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)