def blasr(query, target, format, nproc = 1, outname = "out.m5", consensus=True):
"""
Simple mapper
"""
cmd = ("blasr %s %s %s -nproc %d -bestn 1 -out %s ") \
% (query, target, format, nproc, outname)
#need to figure out how to m5-pie it...maybe
if consensus:
r, o, e = exe(cmd + BLASRPARAMS)
else:
r, o, e = exe(cmd + EEBLASRPARAMS)
logging.debug("blasr - %d - %s - %s" % (r, o, e))
def blasr(query, target, format, nproc=1, outname="out.m5", consensus=True):
"""
Simple mapper
"""
cmd = ("blasr %s %s %s --nproc %d --bestn 1 --out %s ") \
% (query, target, format, nproc, outname)
#need to figure out how to m5-pie it...maybe
if consensus:
r, o, e = exe(cmd + BLASRPARAMS)
else:
r, o, e = exe(cmd + EEBLASRPARAMS)
logging.debug("blasr - %d - %s - %s" % (r, o, e))
def samToFastq( inSam, outFq ):
"""
Creates input.fastq from SAM file
"""
return exe(('grep -v "^@" %s | '
'awk \'{print "@" $1 "\\n" $10 "\\n+\\n" $11}\' '
'> %s') % (inSam, outFq))
def parseArgs(self):
"""
Uses OptionParser to parse out input
Jelly.py <stage> <protocol>
"""
parser = OptionParser(USAGE)
parser.remove_option("-h")
parser.add_option("-h", "--help", action="store_true", default=False)
parser.add_option("--debug", action="store_true", default=False)
parser.add_option("-x", dest="extras", type="string", default="", \
help="-x \"<options>\" are options to pass into the stage you're running")
self.options, args = parser.parse_args()
if self.options.help == True:
if len(args) == 1:
if args[0] in STAGES:
print(exe(Stages.PRINT_HELPS[args[0]])[1])
sys.exit(0)
#Else, this will drop down to the next parser.error
else:
print(parser.format_help())
sys.exit(0)
if len(args) != 2 or args[0] not in STAGES:
parser.error("Invalid Arguments. Expected one of\n'%s'" %
"', '".join(STAGES))
sys.exit(1)
self.executeStage = args[0]
self.protocolName = os.path.abspath(args[1])
def blasr(query, target, nproc=1, outname="out.m5"):
"""
Simple overlapper
"""
r, o, e = exe(("blasr %s %s -m 5 --bestn 200 --nCandidates 200 --minMatch 12 "
"--affineExtend 3 --nproc %d --noSplitSubreads --out %s --maxScore -1000") % \
(query, target, nproc, outname))
def remapReads(reads, outName):
"""
remaps reads to the provided reference (only setup for hg19 -- see
global variable reference)
"""
return exe("blasr {0} {1} -sa {1}.sa -nproc 4 -out {2} -sam -bestn 1"\
.format(reads, reference, outName))
def callBlasr(inFile, refFile, params, nproc=1, outFile="map.sam"):
"""
fq = input file
automatically search for .sa
"""
if os.path.exists(refFile+".sa"):
sa = "-sa " + refFile + ".sa"
else:
sa = ""
logging.info("Running Blasr")
cmd = ("blasr %s %s %s -nproc %d -bestn 1 "
"-sam -clipping subread -out %s ") \
% (inFile, refFile, sa, nproc, outFile)
r, o, e = exe(cmd + params)
#r,o,e = exe(("blasr %s %s %s -nproc %d -sam -bestn 1 -nCandidates 20 "
#"-out %s -clipping soft -minPctIdentity 75 "
#" -noSplitSubreads") % (fq, ref, sa, nproc, out))
if r != 0:
logging.error("blasr mapping failed!")
logging.error("RETCODE %d" % (r))
logging.error("STDOUT %s" % (str(o)))
logging.error("STDERR %s" % (str(e)))
logging.error("Exiting")
exit(r)
logging.info(str([r, o, e]))
def remapReads( reads, outName):
"""
remaps reads to the provided reference (only setup for hg19 -- see
global variable reference)
"""
return exe("blasr {0} {1} -sa {1}.sa -nproc 4 -out {2} -sam -bestn 1"\
.format(reads, reference, outName))
def samToFastq(inSam, outFq):
"""
Creates input.fastq from SAM file
"""
return exe(('grep -v "^@" %s | '
'awk \'{print "@" $1 "\\n" $10 "\\n+\\n" $11}\' '
'> %s') % (inSam, outFq))
def sam2bam( fn ):
"""
Creates BAM from SAM (only setup for hg19 -- see global variable reference)
"""
name = fn[:-4]
return exe(("samtools view -bt {0} {1} | samtools sort - {2}.sort && "
"mv {2}.sort.bam {2}.bam && "
"samtools index {2}.bam").format(reference, fn, name))
def blasr(query, target, format, nproc=1, outname="out.m5", consensus=True):
"""
Simple mapper
"""
cmd = ("blasr %s %s %s -nproc %d -bestn 1 -out %s ") \
% (query, target, format, nproc, outname)
#need to figure out how to m5-pie it...maybe
if consensus:
r, o, e = exe(cmd + " -noSplitSubreads -minMatch 5 " + \
"-nCandidates 20 -sdpTupleSize 6 -insertion 1 -deletion 1 -bestn 1")
else:
r, o, e = exe(cmd + " -maxAnchorsPerPosition 100 "
"-affineAlign -affineOpen 100 -affineExtend 0 "
"-insertion 10 -deletion 10 "
"-noSplitSubreads -nCandidates 20 ")
logging.debug("blasr - %d - %s - %s" % (r, o, e))
def sam2bam(fn):
"""
Creates BAM from SAM (only setup for hg19 -- see global variable reference)
"""
name = fn[:-4]
return exe(("samtools view -bt {0} {1} | samtools sort - {2}.sort && "
"mv {2}.sort.bam {2}.bam && "
"samtools index {2}.bam").format(reference, fn, name))
def blasr(query, target, format, nproc = 1, outname = "out.m5", consensus=True):
"""
Simple mapper
"""
cmd = ("blasr %s %s %s -nproc %d -bestn 1 -out %s ") \
% (query, target, format, nproc, outname)
#need to figure out how to m5-pie it...maybe
if consensus:
r, o, e = exe(cmd + " -noSplitSubreads -minMatch 5 " + \
"-nCandidates 20 -sdpTupleSize 6 -insertion 1 -deletion 1 -bestn 1")
else:
r, o, e = exe(cmd + " -maxAnchorsPerPosition 100 "
"-affineAlign -affineOpen 100 -affineExtend 0 "
"-insertion 10 -deletion 10 "
"-noSplitSubreads -nCandidates 20 ")
logging.debug("blasr - %d - %s - %s" % (r, o, e))
def mapTails(fq, ref, nproc=1, out="tailmap.sam", useSa=True):
"""
automatically search for .sa
"""
if os.path.exists(ref + ".sa") and useSa:
sa = "--sa " + ref + ".sa"
else:
sa = ""
cmd = ("blasr %s %s %s --nproc %d -m 4 --bestn 1 --nCandidates 20 --out %s"
" --minPctIdentity 75 --sdpTupleSize 6 --noSplitSubreads") \
% (fq, ref, sa, nproc, out)
logging.debug(cmd)
r, o, e = exe(cmd)
if r != 0:
logging.error("blasr mapping failed!")
