTest)
print(SlamSeqInterval.Header, file=fileTest)
fileCSV = open(outputCSV, 'w')
print("#slamdunk v" + __version__,
__count_version__,
"sample info:",
sampleInfo.Name,
sampleInfo.ID,
sampleInfo.Type,
sampleInfo.Time,
sep="\t",
file=fileCSV)
print("#annotation:",
os.path.basename(bed),
bedMD5,
sep="\t",
file=fileCSV)
print(SlamSeqInterval.Header, file=fileCSV)
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
#Go through one chr after the other
testFile = SlamSeqBamFile(bam, ref, snps)
if not testFile.bamVersion == __bam_version__:
raise RuntimeError("Wrong filtered BAM file version detected (" +
testFile.bamVersion + "). Expected version " +
__bam_version__ + ". Please rerun slamdunk filter.")
bedMD5 = md5(bed)
if slamseqInfo.AnnotationMD5 != bedMD5:
print(
"Warning: MD5 checksum of annotation (" + bedMD5 +
") does not matched MD5 in filtered BAM files (" +
slamseqInfo.AnnotationMD5 +
"). Most probably the annotation filed changed after the filtered BAM files were created.",
file=log)
conversionBedGraph = {}
for utr in BedIterator(bed):
Tcontent = 0
slamSeqUtr = SlamSeqInterval(utr.chromosome, utr.start, utr.stop,
utr.strand, utr.name, Tcontent, 0, 0, 0,
0, 0, 0, 0)
slamSeqUtrMLE = SlamSeqInterval(utr.chromosome, utr.start, utr.stop,
utr.strand, utr.name, Tcontent, 0, 0,
0, 0, 0, 0, 0)
if (not utr.hasStrand()):
raise RuntimeError(
"Input BED file does not contain stranded intervals.")
if utr.start < 0:
raise RuntimeError(
"Negativ start coordinate found. Please check the following entry in your BED file: "
+ utr)
# Retreive reference sequence
region = utr.chromosome + ":" + str(utr.start + 1) + "-" + str(
utr.stop)
if (utr.chromosome in list(referenceFile.references)):
#print(refRegion,file=sys.stderr)
# pysam-0.15.0.1
#refSeq = referenceFile.fetch(region=region).upper()
refSeq = referenceFile.fetch(reference=utr.chromosome,
start=utr.start,
end=utr.stop).upper()
if (utr.strand == "-"):
#refSeq = complement(refSeq[::-1])
Tcontent = refSeq.count("A")
else:
Tcontent = refSeq.count("T")
slamSeqUtr._Tcontent = Tcontent
readIterator = testFile.readInRegion(utr.chromosome, utr.start,
utr.stop, utr.strand,
maxReadLength, minQual,
conversionThreshold)
tcCountUtr = [0] * utr.getLength()
coverageUtr = [0] * utr.getLength()
tInReads = []
tcInRead = []
countFwd = 0
tcCountFwd = 0
countRev = 0
tCountRev = 0
multiMapFwd = 0
multiMapRev = 0
for read in readIterator:
# Overwrite any conversions for non-TC reads (reads with < 2 TC conversions)
if (not read.isTcRead):
read.tcCount = 0
read.mismatches = []
read.conversionRates = 0.0
read.tcRate = 0.0
if (read.direction == ReadDirection.Reverse):
countRev += 1
if read.tcCount > 0:
tCountRev += 1
if read.isMultimapper:
multiMapRev += 1
else:
countFwd += 1
if read.tcCount > 0:
tcCountFwd += 1
if read.isMultimapper:
multiMapFwd += 1
for mismatch in read.mismatches:
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)
and mismatch.referencePosition >= 0
and mismatch.referencePosition < utr.getLength()):
tcCountUtr[mismatch.referencePosition] += 1
testN = read.getTcount()
testk = 0
for mismatch in read.mismatches:
if (mismatch.referencePosition >= 0
and mismatch.referencePosition < utr.getLength()):
if (mismatch.isT(read.direction == ReadDirection.Reverse)):
testN += 1
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)):
testk += 1
#print(utr.name, read.name, read.direction, testN, testk, read.sequence, sep="\t")
tInReads.append(testN)
tcInRead.append(testk)
#print(utr.name, testN, testk, sep="\t", file=fileTest)
for i in xrange(read.startRefPos, read.endRefPos):
if (i >= 0 and i < utr.getLength()):
coverageUtr[i] += 1
if ((utr.strand == "+" and countFwd > 0)
