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 readSummary(filteredFiles,
countDirectory,
outputFile,
log,
printOnly=False,
verbose=True,
force=False):
# Print sort by ID
contentDict = {}
tsvFile = open(outputFile, "w")
print("# slamdunk summary v" + __version__, file=tsvFile)
if (countDirectory != None):
f = tempfile.NamedTemporaryFile(delete=False)
for bam in filteredFiles:
slamseqInfo = SlamSeqInfo(bam)
sampleInfo = getSampleInfo(bam)
if (countDirectory != None):
countedReads = 0
countFile = os.path.join(
countDirectory,
replaceExtension(os.path.basename(bam), ".tsv", "_tcount"))
if not os.path.exists(countFile):
print(
"TCount directory does not seem to contain tcount file for:\t"
+ countFile)
else:
print(sampleInfo.Name, countFile, sep='\t', file=f)
countedReads = sumCounts(countFile)
if (sampleInfo.ID in contentDict):
ID = len(contentDict) + 1
else:
ID = sampleInfo.ID
contentDict[int(ID)] = "\t".join([
bam, sampleInfo.Name, sampleInfo.Type, sampleInfo.Time,
str(slamseqInfo.SequencedReads),
str(slamseqInfo.MappedReads),
str(slamseqInfo.DedupReads),
str(slamseqInfo.MQFilteredReads),
str(slamseqInfo.IdFilteredReads),
str(slamseqInfo.NmFilteredReads),
str(slamseqInfo.MultimapperReads),
str(slamseqInfo.FilteredReads),
str(countedReads), slamseqInfo.AnnotationName
])
else:
if (sampleInfo.ID in contentDict):
ID = len(contentDict) + 1
else:
ID = sampleInfo.ID
contentDict[int(ID)] = "\t".join([
bam, sampleInfo.Name, sampleInfo.Type, sampleInfo.Time,
str(slamseqInfo.SequencedReads),
str(slamseqInfo.MappedReads),
str(slamseqInfo.DedupReads),
str(slamseqInfo.MQFilteredReads),
str(slamseqInfo.IdFilteredReads),
str(slamseqInfo.NmFilteredReads),
str(slamseqInfo.MultimapperReads),
str(slamseqInfo.FilteredReads), slamseqInfo.AnnotationName
])
if (countDirectory != None):
f.close()
callR(getPlotter("PCAPlotter") + " -f " + f.name + " -O " +
replaceExtension(outputFile, ".pdf", "_PCA") + " -P " +
replaceExtension(outputFile, ".txt", "_PCA"),
log,
dry=printOnly,
verbose=verbose)
print("FileName",
"SampleName",
"SampleType",
"SampleTime",
"Sequenced",
"Mapped",
"Deduplicated",
"MQ-Filtered",
"Identity-Filtered",
"NM-Filtered",
"Multimap-Filtered",
"Retained",
"Counted",
"Annotation",
sep="\t",
file=tsvFile)
else:
print("FileName",
"SampleName",
"SampleType",
"SampleTime",
"Sequenced",
"Mapped",
"Deduplicated",
"MQ-Filtered",
"Identity-Filtered",
"NM-Filtered",
"Multimap-Filtered",
"Retained",
"Annotation",
sep="\t",
file=tsvFile)
for key in sorted(contentDict):
print(contentDict[key], file=tsvFile)
tsvFile.close()
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 Filter(inputBAM, outputBAM, log, bed, MQ=2, minIdentity=0.8, NM=-1, printOnly=False, verbose=True, force=False):
if(printOnly or checkStep([inputBAM], [outputBAM], force)):
mappedReads = 0
unmappedReads = 0
filteredReads = 0
mqFiltered = 0
idFiltered = 0
nmFiltered = 0
multimapper = 0
infile = pysam.AlignmentFile(inputBAM, "rb")
outfile = pysam.AlignmentFile(outputBAM, "wb", template=infile)
# Default filtering without bed
if (bed == None) :
print("#No bed-file supplied. Running default filtering on " + inputBAM + ".",file=log)