logging.error("RETCODE %d" % (r))
logging.error("STDOUT %s" % (str(o)))
logging.error("STDERR %s" % (str(e)))
logging.error("Exiting")
exit(r)
logging.info(str([r, o, e]))
def mapTails(fq, ref, nproc=1, out="tailmap.sam", useSa=True):
"""
automatically search for .sa
"""
if os.path.exists(ref+".sa") and useSa:
sa = "--sa " + ref + ".sa"
else:
sa = ""
cmd = ("blasr %s %s %s --nproc %d -m 4 --bestn 1 --nCandidates 20 --out %s"
" --minPctIdentity 75 --sdpTupleSize 6 --noSplitSubreads") \
% (fq, ref, sa, nproc, out)
logging.debug(cmd)
r,o,e = exe(cmd)
if r != 0:
logging.error("blasr mapping failed!")
logging.error("RETCODE %d" % (r))
logging.error("STDOUT %s" % (str(o)))
logging.error("STDERR %s" % (str(e)))
logging.error("Exiting")
exit(r)
logging.info(str([r, o, e]))
def consensusCalling(self, spot, bam, reference, args):
"""
Make a consensus of all the reads in the region and identify all of the SVs in the region
"""
#
MAXNUMREADS = 100 #I don't think we'll need more than this many reads
MAXATTEMPTS = MAXNUMREADS/2 #I don't feel like trying 100 times
SPANBUFFER = 100 #number of bases I want a read to span
chrom, start, end = spot.chrom, spot.start, spot.end
buffer = args.buffer
supportReads = []
spanReads = []
#Fetch reads and trim
totCnt = 0
for read in bam.fetch(chrom, max(0, start-buffer-SPANBUFFER), end+buffer+SPANBUFFER):
if read.qname not in spot.varReads:
continue
seq, qual = self.readTrim(read, start-buffer, end+buffer)
if read.pos < start-SPANBUFFER and read.aend > end+SPANBUFFER:
spanReads.append((len(seq), seq, qual))
else:
supportReads.append((seq, qual))
totCnt += 1
if len(spanReads) == 0:
logging.debug("noone spans - consensus aborted. %s" % (str(spot)))
spot.tags["noSpan"] = True
return [spot]
spanReads.sort(reverse=True)
if len(spanReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads[:MAXNUMREADS]]
elif len(spanReads) + len(supportReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads] + supportReads[:MAXNUMREADS-len(spanReads)]
else:
origSupportReads = [(x[1], x[2]) for x in spanReads] + supportReads
logging.debug("Alt reads: %d total, %d extra support" % (totCnt, len(origSupportReads)))
mySpots = []
refReadId = 0
haveVar = False
#Attempt each spanRead until we get one that passes
#while refReadId < len(spanReads) and not haveVar and refReadId < MAXATTEMPTS:
#refread = spanReads[refReadId]
#supportReads = origSupportReads[:refReadId] + origSupportReads[refReadId+1:]
refReadId += 1
#read that spans most of the region goes first
#use the rest for cleaning
#building consensus sequence
foutreads = NamedTemporaryFile(suffix=".fasta")
qoutreads = open(foutreads.name + '.qual', 'w')
for id, i in enumerate(origSupportReads):
foutreads.write(">%d\n%s\n" % (id, i[0]))
qoutreads.write(">%d\n%s\n" % (id, " ".join(str(ord(j)-33) for j in i[1])))
foutreads.flush()
qoutreads.flush()
#foutref = NamedTemporaryFile(suffix=".fasta")
#foutref.write(">%s:%d-%d\n%s" % (spot.chrom, start, end, refread[1]))
#foutref.flush()
logging.debug("Making the contig....")
#run it through phrap
#make out.fasta and out.fasta.qual
#run phrap
#if asm -- consensus only
r, o, e = exe("phrap %s -minmatch 6 -minscore 20" % (foutreads.name), timeout=3)
if r != 0:#failed
logging.warning('phrap failed ' + self.name)
logging.warning(o)
logging.warning(e)
return [] #here is where I'd like to add just the no-consensus spot
results = mergeFastaQual(foutreads.name + ".contigs", foutreads.name + ".contigs.qual")
if len(results) == 0:
logging.warning('no asm made ' + self.name)
return [] #here is where I'd like to add just the no-consensus spot
logging.info('%d contigs made %s' % (len(results), self.name))
#then run it through consensus
logging.debug("Polishing contigs")
alignOut = NamedTemporaryFile(suffix=".m5")
blasr(foutreads.name, foutreads.name + ".contigs", format="-m 5", nproc=1, outname=alignOut.name)
# elif no asm and consensus only (faster)
if args.polish == "pbbanana":
aligns = M5File(alignOut.name)
con = ">con\n%s\n" % consensus(aligns).sequence
conName = "pbbanana"
elif args.polish == "pbdagcon":
logging.debug("pbdagcon is running")
#using minerrreads - 1 because one f them is already being used as seed!
r, con, e = exe("pbdagcon -c %d -t 0 %s" % (max(0, args.minErrReads - 1), alignOut.name), timeout=1)
#r, con, e = exe("pbdagcon %s" % (alignOut.name), timeout=2)
logging.debug("back from pbdagcon")
logging.debug((r,e))
#raw_input("press ent")
if con is not None:
con = con[con.index("\n")+1:]
else:
con = ""
conName = "pbdagcon"
alignOut.close()
#foutref.close()
foutreads.close()
#we don't have a consensus - retry
if len(con) == 0:
logging.debug("Trying another seed read for consensus")
con = results.values()[0].seq
logging.debug("%s %d bp seq" % (conName, len(con.split('\n')[1])))
#try improving consensus
conOut = NamedTemporaryFile(suffix=".fasta")
conOut.write(con)
#conOut.close()
conOut.flush()
refOut = NamedTemporaryFile(suffix=".fasta")
#j = reference.fetch(chrom, max(0, start-buffer), end+buffer)
#fout = open("f****e.ref.fasta",'w')
#fout.write(j)
#fout.close()
refOut.write(">%s:%d-%d\n%s\n" % (chrom, start, end, \
reference.fetch(chrom, max(0, start-buffer), end+buffer)))
refOut.flush()
#map consensus to refregion
varSam = NamedTemporaryFile(suffix=".sam")
blasr(conOut.name, refOut.name, format="-sam", outname=varSam.name)
#consensus=False) -- would this help?