or (utr.strand == "-" and countRev > 0)):
tcRateUtr = [
x * 100.0 / y if y > 0 else 0
for x, y in zip(tcCountUtr, coverageUtr)
]
readCount = countFwd
tcReadCount = tcCountFwd
multiMapCount = multiMapFwd
if (utr.strand == "-"):
readCount = countRev
tcReadCount = tCountRev
multiMapCount = multiMapRev
if ((utr.strand == "-" and countFwd > countRev)
or (utr.strand == "+" and countRev > countFwd)):
print(
"Warning: " + utr.name + " is located on the " +
utr.strand +
" strand but read counts are higher for the opposite strand (fwd: "
+ countFwd + ", rev: " + countRev + ")",
file=sys.stderr)
refSeq = readIterator.getRefSeq()
# Get number of covered Ts/As in the UTR and compute average conversion rate for all covered Ts/As
coveredTcount = 0
avgConversationRate = 0
coveredPositions = 0
# Get number of reads on T positions and number of reads with T->C conversions on T positions
coverageOnTs = 0
conversionsOnTs = 0
for position in xrange(0, len(coverageUtr)):
if (coverageUtr[position] > 0
and ((utr.strand == "+" and refSeq[position] == "T") or
(utr.strand == "-" and refSeq[position] == "A"))):
coveredTcount += 1
avgConversationRate += tcRateUtr[position]
coverageOnTs += coverageUtr[position]
conversionsOnTs += tcCountUtr[position]
conversionBedGraph[utr.chromosome + ":" +
str(utr.start + position) + ":" +
str(utr.strand)] = tcRateUtr[position]
if (coverageUtr[position] > 0):
coveredPositions += 1
if (coveredTcount > 0):
avgConversationRate = avgConversationRate / coveredTcount
else:
avgConversationRate = 0
# reads per million mapped to the UTR
readsCPM = 0
if (readNumber > 0):
readsCPM = readCount * 1000000.0 / readNumber
# Convert to SlamSeqInterval and print
conversionRate = 0
if (coverageOnTs > 0):
conversionRate = float(conversionsOnTs) / float(coverageOnTs)
slamSeqUtr = SlamSeqInterval(utr.chromosome, utr.start, utr.stop,
utr.strand, utr.name, Tcontent,
readsCPM, coverageOnTs,
conversionsOnTs, conversionRate,
readCount, tcReadCount, multiMapCount)
slamSeqUtrMLE = SlamSeqInterval(
utr.chromosome, utr.start, utr.stop, utr.strand, utr.name,
Tcontent, readsCPM, coverageOnTs, conversionsOnTs,
conversionRate, ",".join(str(x) for x in tInReads),
",".join(str(x) for x in tcInRead), multiMapCount)
print(slamSeqUtr, file=fileCSV)
if (mle):
print(slamSeqUtrMLE, file=fileTest)
fileCSV.close()
if (mle):
fileTest.close()
fileBedgraphPlus = open(outputBedgraphPlus, 'w')
fileBedgraphMinus = open(outputBedgraphMinus, 'w')
for position in conversionBedGraph:
positionData = position.split(":")
if (positionData[2] == "+"):
print(positionData[0],
positionData[1],
int(positionData[1]) + 1,
conversionBedGraph[position],
file=fileBedgraphPlus)
else:
print(positionData[0],
positionData[1],
int(positionData[1]) + 1,
conversionBedGraph[position],
file=fileBedgraphMinus)
fileBedgraphPlus.close()
fileBedgraphMinus.close()
if (mle):
fileNameMLE = replaceExtension(outputCSV, ".tsv", "_mle")
callR(
getPlotter("compute_conversion_rate_mle") + " -f " + fileNameTest +
" -r " + "0.024" + " -o " + fileNameMLE + " &> /dev/null")
def computeSNPMaskedRates(ref,
bed,
snpsFile,
bam,
maxReadLength,
minQual,
coverageCutoff,
variantFraction,
outputCSV,
outputPDF,
strictTCs,
log,
printOnly=False,
verbose=True,
force=False):
if (not checkStep([bam, ref], [outputCSV], force)):
print("Skipped computing T->C per UTR with SNP masking for file " +
bam,
file=log)
else:
fileCSV = open(outputCSV, 'w')
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
#Go through one chr after the other
testFile = SlamSeqBamFile(bam, ref, snps)
progress = 0
for utr in BedIterator(bed):
if (not utr.hasStrand()):
raise RuntimeError(
"Input BED file does not contain stranded intervals.")