for read in infile:
if(not read.is_secondary and not read.is_supplementary):
if(read.is_unmapped):
unmappedReads += 1
else:
mappedReads += 1
if(read.is_unmapped):
continue
if(read.mapping_quality < MQ):
mqFiltered += 1
continue
if(float(read.get_tag("XI")) < minIdentity):
idFiltered += 1
continue
if(NM > -1 and int(read.get_tag("NM")) > NM):
nmFiltered += 1
continue
if(not read.is_secondary and not read.is_supplementary):
filteredReads += 1
outfile.write(read)
print("Criterion\tFiltered reads",file=log)
print("MQ < " + str(MQ) + "\t" + str(mqFiltered),file=log)
print("ID < " + str(minIdentity) + "\t" + str(idFiltered),file=log)
print("NM > " + str(NM) + "\t" + str(nmFiltered),file=log)
print("MM\t0",file=log)
else :
# Multimap retention strategy filtering when bed is supplied
random.seed(1)
print("#Bed-file supplied. Running multimap retention filtering strategy on " + inputBAM + ".",file=log)
mappedReads, unmappedReads, filteredReads, mqFiltered, idFiltered, nmFiltered, multimapper = multimapUTRRetainment (infile, outfile, bed, minIdentity, NM, log)
#mappedReads, unmappedReads, filteredReads = multimapUTRRetainment (infile, outfile, bed, minIdentity, NM, log)
# Add number of sequenced and number of mapped reads to the read group description
# Used for creating summary file
inFileBamHeader = outfile.header
if('RG' in inFileBamHeader and len(inFileBamHeader['RG']) > 0):
slamseqInfo = SlamSeqInfo()
slamseqInfo.SequencedReads = mappedReads + unmappedReads
slamseqInfo.MappedReads = mappedReads
slamseqInfo.FilteredReads = filteredReads
slamseqInfo.MQFilteredReads = mqFiltered
slamseqInfo.IdFilteredReads = idFiltered
slamseqInfo.NmFilteredReads = nmFiltered
slamseqInfo.MultimapperReads = multimapper
if (bed != None) :
slamseqInfo.AnnotationName = os.path.basename(bed)
slamseqInfo.AnnotationMD5 = md5(bed)
else :
slamseqInfo.AnnotationName = ""
slamseqInfo.AnnotationMD5 = ""
if not isinstance(inFileBamHeader, dict):
inFileBamHeader = inFileBamHeader.to_dict()
inFileBamHeader['RG'][0]['DS'] = str(slamseqInfo)
#inFileBamHeader['RG'][0]['DS'] = "{'sequenced':" + str(mappedReads + unmappedReads) + "," + "'mapped':" + str(mappedReads) + "," + "'filtered':" + str(filteredReads) + "}"
slamDunkPG = { 'ID': 'slamdunk', 'PN': 'slamdunk filter v' + __version__, 'VN': __bam_version__ }
if('PG' in inFileBamHeader):
inFileBamHeader['PG'].append(slamDunkPG)
else:
inFileBamHeader['PG'] = [ slamDunkPG ]
infile.close()
outfile.close()
# Sort afterwards
bamSort(outputBAM, log, inFileBamHeader, verbose)
pysamIndex(outputBAM)
#pysamFlagstat(outputBAM)
#runFlagstat(outputBAM, log, verbose=verbose, dry=printOnly)
else:
print("Skipped filtering for " + inputBAM, file=log)
def Filter(inputBAM,
outputBAM,
log,
bed,
MQ=2,
minIdentity=0.8,
NM=-1,
printOnly=False,
verbose=True,
force=False):
if (printOnly or checkStep([inputBAM], [outputBAM], force)):
mappedReads = 0
unmappedReads = 0
filteredReads = 0
mqFiltered = 0
idFiltered = 0
nmFiltered = 0
multimapper = 0
infile = pysam.AlignmentFile(inputBAM, "rb")
outfile = pysam.AlignmentFile(outputBAM, "wb", template=infile)
# Default filtering without bed
if (bed == None):