#or what if I fed it through leftalign?
sam = pysam.Samfile(varSam.name)
matches = 0.0
bases = 0.0
nReads = 0
mySpots = []
for read in sam:
nReads += 1
spot.tags["consensusCreated"] = True
for svstart, svsize, svtype, altseq in expandCigar(read, args.minIndelSize, CONFIRMCOLLAPSE, True):
newspot = copy.deepcopy(spot)
if spot.svtype == svtype and svtype == "INS":
haveVar = True
newspot.start = svstart + start - buffer
newspot.end = svstart + start - buffer
newspot.tags["seq"] = altseq
newspot.size = svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
mySpots.append(newspot)
elif spot.svtype == svtype and svtype == "DEL":
haveVar = True
newspot.start = svstart + start - buffer
newspot.end = svstart + svsize + start - buffer
newspot.size = -svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
newspot.tags["seq"] = reference.fetch(chrom, newspot.start, newspot.end)
mySpots.append(newspot)
#identity = matches/bases
#If no var, nothing is returned.
#for newspot in mySpots:
#newspot.tags["alnIdentityEstimate"] = identity
#Keep reporting the actual contigs out until we
#find a reason to need it (and also we can get quals...)
#vbam.reset()
#for id, read in enumerate(vbam):
#newspot.tags["contigSeq%d" % (id)] = read.seq
#newspot.tags["contigQual%d" % (id)] = read.qual
#vbam.close()
#varBam.close()
refOut.close()
logging.debug("%d consensus reads created %d spots" % (nReads, len(mySpots)))
return mySpots
def consensusCalling(self, spot, bam, reference, args):
"""
Make a consensus of all the reads in the region and identify all of the SVs in the region
"""
#
MAXNUMREADS = 100 #I don't think we'll need more than this many reads
MAXATTEMPTS = 5 #MAXNUMREADS/2 #I don't feel like trying 100 times
SPANBUFFER = 100 #number of bases I want a read to span
chrom, start, end = spot.chrom, spot.start, spot.end
buffer = args.buffer
supportReads = []
spanReads = []
#Fetch reads and trim
totCnt = 0
for read in bam.fetch(chrom, max(0, start - buffer - SPANBUFFER),
end + buffer + SPANBUFFER):
if read.qname not in spot.varReads:
continue
seq, qual = self.readTrim(read, start - buffer, end + buffer)
if read.pos < start - SPANBUFFER and read.aend > end + SPANBUFFER:
sz = spot.varReadsSize[spot.varReads.index(read.qname)]
spanReads.append((abs(sz - spot.tags["szMedian"]), seq, qual))
else:
supportReads.append((seq, qual))
totCnt += 1
if len(spanReads) == 0:
logging.debug("noone spans - consensus aborted. %s" % (str(spot)))
spot.tags["noSpan"] = True
return [spot]
#spanReads.sort(reverse=True)
spanReads.sort()
if len(spanReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads[:MAXNUMREADS]]
elif len(spanReads) + len(supportReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads
] + supportReads[:MAXNUMREADS - len(spanReads)]
else:
origSupportReads = [(x[1], x[2]) for x in spanReads] + supportReads
mySpots = []
refReadId = 0
haveVar = False
#Attempt each spanRead until we get one that passes
while refReadId < len(
spanReads) and not haveVar and refReadId < MAXATTEMPTS:
refread = spanReads[refReadId]
supportReads = origSupportReads[:refReadId] + origSupportReads[
refReadId + 1:]
refReadId += 1
#read that spans most of the region goes first
#use the rest for cleaning
#building consensus sequence
foutreads = NamedTemporaryFile(suffix=".fastq")
for id, i in enumerate(supportReads):
foutreads.write("@%d\n%s\n+\n%s\n" % (id, i[0], i[1]))
foutreads.flush()
foutref = NamedTemporaryFile(suffix=".fasta")
foutref.write(">%s:%d-%d\n%s" %
(spot.chrom, start, end, refread[1]))
foutref.flush()
alignOut = NamedTemporaryFile(suffix=".m5")
logging.debug("making the contig....")
#run it through phrap
#then run it through consensus
blasr(foutreads.name,
foutref.name,
format="-m 5",
nproc=1,
outname=alignOut.name)
if args.consensus == "pbbanana":
aligns = M5File(alignOut.name)
con = ">con\n%s\n" % consensus(aligns).sequence
conName = "pbbanana"
elif args.consensus == "pbdagcon":
logging.debug("pbdagcon is running")
#using minerreads - 1 because one f them is already being used as seed!
#I want to be sure I get something out... so just require somebody on there
#r, con, e = exe("pbdagcon -c %d -t 0 %s" % (1, alignOut.name), timeout=1)
#r, con, e = exe("pbdagcon -m 100 -c %d -t 0 %s" % (max(args.minErrReads - 1, 0), alignOut.name), timeout=1)
r, con, e = exe("pbdagcon -m 100 -c %d -t 0 %s" %
(3, alignOut.name),
timeout=1)
logging.debug("back from pbdagcon")
logging.debug((r, e))
#raw_input("press ent")
if con is not None:
con = con[con.index("\n") + 1:]
else:
con = ""
conName = "pbdagcon"
alignOut.close()
foutref.close()
foutreads.close()
#we don't have a consensus - retry
if len(con) == 0:
logging.debug("Trying another seed read for consensus")
continue
logging.debug("%s %d bp seq" % (conName, len(con.split('\n')[1])))
#try improving consensus
conOut = NamedTemporaryFile(suffix=".fasta")
conOut.write(con)
#conOut.close()
conOut.flush()
refOut = NamedTemporaryFile(suffix=".fasta")
#j = reference.fetch(chrom, max(0, start-buffer), end+buffer)
#fout = open("f****e.ref.fasta",'w')
#fout.write(j)
#fout.close()
refOut.write(">%s:%d-%d\n%s\n" % (chrom, start, end, \
reference.fetch(chrom, max(0, start-(buffer*2)), end+(buffer*2))))
refOut.flush()
#map consensus to refregion
varSam = NamedTemporaryFile(suffix=".sam")
blasr(conOut.name, refOut.name, format="-sam", outname=varSam.name,\
consensus=False) #-- would this help?