if utr.start < 0:
raise RuntimeError(
"Negativ start coordinate found. Please check the following entry in your BED file: "
+ utr)
readIterator = testFile.readInRegion(utr.chromosome, utr.start,
utr.stop, utr.strand,
maxReadLength, minQual)
unmaskedTCCount = 0
maskedTCCount = 0
readCount = 0
for read in readIterator:
# Overwrite any conversions for non-TC reads (reads with < 2 TC conversions)
if (not read.isTcRead and strictTCs):
read.tcCount = 0
read.mismatches = []
read.conversionRates = 0.0
read.tcRate = 0.0
isTC = False
isTrueTC = False
for mismatch in read.mismatches:
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)
and mismatch.referencePosition >= 0
and mismatch.referencePosition < utr.getLength()):
isTrueTC = True
unmasked = False
if (read.direction == ReadDirection.Reverse
and mismatch.referenceBase == "A"
and mismatch.readBase == "G"):
unmasked = True
elif (read.direction != ReadDirection.Reverse
and mismatch.referenceBase == "T"
and mismatch.readBase == "C"):
unmasked = True
if (unmasked and mismatch.referencePosition >= 0
and mismatch.referencePosition < utr.getLength()):
isTC = True
readCount += 1
if (isTC):
unmaskedTCCount += 1
if (isTrueTC):
maskedTCCount += 1
containsSNP = 0
if (unmaskedTCCount != maskedTCCount):
containsSNP = 1
print(utr.name + "\t" + str(readCount) + "\t" +
str(unmaskedTCCount) + "\t" + str(maskedTCCount) + "\t" +
str(containsSNP),
file=fileCSV)
progress += 1
fileCSV.close()
if (not checkStep([outputCSV], [outputPDF], force)):
print("Skipped computing T->C per UTR position plot for file " + bam,
file=log)
else:
callR(getPlotter("SNPeval") + " -i " + outputCSV + " -c " +
str(coverageCutoff) + " -v " + str(variantFraction) + " -o " +
outputPDF,
log,
dry=printOnly,
verbose=verbose)
def statsComputeTCContext(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
# combinations = ["AT","CT","GT","TT","NT","AA","CA","GA","TA","NA"]
frontCombinations = ["AT", "CT", "GT", "TT", "NT"]
backCombinations = ["TA", "TC", "TG", "TT", "TN"]
counts = {}
counts['5prime'] = {}
counts['3prime'] = {}
counts['5prime']['fwd'] = {}
counts['5prime']['rev'] = {}
counts['3prime']['fwd'] = {}
counts['3prime']['rev'] = {}
for combination in frontCombinations :
counts['5prime']['fwd'][combination] = 0
counts['5prime']['rev'][combination] = 0
for combination in backCombinations:
counts['3prime']['fwd'][combination] = 0
counts['3prime']['rev'][combination] = 0
bamFile = pysam.AlignmentFile(bam, "rb")
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
for read in bamFile.fetch(region=chromosome):
i = 0
while i < len(read.query_sequence):
if(read.query_sequence[i] == "T" and not read.is_reverse) :
frontContext = None
backContext = None
if (i > 0) :
frontContext = read.query_sequence[i - 1]
if (i < (len(read.query_sequence) - 1)) :
backContext = read.query_sequence[i + 1]
if (frontContext != None) :
counts['5prime']['fwd'][frontContext + "T"] += 1
if (backContext != None) :
counts['3prime']['fwd']["T" + backContext] += 1
if(read.query_sequence[i] == "A" and read.is_reverse) :
frontContext = None
backContext = None
if (i > 0) :
backContext = read.query_sequence[i - 1]
if (i < (len(read.query_sequence) - 1)) :
frontContext = read.query_sequence[i + 1]
if (frontContext != None) :
counts['5prime']['rev'][complement(frontContext + "A")] += 1
if (backContext != None) :
counts['3prime']['rev'][complement("A" + backContext)] += 1
i += 1
# Print rates in correct format for plotting
fo = open(outputCSV, "w")
print("\t".join(frontCombinations), file=fo)
frontFwdLine = ""
frontRevLine = ""
backFwdLine = ""
backRevLine = ""
for combination in frontCombinations :
frontFwdLine += str(counts['5prime']['fwd'][combination]) + "\t"
frontRevLine += str(counts['5prime']['rev'][combination]) + "\t"
print(frontFwdLine.rstrip(), file=fo)
print(frontRevLine.rstrip(), file=fo)
print("\t".join(backCombinations), file=fo)
for combination in backCombinations :