print("#No bed-file supplied. Running default filtering on " +
inputBAM + ".",
file=log)
for read in infile:
if (not read.is_secondary and not read.is_supplementary):
if (read.is_unmapped):
unmappedReads += 1
else:
mappedReads += 1
if (read.is_unmapped):
continue
if (read.mapping_quality < MQ):
mqFiltered += 1
continue
if (float(read.get_tag("XI")) < minIdentity):
idFiltered += 1
continue
if (NM > -1 and int(read.get_tag("NM")) > NM):
nmFiltered += 1
continue
if (not read.is_secondary and not read.is_supplementary):
filteredReads += 1
outfile.write(read)
print("Criterion\tFiltered reads", file=log)
print("MQ < " + str(MQ) + "\t" + str(mqFiltered), file=log)
print("ID < " + str(minIdentity) + "\t" + str(idFiltered),
file=log)
print("NM > " + str(NM) + "\t" + str(nmFiltered), file=log)
print("MM\t0", file=log)
else:
# Multimap retention strategy filtering when bed is supplied
random.seed(1)
print(
"#Bed-file supplied. Running multimap retention filtering strategy on "
+ inputBAM + ".",
file=log)
mappedReads, unmappedReads, filteredReads, mqFiltered, idFiltered, nmFiltered, multimapper = multimapUTRRetainment(
infile, outfile, bed, minIdentity, NM, log)
#mappedReads, unmappedReads, filteredReads = multimapUTRRetainment (infile, outfile, bed, minIdentity, NM, log)
# Add number of sequenced and number of mapped reads to the read group description
# Used for creating summary file
inFileBamHeader = outfile.header
if ('RG' in inFileBamHeader and len(inFileBamHeader['RG']) > 0):
slamseqInfo = SlamSeqInfo()
slamseqInfo.SequencedReads = mappedReads + unmappedReads
slamseqInfo.MappedReads = mappedReads
slamseqInfo.FilteredReads = filteredReads
slamseqInfo.MQFilteredReads = mqFiltered
slamseqInfo.IdFilteredReads = idFiltered
slamseqInfo.NmFilteredReads = nmFiltered
slamseqInfo.MultimapperReads = multimapper
if (bed != None):
slamseqInfo.AnnotationName = os.path.basename(bed)
slamseqInfo.AnnotationMD5 = md5(bed)
else:
slamseqInfo.AnnotationName = ""
slamseqInfo.AnnotationMD5 = ""
if not isinstance(inFileBamHeader, dict):
inFileBamHeader = inFileBamHeader.to_dict()
inFileBamHeader['RG'][0]['DS'] = str(slamseqInfo)
#inFileBamHeader['RG'][0]['DS'] = "{'sequenced':" + str(mappedReads + unmappedReads) + "," + "'mapped':" + str(mappedReads) + "," + "'filtered':" + str(filteredReads) + "}"
slamDunkPG = {
'ID': 'slamdunk',
'PN': 'slamdunk filter v' + __version__,
'VN': __bam_version__
}
if ('PG' in inFileBamHeader):
inFileBamHeader['PG'].append(slamDunkPG)
else:
inFileBamHeader['PG'] = [slamDunkPG]
infile.close()
outfile.close()
# Sort afterwards
bamSort(outputBAM, log, inFileBamHeader, verbose)
pysamIndex(outputBAM)
#pysamFlagstat(outputBAM)
#runFlagstat(outputBAM, log, verbose=verbose, dry=printOnly)
else:
print("Skipped filtering for " + inputBAM, file=log)
def Filter(inputBAM,
outputBAM,
log,
bed,
MQ=2,
minIdentity=0.8,
NM=-1,
printOnly=False,
verbose=True,
force=False,
paired=False):
inputBAM = os.path.expanduser(inputBAM)
outputBAM = os.path.expanduser(outputBAM)
if printOnly or checkStep([inputBAM], [outputBAM], force):
(mappedReads, unmappedReads, filteredReads, mqFiltered, idFiltered,
nmFiltered, multimapper) = 0, 0, 0, 0, 0, 0, 0