#or what if I fed it through leftalign?
#os.system("cp %s ." % (refOut.name))
#os.system("cp %s ." % (varSam.name))
sam = pysam.Samfile(varSam.name)
matches = 0.0
bases = 0.0
nReads = 0
minVarDiff = 10000
for read in sam:
localSpots = []
nReads += 1
spot.tags["consensusCreated"] = True
for svstart, svsize, svtype, altseq in expandCigar(
read, args.minIndelSize, CONFIRMCOLLAPSE, True):
newspot = copy.deepcopy(spot)
if spot.svtype == svtype and svtype == "INS":
#haveVar = True
newspot.start = svstart + start - (buffer * 2)
newspot.end = svstart + start - (buffer * 2)
newspot.tags["seq"] = altseq
newspot.size = svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
if abs(spot.tags["szMedian"] -
newspot.size) < minVarDiff:
minVarDiff = abs(spot.tags["szMedian"] -
newspot.size)
if args.reportContig:
newspot.tags["contigseq"] = read.seq
newspot.tags["contigqual"] = read.qual
localSpots.append(newspot)
elif spot.svtype == svtype and svtype == "DEL":
#haveVar = True
newspot.start = svstart + start - (buffer * 2)
newspot.end = svstart + svsize + start - (buffer * 2)
newspot.size = svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
newspot.tags["seq"] = reference.fetch(
chrom, newspot.start, newspot.end)
if abs(spot.tags["szMedian"] -
newspot.size) < minVarDiff:
minVarDiff = abs(spot.tags["szMedian"] -
newspot.size)
if args.reportContig:
newspot.tags["contigseq"] = read.seq
newspot.tags["contigqual"] = read.qual
localSpots.append(newspot)
if len(localSpots) > 0:
mySpots.append((minVarDiff, localSpots))
#identity = matches/bases
#If no var, nothing is returned.
#for newspot in mySpots:
#newspot.tags["alnIdentityEstimate"] = identity
#Keep reporting the actual contigs out until we
#find a reason to need it (and also we can get quals...)
#vbam.reset()
#for id, read in enumerate(vbam):
#newspot.tags["contigSeq%d" % (id)] = read.seq
#newspot.tags["contigQual%d" % (id)] = read.qual
#vbam.close()
#varBam.close()
refOut.close()
#logging.debug("%d consensus reads created %d spots" % (nReads, len(localSpots)))
if len(mySpots) == 0:
return []
mySpots.sort()
return mySpots[0][1]
#!/usr/bin/env python
from pbsuite.utils.FileHandlers import FastqFile, M5File
from pbsuite.utils.CommandRunner import exe
"""
This can be run inside of an assembly folder and create our filling sequence into
polish.out.fasta
"""
if __name__ == '__main__':
input = FastqFile("input.fastq")
fout = open("ref.fasta",'w')
for i in input.values():
if i.name.startswith("ref"):
fout.write(">%s\n%s\n" % (i.name, i.seq))
fout.close()
print exe(("blasr input.fastq ref.fasta --bestn 2 -m 5 --noSplitSubreads > out.m5"))
print exe(("python /stornext/snfs5/next-gen/scratch/english/Jelly/"
"DevJelly/branches/consensusDev/GetSubs.py out.m5 input.fastq"))
print exe(("python /stornext/snfs5/next-gen/scratch/english/Jelly/"
"DevJelly/branches/sv/pbjPolish.py "
"reads.fastq seed.fasta -n 4 -l"))
#!/usr/bin/python
from pbsuite.utils.FileHandlers import FastqFile, M5File
from pbsuite.utils.CommandRunner import exe
"""
This can be run inside of an assembly folder and create our filling sequence into
polish.out.fasta
"""
if __name__ == '__main__':
input = FastqFile("input.fastq")
fout = open("ref.fasta",'w')
for i in input.values():
if i.name.startswith("ref"):
fout.write(">%s\n%s\n" % (i.name, i.seq))
fout.close()
print exe(("blasr input.fastq ref.fasta --bestn 2 -m 5 --noSplitSubreads > out.m5"))
print exe(("python /stornext/snfs5/next-gen/scratch/english/Jelly/"
"DevJelly/branches/consensusDev/GetSubs.py out.m5 input.fastq"))
print exe(("python /stornext/snfs5/next-gen/scratch/english/Jelly/"
"DevJelly/branches/sv/pbjPolish.py "
"reads.fastq seed.fasta -n 4 -l"))
def blasr(query, target, nproc=1, bestn=1, outName="map.m5"):
"""
runs blasr
"""
r,o,e = exe("blasr %s %s --bestn %d --affineAlign -m 5 --nproc %d --out %s" \
% (query, target, bestn, nproc, outName))
def blasr(query, target, nproc=1, bestn=1, outName="map.m5"):
"""
runs blasr
"""
r,o,e = exe("blasr %s %s --bestn %d --affineAlign -m 5 --nproc %d --out %s" \
% (query, target, bestn, nproc, outName))
def bam2sam(fn, outName):
"""
Turns a bam to a sam
"""
return exe("samtools view -h %s > %s " % (fn, outName))
def consensusCalling(self, spot, bam, reference, args):
"""
Make a consensus of all the reads in the region and identify all of the SVs in the region
"""
#
MAXNUMREADS = 100 #I don't think we'll need more than this many reads
MAXATTEMPTS = 5 #MAXNUMREADS/2 #I don't feel like trying 100 times
SPANBUFFER = 100 #number of bases I want a read to span
chrom, start, end = spot.chrom, spot.start, spot.end
buffer = args.buffer
supportReads = []
spanReads = []
#Fetch reads and trim
totCnt = 0
for read in bam.fetch(chrom, max(0, start-buffer-SPANBUFFER), end+buffer+SPANBUFFER):
if read.qname not in spot.varReads:
continue
seq, qual = self.readTrim(read, start-buffer, end+buffer)
if read.pos < start-SPANBUFFER and read.aend > end+SPANBUFFER:
sz = spot.varReadsSize[spot.varReads.index(read.qname)]
spanReads.append((abs(sz - spot.tags["szMedian"]), seq, qual))
else:
supportReads.append((seq, qual))
totCnt += 1
if len(spanReads) == 0:
logging.debug("noone spans - consensus aborted. %s" % (str(spot)))
spot.tags["noSpan"] = True
return [spot]
#spanReads.sort(reverse=True)
spanReads.sort()
if len(spanReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads[:MAXNUMREADS]]
elif len(spanReads) + len(supportReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads] + supportReads[:MAXNUMREADS-len(spanReads)]
else:
origSupportReads = [(x[1], x[2]) for x in spanReads] + supportReads
mySpots = []
refReadId = 0
haveVar = False
#Attempt each spanRead until we get one that passes
while refReadId < len(spanReads) and not haveVar and refReadId < MAXATTEMPTS:
refread = spanReads[refReadId]
supportReads = origSupportReads[:refReadId] + origSupportReads[refReadId+1:]
refReadId += 1
#read that spans most of the region goes first
#use the rest for cleaning
#building consensus sequence
foutreads = NamedTemporaryFile(suffix=".fastq")
for id, i in enumerate(supportReads):
foutreads.write("@%d\n%s\n+\n%s\n" % (id, i[0], i[1]))
foutreads.flush()
foutref = NamedTemporaryFile(suffix=".fasta")
foutref.write(">%s:%d-%d\n%s" % (spot.chrom, start, end, refread[1]))
foutref.flush()
alignOut = NamedTemporaryFile(suffix=".m5")
logging.debug("making the contig....")