backFwdLine += str(counts['3prime']['fwd'][combination]) + "\t"
backRevLine += str(counts['3prime']['rev'][combination]) + "\t"
print(backFwdLine.rstrip(), file=fo)
print(backRevLine.rstrip(), file=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(os.path.basename(bam)), outputCSV, sep='\t', file=f)
f.close()
callR(getPlotter("compute_context_TC_rates") + " -f " + f.name + " -O " + outputPDF, log, dry=printOnly, verbose=verbose)
def tcPerUtr(referenceFile,
utrBed,
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 UTR position for file " + bam,
file=log)
else:
counter = 0
totalUtrCountFwd = [0] * utrNormFactor
totalUtrCountRev = [0] * utrNormFactor
tcPerPosRev = [0] * utrNormFactor
tcPerPosFwd = [0] * utrNormFactor
allPerPosRev = [0] * utrNormFactor
allPerPosFwd = [0] * utrNormFactor
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
# Go through one utr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
for utr in BedIterator(utrBed):
readIterator = testFile.readInRegion(utr.chromosome, utr.start,
utr.stop, utr.strand,
maxReadLength, minQual)
tcForwardCounts = [0] * utrNormFactor
mutForwardCounts = [0] * utrNormFactor
tcReverseCounts = [0] * utrNormFactor
mutReverseCounts = [0] * utrNormFactor
for read in readIterator:
tcCounts = [0] * utrNormFactor
mutCounts = [0] * utrNormFactor
for mismatch in read.mismatches:
mismatchPos = mismatch.referencePosition
# mismatchPos = read.startRefPos
if (utr.strand == "+"):
# New try for UTRs (remove + 1
if (mismatchPos >= (utr.getLength() - utrNormFactor)
and mismatchPos < utr.getLength()):
# if (mismatchPos >= (utr.getLength() - utrNormFactor) and mismatchPos < utr.getLength() + 1) :
mismatchPos = utrNormFactor - (utr.getLength() -
mismatchPos)
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)):
tcCounts[mismatchPos] += 1
else:
mutCounts[mismatchPos] += 1
else:
if (mismatchPos >= 0 and mismatchPos < min(
utr.getLength(), utrNormFactor)):
if (mismatch.isTCMismatch(
read.direction == ReadDirection.Reverse)):
tcCounts[mismatchPos] += 1
else:
mutCounts[mismatchPos] += 1
if (read.direction == ReadDirection.Reverse):
tcReverseCounts = sumLists(tcReverseCounts, tcCounts)
mutReverseCounts = sumLists(mutReverseCounts, mutCounts)
start = max(
0,
min(min(utr.getLength(), utrNormFactor),
read.startRefPos))
end = max(
0,
min(min(utr.getLength(), utrNormFactor),
read.endRefPos))
for i in range(start, end):
totalUtrCountRev[i] += 1
else:
tcForwardCounts = sumLists(tcForwardCounts, tcCounts)
mutForwardCounts = sumLists(mutForwardCounts, mutCounts)
start = min(
utr.getLength(),
max(utr.getLength() - utrNormFactor, read.startRefPos))
end = min(
utr.getLength(),
max(utr.getLength() - utrNormFactor, read.endRefPos))
for i in range(start, end):
normPos = utrNormFactor - (utr.getLength() - i)
totalUtrCountFwd[normPos] += 1
tcPerPosFwd = sumLists(tcPerPosFwd, tcForwardCounts)
allPerPosFwd = sumLists(allPerPosFwd, mutForwardCounts)
tcPerPosRev = sumLists(tcPerPosRev, tcReverseCounts)
allPerPosRev = sumLists(allPerPosRev, mutReverseCounts)
counter += 1
if (verbose and counter % 10000 == 0):
print("Handled " + str(counter) + " UTRs.", file=log)
foTC = open(outputCSV, "w")
print("# slamdunk tcperutr v" + __version__, file=foTC)
reverseAllPerPosRev = allPerPosRev[::-1]
reverseTcPerPosRev = tcPerPosRev[::-1]
reverseTotalUtrCountRev = totalUtrCountRev[::-1]
for i in range(0, utrNormFactor):
print(allPerPosFwd[i],
reverseAllPerPosRev[i],
tcPerPosFwd[i],
reverseTcPerPosRev[i],
totalUtrCountFwd[i],
reverseTotalUtrCountRev[i],
sep='\t',
file=foTC)
foTC.close()
if (not checkStep([outputCSV], [outputPDF], force)):
print("Skipped computing T->C per UTR position plot for file " + bam,
file=log)
else:
callR(getPlotter("conversion_per_read_position") + " -u -i " +
outputCSV + " -o " + outputPDF,
log,
dry=printOnly,
verbose=verbose)
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 statsComputeOverallRatesPerUTR(referenceFile,
bam,
minBaseQual,
strictTCs,
outputCSV,