infile = pysam.AlignmentFile(inputBAM, "rb")
outfile = pysam.AlignmentFile(outputBAM, "wb", template=infile)
# Default filtering without bed
if bed is None:
print("#No bed-file supplied. Running default filtering on " +
inputBAM + ".",
file=log)
if paired:
read1 = None
read2 = None
for read in infile:
if paired:
if not read.is_paired or read.mate_is_unmapped or read.is_duplicate:
unmappedReads += 1
continue
if read.is_read2:
read2 = read
else:
read1 = read
read2 = None
continue
if not read.is_secondary and not read.is_supplementary:
if read.is_unmapped:
unmappedReads += 1
continue
else:
mappedReads += 1
if not paired:
if read.mapping_quality < MQ:
mqFiltered += 1
continue
if float(read.get_tag("XI")) < minIdentity:
idFiltered += 1
continue
if -1 < NM < int(read.get_tag("NM")):
nmFiltered += 1
continue
filteredReads += 1
outfile.write(read)
else:
if read1 is None or read2 is None:
continue
if read1.query_name != read2.query_name:
continue
if read1.mapping_quality < MQ and read2.mapping_quality < MQ:
mqFiltered += 1
continue
if float(read1.get_tag("XI")) < minIdentity and float(
read2.get_tag("XI")) < minIdentity:
idFiltered += 1
continue
if -1 < NM < int(read1.get_tag("NM")) and -1 < NM < int(
read2.get_tag("NM")):
nmFiltered += 1
continue
filteredReads += 1
outfile.write(read1)
outfile.write(read2)
print("Criterion\tFiltered reads", file=log)
print("MQ < 0\t0", file=log)
print("ID < %s\t%s" % (minIdentity, idFiltered), file=log)
print("NM > %s\t%s" % (NM, nmFiltered), file=log)
print("MM\t0", file=log)
else:
# Multimap retention strategy filtering when bed is supplied
print(
"#Bed-file supplied. Running multimap retention filtering strategy on "
+ inputBAM + ".",
file=log)
(mappedReads, unmappedReads, filteredReads, mqFiltered, idFiltered,
nmFiltered,
multimapper) = multimapUTRRetainment(infile, outfile, bed,
minIdentity, NM, MQ, log)
# Add number of sequenced and number of mapped reads to the read group description
# Used for creating summary file
inFileBamHeader = outfile.header
if "RG" in inFileBamHeader and len(inFileBamHeader["RG"]) > 0:
slamseqInfo = SlamSeqInfo()
slamseqInfo.SequencedReads = mappedReads + unmappedReads
slamseqInfo.MappedReads = mappedReads
slamseqInfo.FilteredReads = filteredReads
slamseqInfo.MQFilteredReads = mqFiltered
slamseqInfo.IdFilteredReads = idFiltered
slamseqInfo.NmFilteredReads = nmFiltered
slamseqInfo.MultimapperReads = multimapper
if bed:
slamseqInfo.AnnotationName = os.path.basename(bed)
slamseqInfo.AnnotationMD5 = md5(bed)
else:
slamseqInfo.AnnotationName = ""
slamseqInfo.AnnotationMD5 = ""
if not isinstance(inFileBamHeader, dict):
inFileBamHeader = inFileBamHeader.to_dict()
inFileBamHeader["RG"][0]["DS"] = str(slamseqInfo)
slamDunkPG = {
"ID": "slamdunk",
"PN": "slamdunk filter v" + __version__,
"VN": __bam_version__
}
if "PG" in inFileBamHeader:
inFileBamHeader["PG"].append(slamDunkPG)
else:
inFileBamHeader["PG"] = [slamDunkPG]
infile.close()
outfile.close()
# Sort afterwards
bamSort(outputBAM, log, inFileBamHeader, paired=False, verbose=verbose)
if not paired:
pysamIndex(outputBAM)
else:
print("Skipped filtering for " + inputBAM, file=log)