#run it through phrap
#then run it through consensus
blasr(foutreads.name, foutref.name, format="-m 5", nproc=1, outname=alignOut.name)
if args.consensus == "pbbanana":
aligns = M5File(alignOut.name)
con = ">con\n%s\n" % consensus(aligns).sequence
conName = "pbbanana"
elif args.consensus == "pbdagcon":
logging.debug("pbdagcon is running")
#using minerreads - 1 because one f them is already being used as seed!
#I want to be sure I get something out... so just require somebody on there
#r, con, e = exe("pbdagcon -c %d -t 0 %s" % (1, alignOut.name), timeout=1)
#r, con, e = exe("pbdagcon -m 100 -c %d -t 0 %s" % (max(args.minErrReads - 1, 0), alignOut.name), timeout=1)
r, con, e = exe("pbdagcon -m 100 -c %d -t 0 %s" % (3, alignOut.name), timeout=1)
logging.debug("back from pbdagcon")
logging.debug((r,e))
#raw_input("press ent")
if con is not None:
con = con[con.index("\n")+1:]
else:
con = ""
conName = "pbdagcon"
alignOut.close()
foutref.close()
foutreads.close()
#we don't have a consensus - retry
if len(con) == 0:
logging.debug("Trying another seed read for consensus")
continue
logging.debug("%s %d bp seq" % (conName, len(con.split('\n')[1])))
#try improving consensus
conOut = NamedTemporaryFile(suffix=".fasta")
conOut.write(con)
#conOut.close()
conOut.flush()
refOut = NamedTemporaryFile(suffix=".fasta")
#j = reference.fetch(chrom, max(0, start-buffer), end+buffer)
#fout = open("f****e.ref.fasta",'w')
#fout.write(j)
#fout.close()
refOut.write(">%s:%d-%d\n%s\n" % (chrom, start, end, \
reference.fetch(chrom, max(0, start-(buffer*2)), end+(buffer*2))))
refOut.flush()
#map consensus to refregion
varSam = NamedTemporaryFile(suffix=".sam")
blasr(conOut.name, refOut.name, format="-sam", outname=varSam.name,\
consensus=False) #-- would this help?
#or what if I fed it through leftalign?
#os.system("cp %s ." % (refOut.name))
#os.system("cp %s ." % (varSam.name))
sam = pysam.Samfile(varSam.name)
matches = 0.0
bases = 0.0
nReads = 0
minVarDiff = 10000
for read in sam:
localSpots = []
nReads += 1
spot.tags["consensusCreated"] = True
for svstart, svsize, svtype, altseq in expandCigar(read, args.minIndelSize, CONFIRMCOLLAPSE, True):
newspot = copy.deepcopy(spot)
if spot.svtype == svtype and svtype == "INS":
#haveVar = True
newspot.start = svstart + start - (buffer*2)
newspot.end = svstart + start - (buffer*2)
newspot.tags["seq"] = altseq
newspot.size = svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
if abs(spot.tags["szMedian"] - newspot.size) < minVarDiff:
minVarDiff = abs(spot.tags["szMedian"] - newspot.size)
if args.reportContig:
newspot.tags["contigseq"] = read.seq
newspot.tags["contigqual"] = read.qual
localSpots.append(newspot)
elif spot.svtype == svtype and svtype == "DEL":
#haveVar = True
newspot.start = svstart + start - (buffer*2)
newspot.end = svstart + svsize + start - (buffer*2)
newspot.size = svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
newspot.tags["seq"] = reference.fetch(chrom, newspot.start, newspot.end)
if abs(spot.tags["szMedian"] - newspot.size) < minVarDiff:
minVarDiff = abs(spot.tags["szMedian"] - newspot.size)
if args.reportContig:
newspot.tags["contigseq"] = read.seq
newspot.tags["contigqual"] = read.qual
localSpots.append(newspot)
if len(localSpots) > 0:
mySpots.append((minVarDiff, localSpots))
#identity = matches/bases
#If no var, nothing is returned.
#for newspot in mySpots:
#newspot.tags["alnIdentityEstimate"] = identity
#Keep reporting the actual contigs out until we
#find a reason to need it (and also we can get quals...)