outputPDF,
utrBed,
maxReadLength,
log,
printOnly=False,
verbose=True,
force=False):
sampleInfo = getSampleInfo(bam)
slamseqInfo = SlamSeqInfo(bam)
if (not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing overall rates for file " + bam, file=log)
else:
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
# UTR stats for MultiQC
utrStats = dict()
plotConversions = [
'A>T',
'A>G',
'A>C',
'C>A',
'C>G',
'C>T',
'G>A',
'G>C',
'G>T',
'T>A',
'T>G',
'T>C',
]
for conversion in plotConversions:
utrStats[conversion] = list()
f = tempfile.NamedTemporaryFile(delete=False)
for utr in BedIterator(utrBed):
readIterator = testFile.readInRegion(utr.chromosome, utr.start,
utr.stop, utr.strand,
maxReadLength, minBaseQual)
# Init
totalRates = [0] * 25
readCount = 0
for read in readIterator:
if (not read.isTcRead and strictTCs and read.tcCount > 0):
pass
else:
# Compute rates for current read
rates = read.conversionRates
# Add rates from read to total rates
totalRates = sumLists(totalRates, rates)
readCount += 1
print(utr.name,
utr.chromosome,
utr.start,
utr.stop,
utr.strand,
readCount,
"\t".join(str(x) for x in totalRates),
sep="\t",
file=f)
# Process rates for MultiQC
# Copied directly, too lazy to do it properly now
utrDict = {}
conversionSum = 0
A_A = totalRates[0]
conversionSum = +A_A
A_C = totalRates[1]
conversionSum = +A_C
A_G = totalRates[2]
conversionSum = +A_G
A_T = totalRates[3]
conversionSum = +A_T
C_A = totalRates[5]
conversionSum = +C_A
C_C = totalRates[6]
conversionSum = +C_C
C_G = totalRates[7]
conversionSum = +C_G
C_T = totalRates[8]
conversionSum = +C_T
G_A = totalRates[10]
conversionSum = +G_A
G_C = totalRates[11]
conversionSum = +G_C
G_G = totalRates[12]
conversionSum = +G_G
G_T = totalRates[13]
conversionSum = +G_T
T_A = totalRates[15]
conversionSum = +T_A
T_C = totalRates[16]
conversionSum = +T_C
T_G = totalRates[17]
conversionSum = +T_G
T_T = totalRates[18]
conversionSum = +T_T
if utr.strand == "-":
A_A, T_T = T_T, A_A
G_G, C_C = C_C, G_G
A_C, T_G = T_G, A_C
A_G, T_C = T_C, A_G
A_T, T_A = T_A, A_T
C_A, G_T = G_T, C_A
C_G, G_C = G_C, C_G
C_T, G_A = G_A, C_T
if conversionSum > 0:
Asum = A_A + A_C + A_G + A_T
Csum = C_A + C_C + C_G + C_T
Gsum = G_A + G_C + G_G + G_T
Tsum = T_A + T_C + T_G + T_T
if Asum > 0:
A_T = A_T / float(Asum) * 100
A_G = A_G / float(Asum) * 100
A_C = A_C / float(Asum) * 100
else:
A_T = 0
A_G = 0
A_C = 0
if Csum > 0:
C_A = C_A / float(Csum) * 100
C_G = C_G / float(Csum) * 100
C_T = C_T / float(Csum) * 100
else:
C_A = 0
C_G = 0
C_T = 0
if Gsum > 0:
G_A = G_A / float(Gsum) * 100
G_C = G_C / float(Gsum) * 100
G_T = G_T / float(Gsum) * 100
else:
G_A = 0
G_C = 0
G_T = 0
if Tsum > 0:
T_A = T_A / float(Tsum) * 100
T_G = T_G / float(Tsum) * 100
T_C = T_C / float(Tsum) * 100
else:
T_A = 0
T_G = 0
T_C = 0
utrStats['A>T'].append(A_T)
utrStats['A>G'].append(A_G)
utrStats['A>C'].append(A_C)
utrStats['C>A'].append(C_A)
utrStats['C>G'].append(C_G)
utrStats['C>T'].append(C_T)
utrStats['G>A'].append(G_A)
utrStats['G>T'].append(G_T)
utrStats['G>C'].append(G_C)
utrStats['T>A'].append(T_A)
utrStats['T>G'].append(T_G)
utrStats['T>C'].append(T_C)
f.close()
fo = open(outputCSV, "w")
print("# slamdunk utrrates v" + __version__, file=fo)
print("# Median-Conversions=", end="", file=fo)
first = True
for conversion in plotConversions:
if (not first):
print(',', file=fo, end="")
else:
first = False
print(conversion + ":" + str(np.median(utrStats[conversion])),
file=fo,
end="")
print(file=fo)
print("Name",
"Chr",
"Start",
"End",
"Strand",
"ReadCount",
sep="\t",
end="\t",
file=fo)
for i in range(0, 5):
for j in range(0, 5):
print(toBase[i].upper() + "_" + toBase[j].upper(),
end="",
file=fo)
if (i != 4 or j != 4):
print("\t", end="", file=fo)
print(file=fo)
with open(f.name, "rb") as valueFile:
fo.write(valueFile.read())
fo.close()
if (not checkStep([bam, referenceFile], [outputPDF], force)):
print("Skipped computing global rate pdfs for file " + bam, file=log)
else:
f = tempfile.NamedTemporaryFile(delete=False)
print(sampleInfo.Name, outputCSV, sep='\t', file=f)
f.close()
callR(getPlotter("globalRatePlotter") + " -f " + f.name + " -O " +
outputPDF,
log,
dry=printOnly,
verbose=verbose)
def statsComputeTCContext(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
# combinations = ["AT","CT","GT","TT","NT","AA","CA","GA","TA","NA"]
frontCombinations = ["AT", "CT", "GT", "TT", "NT"]
backCombinations = ["TA", "TC", "TG", "TT", "TN"]
counts = {}
counts['5prime'] = {}
counts['3prime'] = {}
counts['5prime']['fwd'] = {}
counts['5prime']['rev'] = {}
counts['3prime']['fwd'] = {}
counts['3prime']['rev'] = {}
for combination in frontCombinations:
counts['5prime']['fwd'][combination] = 0
counts['5prime']['rev'][combination] = 0
for combination in backCombinations:
counts['3prime']['fwd'][combination] = 0
counts['3prime']['rev'][combination] = 0
bamFile = pysam.AlignmentFile(bam, "rb")
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
chromosomes = testFile.getChromosomes()
for chromosome in chromosomes:
for read in bamFile.fetch(region=chromosome):
i = 0
while i < len(read.query_sequence):
if (read.query_sequence[i] == "T" and not read.is_reverse):
frontContext = None
backContext = None
if (i > 0):
frontContext = read.query_sequence[i - 1]
if (i < (len(read.query_sequence) - 1)):
backContext = read.query_sequence[i + 1]
if (frontContext != None):
counts['5prime']['fwd'][frontContext + "T"] += 1
if (backContext != None):
counts['3prime']['fwd']["T" + backContext] += 1
if (read.query_sequence[i] == "A" and read.is_reverse):
frontContext = None
backContext = None
if (i > 0):
backContext = read.query_sequence[i - 1]
if (i < (len(read.query_sequence) - 1)):
frontContext = read.query_sequence[i + 1]
if (frontContext != None):
counts['5prime']['rev'][complement(frontContext +
"A")] += 1
if (backContext != None):
counts['3prime']['rev'][complement(
"A" + backContext)] += 1
i += 1
# Print rates in correct format for plotting
fo = open(outputCSV, "w")
print("\t".join(frontCombinations), file=fo)
frontFwdLine = ""
frontRevLine = ""
backFwdLine = ""
backRevLine = ""
for combination in frontCombinations:
frontFwdLine += str(counts['5prime']['fwd'][combination]) + "\t"
frontRevLine += str(counts['5prime']['rev'][combination]) + "\t"
print(frontFwdLine.rstrip(), file=fo)
print(frontRevLine.rstrip(), file=fo)
print("\t".join(backCombinations), file=fo)
for combination in backCombinations:
backFwdLine += str(counts['3prime']['fwd'][combination]) + "\t"
backRevLine += str(counts['3prime']['rev'][combination]) + "\t"
print(backFwdLine.rstrip(), file=fo)
print(backRevLine.rstrip(), file=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(os.path.basename(bam)),
outputCSV,
sep='\t',
file=f)
f.close()
callR(getPlotter("compute_context_TC_rates") + " -f " + f.name +
" -O " + outputPDF,
log,
dry=printOnly,
verbose=verbose)
def statsComputeOverallRatesPerUTR(referenceFile, bam, minBaseQual, strictTCs, outputCSV, outputPDF, utrBed, maxReadLength, log, printOnly=False, verbose=True, force=False):
sampleInfo = getSampleInfo(bam)
slamseqInfo = SlamSeqInfo(bam)
if(not checkStep([bam, referenceFile], [outputCSV], force)):
print("Skipped computing overall rates for file " + bam, file=log)
else:
# Go through one chr after the other
testFile = SlamSeqBamFile(bam, referenceFile, None)
# UTR stats for MultiQC
utrStats = dict()
plotConversions = ['A>T', 'A>G', 'A>C',
'C>A', 'C>G', 'C>T',
'G>A', 'G>C', 'G>T',
'T>A', 'T>G', 'T>C',
]
for conversion in plotConversions:
utrStats[conversion] = list()
f = tempfile.NamedTemporaryFile(delete=False)
for utr in BedIterator(utrBed):
readIterator = testFile.readInRegion(utr.chromosome, utr.start, utr.stop, utr.strand, maxReadLength, minBaseQual)
# Init
totalRates = [0] * 25
readCount = 0
for read in readIterator:
if (not read.isTcRead and strictTCs and read.tcCount > 0) :
pass
else :
# Compute rates for current read
rates = read.conversionRates
