def _mainLoop(self):
# Split up reference genome into chunks and farm out the
# a chunk as a unit of work.
logging.debug("Starting main loop.")
ids = reference.enumerateIds(options.referenceWindows)
for _id in ids:
if options.fancyChunking:
chunks = reference.fancyEnumerateChunks(self._inCmpH5,
_id,
options.referenceChunkSize,
options.minCoverage,
options.minMapQV,
options.referenceWindows)
else:
chunks = reference.enumerateChunks(_id,
options.referenceChunkSize,
options.referenceWindows)
for chunk in chunks:
if self._aborting: return
try:
self._workQueue.put(chunk, True, options.queueTimeout)
except:
return
# Write sentinels ("end-of-work-stream")
for i in xrange(options.numWorkers):
self._workQueue.put(None)
def _mainLoop(self):
# Split up reference genome into chunks and farm out the
# a chunk as a unit of work.
logging.debug("Starting main loop.")
ids = reference.enumerateIds(options.referenceWindows)
for _id in ids:
if options.fancyChunking:
chunks = reference.fancyEnumerateChunks(
self._inAlnFile, _id, options.referenceChunkSize,
options.minCoverage, options.minMapQV,
options.referenceWindows)
else:
chunks = reference.enumerateChunks(_id,
options.referenceChunkSize,
options.referenceWindows)
for chunk in chunks:
if self._aborting: return
self._workQueue.put(chunk)
# Write sentinels ("end-of-work-stream")
for i in xrange(options.numWorkers):
self._workQueue.put(None)
def getReads(cmpH5, reference, interval, paddedTemplateWidth, depthLimit, real_quiver=False):
minMapQV = 10
minPoaCoverage = 3
maxPoaCoverage = 11
mutationSeparation = 10
mutationNeighborhood = 20
maxIterations = 20
refineDinucleotideRepeats = True
noEvidenceConsensus = "nocall"
computeConfidence = True
readStumpinessThreshold = 0.1
refId = [x for x in reference.enumerateIds()][0]
refSeq = reference.byId[refId].sequence
refWindow = (refId, 0, reference.byId[refId].length)
intStart, intEnd = interval
subWin = subWindow(refWindow, interval)
windowRefSeq = refSeq[intStart:intEnd]
rows = readsInWindow(cmpH5, subWin,
depthLimit = depthLimit,
minMapQV = minMapQV,
strategy = "longest",
stratum = None,
barcode = None)
#print([cmpH5[row].alignedLength for row in rows if cmpH5[row].spansReferenceRange(intStart, intEnd)])
spanningRows = [row for row in rows if cmpH5[row].spansReferenceRange(intStart, intEnd) ]
alns = cmpH5[spanningRows]
clippedAlns_ = [ aln.clippedTo(*interval) for aln in alns ]
clippedAlns__ = [ aln for aln in clippedAlns_ if aln.alignedLength <= paddedTemplateWidth - 7]
clippedAlns = filterAlns(subWin, clippedAlns__, readStumpinessThreshold)
# Compute the POA consensus, which is our initial guess, and
# should typically be > 99.5% accurate
fwdSequences = [ a.read(orientation="genomic", aligned=False)
for a in clippedAlns]
p = cc.PoaConsensus.FindConsensus(fwdSequences[:maxPoaCoverage])
template = p.Sequence()
tmplSeq = np.zeros((paddedTemplateWidth), dtype=np.uint8)
tmplOrds = map(ord, template)
tmplSeq[:len(tmplOrds)] = tmplOrds
#read pos y, read x
readSeqs = np.zeros((paddedTemplateWidth, len(clippedAlns)), dtype=np.uint8)
for i in xrange(len(clippedAlns)):
alnOrds = map(ord, fwdSequences[i])
readSeqs[:len(alnOrds), i] = alnOrds
#uint8
#metric z, read pos y, read x
qvInfo = np.zeros((8, paddedTemplateWidth, len(clippedAlns)), dtype=np.uint8)
for i in xrange(len(clippedAlns)):
qvInfo[0, :clippedAlns[i].readLength, i] = clippedAlns[i].InsertionQV(orientation="genomic", aligned=False)
qvInfo[1, :clippedAlns[i].readLength, i] = clippedAlns[i].MergeQV(orientation="genomic", aligned=False)
qvInfo[2, :clippedAlns[i].readLength, i] = clippedAlns[i].DeletionQV(orientation="genomic", aligned=False)
qvInfo[3, :clippedAlns[i].readLength, i] = clippedAlns[i].DeletionTag(orientation="genomic", aligned=False)
qvInfo[4, :clippedAlns[i].readLength, i] = clippedAlns[i].SubstitutionQV(orientation="genomic", aligned=False)
if real_quiver:
return template, len(tmplOrds), fwdSequences, qvInfo
else:
return tmplSeq, len(tmplOrds), readSeqs, qvInfo
def getReads(cmpH5,
reference,
interval,
paddedTemplateWidth,
depthLimit,
real_quiver=False):
minMapQV = 10
minPoaCoverage = 3
maxPoaCoverage = 11
mutationSeparation = 10
mutationNeighborhood = 20
maxIterations = 20
refineDinucleotideRepeats = True
noEvidenceConsensus = "nocall"
computeConfidence = True
readStumpinessThreshold = 0.1
refId = [x for x in reference.enumerateIds()][0]
refSeq = reference.byId[refId].sequence
refWindow = (refId, 0, reference.byId[refId].length)
intStart, intEnd = interval
subWin = subWindow(refWindow, interval)
windowRefSeq = refSeq[intStart:intEnd]
rows = readsInWindow(cmpH5,
subWin,
depthLimit=depthLimit,
minMapQV=minMapQV,
strategy="longest",
stratum=None,
barcode=None)
#print([cmpH5[row].alignedLength for row in rows if cmpH5[row].spansReferenceRange(intStart, intEnd)])
spanningRows = [
row for row in rows
if cmpH5[row].spansReferenceRange(intStart, intEnd)
]
alns = cmpH5[spanningRows]
clippedAlns_ = [aln.clippedTo(*interval) for aln in alns]
clippedAlns__ = [
aln for aln in clippedAlns_
if aln.alignedLength <= paddedTemplateWidth - 7
]
clippedAlns = filterAlns(subWin, clippedAlns__, readStumpinessThreshold)
# Compute the POA consensus, which is our initial guess, and
# should typically be > 99.5% accurate
fwdSequences = [
a.read(orientation="genomic", aligned=False) for a in clippedAlns
]
p = cc.PoaConsensus.FindConsensus(fwdSequences[:maxPoaCoverage])
template = p.Sequence()
tmplSeq = np.zeros((paddedTemplateWidth), dtype=np.uint8)
tmplOrds = map(ord, template)
tmplSeq[:len(tmplOrds)] = tmplOrds
#read pos y, read x
readSeqs = np.zeros((paddedTemplateWidth, len(clippedAlns)),
dtype=np.uint8)
for i in xrange(len(clippedAlns)):
alnOrds = map(ord, fwdSequences[i])
readSeqs[:len(alnOrds), i] = alnOrds
#uint8
#metric z, read pos y, read x
qvInfo = np.zeros((8, paddedTemplateWidth, len(clippedAlns)),
dtype=np.uint8)
for i in xrange(len(clippedAlns)):
qvInfo[0, :clippedAlns[i].readLength,
i] = clippedAlns[i].InsertionQV(orientation="genomic",
aligned=False)
qvInfo[1, :clippedAlns[i].readLength,
i] = clippedAlns[i].MergeQV(orientation="genomic",
aligned=False)
qvInfo[2, :clippedAlns[i].readLength,
i] = clippedAlns[i].DeletionQV(orientation="genomic",
aligned=False)
qvInfo[3, :clippedAlns[i].readLength,
i] = clippedAlns[i].DeletionTag(orientation="genomic",
aligned=False)
qvInfo[4, :clippedAlns[i].readLength,
i] = clippedAlns[i].SubstitutionQV(orientation="genomic",
aligned=False)
if real_quiver:
return template, len(tmplOrds), fwdSequences, qvInfo
else:
return tmplSeq, len(tmplOrds), readSeqs, qvInfo
options = dummy()
options.diploid = False
options.parametersFile = "/home/nick/workspace/btry6790_project/venv/lib/python2.7/site-packages/GenomicConsensus/quiver/resources/2013-09/GenomicConsensus/QuiverParameters.ini"
options.parameterSet = "best"
options.refineDinucleotideRepeats = True
options.noEvidenceConsensusCall = "nocall"
options.minMapQV = 10
options.fastMode = False
cmpH5 = CmpH5Reader('/home/nick/workspace/btry6790_project/PXO99A_ref_wo_one_copy_212kb_repeat.cmp.h5')
quiverConfig = configure(options, cmpH5)
depthLimit = 100
reference.loadFromFile("/home/nick/workspace/btry6790_project/ref_PXO99A_genome_reference_wo_one_copy_212k_repeat/sequence/ref_PXO99A_genome_reference_wo_one_copy_212k_repeat.fasta", cmpH5)
refId = [x for x in reference.enumerateIds()][0]
refSeq = reference.byId[refId].sequence
refWindow = (refId, 0, reference.byId[refId].length)
def run_real_quiver(cmpH5, quiverConfig, interval, depthLimit, refSeq, refWindow, seedConsensus):
intStart, intEnd = interval
subWin = subWindow(refWindow, interval)
windowRefSeq = refSeq[intStart:intEnd]
rows = readsInWindow(cmpH5, subWin,
depthLimit = depthLimit,
minMapQV = quiverConfig.minMapQV,
strategy = "longest",
stratum = None,
barcode = None)