#vbam.reset()
#for id, read in enumerate(vbam):
#newspot.tags["contigSeq%d" % (id)] = read.seq
#newspot.tags["contigQual%d" % (id)] = read.qual
#vbam.close()
#varBam.close()
refOut.close()
#logging.debug("%d consensus reads created %d spots" % (nReads, len(localSpots)))
if len(mySpots) == 0:
return []
mySpots.sort()
return mySpots[0][1]
def assemble(inputFq, workDir):
return exe("OLCAssembly.py %s --nproc 4 --fqOut --workDir %s" %
(inputFq, workDir))
def pileup(bam):
"""
create a pileup from the bam
"""
return exe("samtools mpileup -f {0} {1} > {1}.plup".format(reference, bam))
def grabReads( inputBam, entry, outFn ):
"""
Gets all of the reads for a region and puts them into outFn
"""
return exe("samtools view -h %s %s > %s" % (inputBam, entry.region, outFn))
def bam2sam( fn, outName):
"""
Turns a bam to a sam
"""
return exe("samtools view -h %s > %s " % (fn, outName))
def __assemble(self):
"""
writes temp files
assembles
reads results
clears temp files
returns results as a string
Calls the assembler
"""
self.myTmpFiles = []
#Temporary Files
fout = tempfile.NamedTemporaryFile(prefix="spades_pe1",
suffix=".fastq",
delete=False,
dir=self.tmpDir,
mode="w")
self.myTmpFiles.append(fout.name)
for name, seq, qual in self.leftReads:
fout.write("@%s\n%s\n+\n%s\n" % (name, seq, qual))
fout.close()
fout2 = tempfile.NamedTemporaryFile(prefix="spades_pe2",
suffix=".fastq",
delete=False,
dir=self.tmpDir,
mode="w")
self.myTmpFiles.append(fout2.name)
for name, seq, qual in self.rightReads:
fout2.write("@%s\n%s\n+\n%s\n" % (name, seq, qual))
fout2.close()
foutp = tempfile.NamedTemporaryFile(prefix="spades_pb",
suffix=".fastq",
delete=False,
dir=self.tmpDir,
mode="w")
self.myTmpFiles.append(foutp.name)
for name, seq, qual in self.pbReads:
foutp.write("@%s\n%s\n+\n%s\n" % (name, seq, qual))
foutp.close()
#working here
resultOut = tempfile.mkdtemp(prefix="spades", dir=self.tmpDir)
estSize = self.buffer * 2
if self.data.rest[0] != 'DEL':
estSize += int(self.data.rest[1])
#r, o, e = exe("dipspades.py -1 {pe1} -2 {pe2} --pacbio {pacbio} -o {output} "\
r, o, e = exe("spades.py -1 {pe1} -2 {pe2} --pacbio {pacbio} -o {output} "\
.format(pe1=fout.name, pe2=fout2.name, pacbio=foutp.name, output=resultOut), \
timeout=self.timeout)
logging.debug("RET - %d\nOUT - %s\nERR- %s" % (r, o, e))
#just the output dir, maybe?
self.myTmpFiles.append(resultOut)
if r == 214:
super(SpadesAssembler, self).cleanupTmp()
return "Failure - Assembly Timeout " + self.data.name
outFsta = os.path.join(resultOut, "dipspades",
"consensus_contigs.fasta")
fasta = FastaFile(outFsta)
results = {}
for key in fasta:
results[key] = FastqEntry(key, fasta[key], '?' * len(fasta[key]))
#save to file
fout = tempfile.NamedTemporaryFile(prefix = "asm" + self.data.name, mode="w", \
suffix=".fastq", delete=False, dir=self.tmpDir)
for key in results:
fout.write("@group" + self.data.name + "_" + key + "\n" + \
results[key].seq + '\n+\n' + \
results[key].qual + '\n')
fout.close()
self.results = fout.name
#clean up
super(SpadesAssembler, self).cleanupTmp()
return self.results
def assemble(inputFq, workDir):
return exe("OLCAssembly.py %s --nproc 4 --fqOut --workDir %s" % (inputFq, workDir))
def run(self):
#Fasta Ref Output
scaffTempName = self.scaffInput + ".tempFasta"
scaffOutput = open(scaffTempName, 'w')
#Qual Ref Output
if self.qualInput is not None:
qualTempName = self.qualInput + ".tempQual"
qualOutput = open(qualTempName, 'w')
#Gaps Output
if self.opts.gapOutput is not None:
gapTableOut = open(self.opts.gapOutput, 'w')
else:
gapTableOut = False
logging.info(
"Creating reference sequence index names and identifying gaps")
refTemplate = "ref%07d"
refId = 1
#Read References
reference = FastaFile(self.scaffInput)
if self.qualInput is not None:
qualReference = QualFile(self.qualInput)
for key in reference:
scaffIndex = refTemplate % refId
scaffName = key.replace(' ', '_')
refId += 1
scaffName = scaffName + "|" + scaffIndex
scaffOutput.write(">" + scaffName + "\n" + wrap(reference[key]) +
"\n")
if self.qualInput is not None:
qualOutput.write(">" + scaffName + "\n" +
qwrap(qualReference[key]) + "\n")
gapCoords = []
for gap in re.finditer("[^Nn]([Nn]{%d,%s})[^Nn]" % \
(self.opts.minGap, self.opts.maxGap), reference[key]):
gapCoords.append([gap.start() + 1, gap.end() - 1])
if len(gapCoords) == 0: #no Gaps
gapTableOut.write("\t".join(
[scaffName, 'na', 'na', scaffIndex + "_0_0", '3']) + '\n')
logging.debug("Scaffold %s is empty" % scaffName)
continue
#Consolidate gaps that are too close -- indicating LQ regions.
i = 0
while i < len(gapCoords) - 1:
if gapCoords[i + 1][0] - gapCoords[i][1] < 25:
gapCoords[i + 1][0] = gapCoords[i][0]
del (gapCoords[i])
else:
i += 1
prevEnd = 0 #Contig Start Tracking
idx = 0
#Make the first gap
prevEnd = gapCoords[0][1]
gapCoords[0][1] - gapCoords[0][0]
flag = Gap.BEGIN
if len(gapCoords) == 1:
flag += Gap.END
if gapTableOut:
gapTableOut.write("%s\t%i\t%i\t%s_%i_%i\t%d\n" \
% (scaffName, gapCoords[0][0], gapCoords[0][1], scaffIndex, idx, idx+1, flag))
#Now Go Through the rest of the gaps
for i in range(1, len(gapCoords)):
idx += 1
prevEnd = gapCoords[i][1]
gapCoords[i][1] - gapCoords[i][0]
if gapTableOut:
if i == len(gapCoords) - 1:
flag = Gap.END
else:
flag = 0
gapTableOut.write("%s\t%i\t%i\t%s_%i_%i\t%d\n" \
% (scaffName, gapCoords[i][0], gapCoords[i][1], scaffIndex, idx, idx+1, flag))
#Close shop
scaffOutput.close()
os.rename(self.scaffInput, self.scaffInput + ".original")
os.rename(scaffTempName, self.scaffInput)
if self.qualInput is not None:
qualOutput.close()
os.rename(self.qualInput, self.qualInput + ".original")
os.rename(qualTempName, self.qualInput)
if gapTableOut:
gapTableOut.close()
if self.opts.index:
logging.info("Creating .sa indexes for references")
r, o, e = exe("sawriter %s.sa %s" %
(self.scaffInput, self.scaffInput))
if r != 0:
logging.error("sawriter returned %d" % r)
logging.error("Ensure it's in your path")
exit(1)
logging.debug(str(o) + ' ' + str(e))
logging.info("Finished!")