# Add rates from read to total rates
totalRates = sumLists(totalRates, rates)
readCount += 1
print(utr.name, utr.chromosome, utr.start, utr.stop, utr.strand, readCount, "\t".join(str(x) for x in totalRates), sep="\t", file=f)
# Process rates for MultiQC
# Copied directly, too lazy to do it properly now
utrDict = {}
conversionSum = 0
A_A = totalRates[0]
conversionSum =+ A_A
A_C = totalRates[1]
conversionSum =+ A_C
A_G = totalRates[2]
conversionSum =+ A_G
A_T = totalRates[3]
conversionSum =+ A_T
C_A = totalRates[5]
conversionSum =+ C_A
C_C = totalRates[6]
conversionSum =+ C_C
C_G = totalRates[7]
conversionSum =+ C_G
C_T = totalRates[8]
conversionSum =+ C_T
G_A = totalRates[10]
conversionSum =+ G_A
G_C = totalRates[11]
conversionSum =+ G_C
G_G = totalRates[12]
conversionSum =+ G_G
G_T = totalRates[13]
conversionSum =+ G_T
T_A = totalRates[15]
conversionSum =+ T_A
T_C = totalRates[16]
conversionSum =+ T_C
T_G = totalRates[17]
conversionSum =+ T_G
T_T = totalRates[18]
conversionSum =+ T_T
if utr.strand == "-":
A_A, T_T = T_T,A_A
G_G, C_C = C_C,G_G
A_C, T_G = T_G, A_C
A_G, T_C = T_C, A_G
A_T, T_A = T_A, A_T
C_A, G_T = G_T, C_A
C_G, G_C = G_C, C_G
C_T, G_A = G_A, C_T
if conversionSum > 0:
Asum = A_A + A_C + A_G + A_T
Csum = C_A + C_C + C_G + C_T
Gsum = G_A + G_C + G_G + G_T
Tsum = T_A + T_C + T_G + T_T
if Asum > 0 :
A_T = A_T / float(Asum) * 100
A_G = A_G / float(Asum) * 100
A_C = A_C / float(Asum) * 100
else :
A_T = 0
A_G = 0
A_C = 0
if Csum > 0:
C_A = C_A / float(Csum) * 100
C_G = C_G / float(Csum) * 100
C_T = C_T / float(Csum) * 100
else :
C_A = 0
C_G = 0
C_T = 0
if Gsum > 0:
G_A = G_A / float(Gsum) * 100
G_C = G_C / float(Gsum) * 100
G_T = G_T / float(Gsum) * 100
else :
G_A = 0
G_C = 0
G_T = 0
if Tsum > 0:
T_A = T_A / float(Tsum) * 100
T_G = T_G / float(Tsum) * 100
T_C = T_C / float(Tsum) * 100
else :
T_A = 0
T_G = 0
T_C = 0
utrStats['A>T'].append(A_T)
utrStats['A>G'].append(A_G)
utrStats['A>C'].append(A_C)
utrStats['C>A'].append(C_A)
utrStats['C>G'].append(C_G)
utrStats['C>T'].append(C_T)
utrStats['G>A'].append(G_A)
utrStats['G>T'].append(G_T)
utrStats['G>C'].append(G_C)
utrStats['T>A'].append(T_A)
utrStats['T>G'].append(T_G)
utrStats['T>C'].append(T_C)
f.close()
fo = open(outputCSV, "w")
print("# slamdunk utrrates v" + __version__, file=fo)
print("# Median-Conversions=",end="",file=fo)
first = True
for conversion in plotConversions:
if (not first) :
print(',',file=fo, end="")
else :
first = False
print(conversion + ":" + str(np.median(utrStats[conversion])),file=fo, end="")
print(file=fo)
print("Name", "Chr", "Start", "End", "Strand", "ReadCount", sep="\t", end="\t", file=fo)
for i in range(0, 5):
for j in range(0, 5):
print(toBase[i].upper() + "_" + toBase[j].upper(), end="", file=fo)
if(i != 4 or j != 4):
print("\t", end="", file=fo)
print(file=fo)
with open(f.name, "rb") as valueFile:
fo.write(valueFile.read())
fo.close()
if(not checkStep([bam, referenceFile], [outputPDF], force)):
print("Skipped computing global rate pdfs for file " + bam, file=log)
else:
f = tempfile.NamedTemporaryFile(delete=False)
print(sampleInfo.Name, outputCSV, sep='\t', file=f)
f.close()
callR(getPlotter("globalRatePlotter") + " -f " + f.name + " -O " + outputPDF, log, dry=printOnly, verbose=verbose)
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 computeSNPMaskedRates (ref, bed, snpsFile, bam, maxReadLength, minQual, coverageCutoff, variantFraction, outputCSV, outputPDF, strictTCs, log, printOnly=False, verbose=True, force=False):
if(not checkStep([bam, ref], [outputCSV], force)):
print("Skipped computing T->C per UTR with SNP masking for file " + bam, file=log)
else:
fileCSV = open(outputCSV,'w')
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
#Go through one chr after the other
testFile = SlamSeqBamFile(bam, ref, snps)
progress = 0
for utr in BedIterator(bed):
if(not utr.hasStrand()):
raise RuntimeError("Input BED file does not contain stranded intervals.")