def grabReads(inputBam, entry, outFn):
"""
Gets all of the reads for a region and puts them into outFn
"""
return exe("samtools view -h %s %s > %s" % (inputBam, entry.region, outFn))
def run(self):
#Fasta Ref Output
scaffTempName = self.scaffInput+".tempFasta"
scaffOutput = open(scaffTempName, 'w')
#Qual Ref Output
if self.qualInput is not None:
qualTempName= self.qualInput+".tempQual"
qualOutput = open(qualTempName, 'w')
#Gaps Output
if self.opts.gapOutput is not None:
gapTableOut = open(self.opts.gapOutput,'w')
else:
gapTableOut = False
logging.info("Creating reference sequence index names and identifying gaps")
refTemplate = "ref%07d"
refId = 1
#Read References
reference = FastaFile(self.scaffInput)
if self.qualInput is not None:
qualReference = QualFile(self.qualInput)
for key in reference:
scaffIndex = refTemplate % refId
scaffName = key.replace(' ','_')
refId += 1
scaffName = scaffName + "|" + scaffIndex
scaffOutput.write(">"+scaffName+"\n"+wrap(reference[key])+"\n")
if self.qualInput is not None:
qualOutput.write(">"+scaffName+"\n"+qwrap(qualReference[key])+"\n")
gapCoords = []
for gap in re.finditer("[^Nn]([Nn]{%d,%s})[^Nn]" % \
(self.opts.minGap, self.opts.maxGap), reference[key]):
gapCoords.append([gap.start() + 1, gap.end() - 1])
if len(gapCoords) == 0:#no Gaps
gapTableOut.write("\t".join([scaffName, 'na', 'na', scaffIndex+"_0_0", '3'])+'\n')
logging.debug("Scaffold %s is empty" % scaffName)
continue
#Consolidate gaps that are too close -- indicating LQ regions.
i = 0
while i < len(gapCoords)-1:
if gapCoords[i+1][0] - gapCoords[i][1] < 25:
gapCoords[i+1][0] = gapCoords[i][0]
del(gapCoords[i])
else:
i += 1
prevEnd = 0#Contig Start Tracking
idx = 0
#Make the first gap
prevEnd = gapCoords[0][1]
gapCoords[0][1]-gapCoords[0][0]
flag = Gap.BEGIN
if len(gapCoords) == 1:
flag += Gap.END
if gapTableOut:
gapTableOut.write("%s\t%i\t%i\t%s_%i_%i\t%d\n" \
% (scaffName, gapCoords[0][0], gapCoords[0][1], scaffIndex, idx, idx+1, flag))
#Now Go Through the rest of the gaps
for i in range(1, len(gapCoords)):
idx += 1
prevEnd = gapCoords[i][1]
gapCoords[i][1]-gapCoords[i][0]
if gapTableOut:
if i == len(gapCoords)-1:
flag = Gap.END
else:
flag = 0
gapTableOut.write("%s\t%i\t%i\t%s_%i_%i\t%d\n" \
% (scaffName, gapCoords[i][0], gapCoords[i][1], scaffIndex, idx, idx+1, flag))
#Close shop
scaffOutput.close()
os.rename(self.scaffInput, self.scaffInput+".original")
os.rename(scaffTempName, self.scaffInput)
if self.qualInput is not None:
qualOutput.close()
os.rename(self.qualInput, self.qualInput+".original")
os.rename(qualTempName, self.qualInput)
if gapTableOut:
gapTableOut.close()
if self.opts.index:
logging.info("Creating .sa indexes for references")
r, o, e = exe("sawriter %s.sa %s" % (self.scaffInput, self.scaffInput))
if r != 0:
logging.error("sawriter returned %d" % r)
logging.error("Ensure it's in your path")
exit(1)
logging.debug(str(o) + ' ' + str(e))
logging.info("Finished!")
def pileup( bam ):
"""
create a pileup from the bam
"""
return exe("samtools mpileup -f {0} {1} > {1}.plup".format(reference, bam))
def consensusCalling(self, spot, bam, reference, args):
"""
Make a consensus of all the reads in the region and identify all of the SVs in the region
"""
#
MAXNUMREADS = 100 #I don't think we'll need more than this many reads
MAXATTEMPTS = MAXNUMREADS / 2 #I don't feel like trying 100 times
SPANBUFFER = 100 #number of bases I want a read to span
chrom, start, end = spot.chrom, spot.start, spot.end
buffer = args.buffer
supportReads = []
spanReads = []
#Fetch reads and trim
totCnt = 0
for read in bam.fetch(chrom, max(0, start - buffer - SPANBUFFER),
end + buffer + SPANBUFFER):
if read.qname not in spot.varReads:
continue
seq, qual = self.readTrim(read, start - buffer, end + buffer)
if read.pos < start - SPANBUFFER and read.aend > end + SPANBUFFER:
spanReads.append((len(seq), seq, qual))
else:
supportReads.append((seq, qual))
totCnt += 1
if len(spanReads) == 0:
logging.debug("noone spans - consensus aborted. %s" % (str(spot)))
spot.tags["noSpan"] = True
return [spot]
spanReads.sort(reverse=True)
if len(spanReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads[:MAXNUMREADS]]
elif len(spanReads) + len(supportReads) > MAXNUMREADS:
origSupportReads = [(x[1], x[2]) for x in spanReads
] + supportReads[:MAXNUMREADS - len(spanReads)]
else:
origSupportReads = [(x[1], x[2]) for x in spanReads] + supportReads
logging.debug("Alt reads: %d total, %d extra support" %
(totCnt, len(origSupportReads)))
mySpots = []
refReadId = 0
haveVar = False
#Attempt each spanRead until we get one that passes
#while refReadId < len(spanReads) and not haveVar and refReadId < MAXATTEMPTS:
#refread = spanReads[refReadId]
#supportReads = origSupportReads[:refReadId] + origSupportReads[refReadId+1:]
refReadId += 1
#read that spans most of the region goes first
#use the rest for cleaning
#building consensus sequence
foutreads = NamedTemporaryFile(suffix=".fasta")
qoutreads = open(foutreads.name + '.qual', 'w')
for id, i in enumerate(origSupportReads):
foutreads.write(">%d\n%s\n" % (id, i[0]))
qoutreads.write(">%d\n%s\n" %
(id, " ".join(str(ord(j) - 33) for j in i[1])))
foutreads.flush()
qoutreads.flush()
#foutref = NamedTemporaryFile(suffix=".fasta")
#foutref.write(">%s:%d-%d\n%s" % (spot.chrom, start, end, refread[1]))
#foutref.flush()
logging.debug("Making the contig....")