if utr.start < 0:
raise RuntimeError("Negativ start coordinate found. Please check the following entry in your BED file: " + utr)
readIterator = testFile.readInRegion(utr.chromosome, utr.start, utr.stop, utr.strand, maxReadLength, minQual)
unmaskedTCCount = 0
maskedTCCount = 0
readCount = 0
for read in readIterator:
# Overwrite any conversions for non-TC reads (reads with < 2 TC conversions)
if (not read.isTcRead and strictTCs) :
read.tcCount = 0
read.mismatches = []
read.conversionRates = 0.0
read.tcRate = 0.0
isTC = False
isTrueTC = False
for mismatch in read.mismatches:
if(mismatch.isTCMismatch(read.direction == ReadDirection.Reverse) and mismatch.referencePosition >= 0 and mismatch.referencePosition < utr.getLength()):
isTrueTC = True
unmasked = False
if (read.direction == ReadDirection.Reverse and mismatch.referenceBase == "A" and mismatch.readBase == "G"):
unmasked = True
elif (read.direction != ReadDirection.Reverse and mismatch.referenceBase == "T" and mismatch.readBase == "C") :
unmasked = True
if (unmasked and mismatch.referencePosition >= 0 and mismatch.referencePosition < utr.getLength()) :
isTC = True
readCount += 1
if (isTC) :
unmaskedTCCount += 1
if (isTrueTC) :
maskedTCCount += 1
containsSNP = 0
if (unmaskedTCCount != maskedTCCount) :
containsSNP = 1
print(utr.name + "\t" + str(readCount) + "\t" + str(unmaskedTCCount) + "\t" + str(maskedTCCount) + "\t" + str(containsSNP), file=fileCSV)
progress += 1
fileCSV.close()
if(not checkStep([outputCSV], [outputPDF], force)):
print("Skipped computing T->C per UTR position plot for file " + bam, file=log)
else:
callR(getPlotter("SNPeval") + " -i " + outputCSV + " -c " + str(coverageCutoff) + " -v " + str(variantFraction) + " -o " + outputPDF, log, dry=printOnly, verbose=verbose)
def tcPerUtr(referenceFile, utrBed, 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 UTR position for file " + bam, file=log)
else:
counter = 0
totalUtrCountFwd = [0] * utrNormFactor
totalUtrCountRev = [0] * utrNormFactor
tcPerPosRev = [0] * utrNormFactor
tcPerPosFwd = [0] * utrNormFactor
allPerPosRev = [0] * utrNormFactor
allPerPosFwd = [0] * utrNormFactor
snps = SNPtools.SNPDictionary(snpsFile)
snps.read()
# Go through one utr after the other
testFile = SlamSeqBamFile(bam, referenceFile, snps)
for utr in BedIterator(utrBed):
readIterator = testFile.readInRegion(utr.chromosome, utr.start, utr.stop, utr.strand, maxReadLength, minQual)
tcForwardCounts = [0] * utrNormFactor
mutForwardCounts = [0] * utrNormFactor
tcReverseCounts = [0] * utrNormFactor
mutReverseCounts = [0] * utrNormFactor
for read in readIterator:
tcCounts = [0] * utrNormFactor
mutCounts = [0] * utrNormFactor
for mismatch in read.mismatches:
mismatchPos = mismatch.referencePosition
# mismatchPos = read.startRefPos
if (utr.strand == "+") :
# New try for UTRs (remove + 1
if (mismatchPos >= (utr.getLength() - utrNormFactor) and mismatchPos < utr.getLength()) :
# if (mismatchPos >= (utr.getLength() - utrNormFactor) and mismatchPos < utr.getLength() + 1) :
mismatchPos = utrNormFactor - (utr.getLength() - mismatchPos)
if(mismatch.isTCMismatch(read.direction == ReadDirection.Reverse)):
tcCounts[mismatchPos] += 1
else :
mutCounts[mismatchPos] += 1
else :
if (mismatchPos >= 0 and mismatchPos < min(utr.getLength(), utrNormFactor)) :
if(mismatch.isTCMismatch(read.direction == ReadDirection.Reverse)):
tcCounts[mismatchPos] += 1
else :
mutCounts[mismatchPos] += 1
if(read.direction == ReadDirection.Reverse):
tcReverseCounts = sumLists(tcReverseCounts, tcCounts)
mutReverseCounts = sumLists(mutReverseCounts, mutCounts)
start = max(0, min(min(utr.getLength(), utrNormFactor), read.startRefPos))
end = max(0, min(min(utr.getLength(), utrNormFactor), read.endRefPos))
for i in range(start, end):
totalUtrCountRev[i] += 1
else:
tcForwardCounts = sumLists(tcForwardCounts, tcCounts)
mutForwardCounts = sumLists(mutForwardCounts, mutCounts)
start = min(utr.getLength(), max(utr.getLength() - utrNormFactor, read.startRefPos))
end = min(utr.getLength(), max(utr.getLength() - utrNormFactor, read.endRefPos))
for i in range(start, end):
normPos = utrNormFactor - (utr.getLength() - i)
totalUtrCountFwd[normPos] += 1
tcPerPosFwd = sumLists(tcPerPosFwd, tcForwardCounts)
allPerPosFwd = sumLists(allPerPosFwd, mutForwardCounts)
tcPerPosRev = sumLists(tcPerPosRev, tcReverseCounts)
allPerPosRev = sumLists(allPerPosRev, mutReverseCounts)
counter += 1
if (verbose and counter % 10000 == 0) :
print("Handled " + str(counter) + " UTRs.", file=log)
foTC = open(outputCSV, "w")
print("# slamdunk tcperutr v" + __version__, file=foTC)
reverseAllPerPosRev = allPerPosRev[::-1]
reverseTcPerPosRev = tcPerPosRev[::-1]
reverseTotalUtrCountRev = totalUtrCountRev[::-1]
for i in range(0, utrNormFactor):
print(allPerPosFwd[i], reverseAllPerPosRev[i], tcPerPosFwd[i], reverseTcPerPosRev[i], totalUtrCountFwd[i], reverseTotalUtrCountRev[i], sep='\t', file=foTC)
foTC.close()
if(not checkStep([outputCSV], [outputPDF], force)):
print("Skipped computing T->C per UTR position plot for file " + bam, file=log)
else:
callR(getPlotter("conversion_per_read_position") + " -u -i " + outputCSV + " -o " + outputPDF, log, dry=printOnly, verbose=verbose)