#run it through phrap
#make out.fasta and out.fasta.qual
#run phrap
#if asm -- consensus only
r, o, e = exe("phrap %s -minmatch 6 -minscore 20" % (foutreads.name),
timeout=3)
if r != 0: #failed
logging.warning('phrap failed ' + self.name)
logging.warning(o)
logging.warning(e)
return [
] #here is where I'd like to add just the no-consensus spot
results = mergeFastaQual(foutreads.name + ".contigs",
foutreads.name + ".contigs.qual")
if len(results) == 0:
logging.warning('no asm made ' + self.name)
return [
] #here is where I'd like to add just the no-consensus spot
logging.info('%d contigs made %s' % (len(results), self.name))
#then run it through consensus
logging.debug("Polishing contigs")
alignOut = NamedTemporaryFile(suffix=".m5")
blasr(foutreads.name,
foutreads.name + ".contigs",
format="-m 5",
nproc=1,
outname=alignOut.name)
# elif no asm and consensus only (faster)
if args.polish == "pbbanana":
aligns = M5File(alignOut.name)
con = ">con\n%s\n" % consensus(aligns).sequence
conName = "pbbanana"
elif args.polish == "pbdagcon":
logging.debug("pbdagcon is running")
#using minerrreads - 1 because one f them is already being used as seed!
r, con, e = exe("pbdagcon -c %d -t 0 %s" %
(max(0, args.minErrReads - 1), alignOut.name),
timeout=1)
#r, con, e = exe("pbdagcon %s" % (alignOut.name), timeout=2)
logging.debug("back from pbdagcon")
logging.debug((r, e))
#raw_input("press ent")
if con is not None:
con = con[con.index("\n") + 1:]
else:
con = ""
conName = "pbdagcon"
alignOut.close()
#foutref.close()
foutreads.close()
#we don't have a consensus - retry
if len(con) == 0:
logging.debug("Trying another seed read for consensus")
con = results.values()[0].seq
logging.debug("%s %d bp seq" % (conName, len(con.split('\n')[1])))
#try improving consensus
conOut = NamedTemporaryFile(suffix=".fasta")
conOut.write(con)
#conOut.close()
conOut.flush()
refOut = NamedTemporaryFile(suffix=".fasta")
#j = reference.fetch(chrom, max(0, start-buffer), end+buffer)
#fout = open("f****e.ref.fasta",'w')
#fout.write(j)
#fout.close()
refOut.write(">%s:%d-%d\n%s\n" % (chrom, start, end, \
reference.fetch(chrom, max(0, start-buffer), end+buffer)))
refOut.flush()
#map consensus to refregion
varSam = NamedTemporaryFile(suffix=".sam")
blasr(conOut.name, refOut.name, format="--sam", outname=varSam.name)
#consensus=False) -- would this help?
#or what if I fed it through leftalign?
sam = pysam.Samfile(varSam.name)
matches = 0.0
bases = 0.0
nReads = 0
mySpots = []
for read in sam:
nReads += 1
spot.tags["consensusCreated"] = True
for svstart, svsize, svtype, altseq in expandCigar(
read, args.minIndelSize, CONFIRMCOLLAPSE, True):
newspot = copy.deepcopy(spot)
if spot.svtype == svtype and svtype == "INS":
haveVar = True
newspot.start = svstart + start - buffer
newspot.end = svstart + start - buffer
newspot.tags["seq"] = altseq
newspot.size = svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
mySpots.append(newspot)
elif spot.svtype == svtype and svtype == "DEL":
haveVar = True
newspot.start = svstart + start - buffer
newspot.end = svstart + svsize + start - buffer
newspot.size = -svsize
gt, gq = genotype(newspot)
newspot.tags["GT"] = gt
newspot.tags["GQ"] = gq
newspot.tags["seq"] = reference.fetch(
chrom, newspot.start, newspot.end)
mySpots.append(newspot)
#identity = matches/bases
#If no var, nothing is returned.
#for newspot in mySpots:
#newspot.tags["alnIdentityEstimate"] = identity
#Keep reporting the actual contigs out until we
#find a reason to need it (and also we can get quals...)
#vbam.reset()
#for id, read in enumerate(vbam):
#newspot.tags["contigSeq%d" % (id)] = read.seq
#newspot.tags["contigQual%d" % (id)] = read.qual
#vbam.close()
#varBam.close()
refOut.close()
logging.debug("%d consensus reads created %d spots" %
(nReads, len(mySpots)))
return mySpots
def __assemble(self, reads):
"""
writes temp files
assembles
reads results
clears temp files
returns results as a string
Calls the assembler
"""
self.myTmpFiles = []
#Temporary Files
fout = tempfile.NamedTemporaryFile(suffix=".fasta", mode="w", delete=False, dir=self.tmpDir)
self.myTmpFiles.append(fout.name)
qout = open(fout.name + '.qual', 'w')
self.myTmpFiles.append(fout.name + '.qual')
for name, seq, qual in reads:
fout.write(">{0}\n{1}\n".format(name, seq))
qout.write(">{0}\n{1}\n".format(name, qual))
fout.close()
qout.close()
r, o, e = exe("phrap %s -minmatch 6 -minscore 20" % (fout.name),\
timeout=self.timeout)
self.myTmpFiles.extend([fout.name + ".contigs", fout.name + ".contigs.qual", \
fout.name + ".problems", fout.name + ".problems.qual", \
fout.name + ".log", fout.name + ".singlets"])
if r == 214:
super(PhrapAssembler, self).cleanupTmp()
return "Failure - Assembly Timeout " + self.data.name
results = mergeFastaQual(fout.name + ".contigs", fout.name + ".contigs.qual")
#Try to push the problems through, too
if os.stat(fout.name + '.problems').st_size != 0:
pfile = fout.name + ".problems"
r, o, e = exe("phrap %s -minmatch 6 -minscore 20" % (pfile), \
timeout=self.timeout)
self.myTmpFiles.extend([pfile + ".contigs", pfile + ".contigs.qual", \
pfile + ".problems", pfile + ".problems.qual", \
pfile + ".log", pfile + ".singlets"])
if r == 214:
super(PhrapAssembler, self).cleanupTmp()
return "Failure - Assembly Timeout " + self.data.name
results.update(mergeFastaQual(fout.name + ".problems.contigs", fout.name + ".problems.contigs.qual"))
#save to file
fout = tempfile.NamedTemporaryFile(prefix = "asm" + self.data.name, mode="w",\
suffix=".fastq", delete=False, dir=self.tmpDir)
for key in results:
fout.write("@group" + self.data.name + "_" + key + "\n" + \
results[key].seq + '\n+\n' + \
results[key].qual + '\n')
fout.close()
self.results = fout.name
#clean up
super(PhrapAssembler, self).cleanupTmp()
return self.results
