def reverseComplement(self, preserveHeader=False):
"""
Return a new FastaRecord with the reverse-complemented DNA sequence.
Optionally, supply a name
"""
rcSequence = sequence.reverseComplement(self.sequence)
if preserveHeader:
return FastaRecord(self.header, rcSequence)
else:
rcName = '{0} [revcomp]'.format(self.header.strip())
return FastaRecord(rcName, rcSequence)
def reverseComplement(self, preserveHeader=False):
"""
Return a new FastaRecord with the reverse-complemented DNA sequence.
Optionally, supply a name
"""
rcSequence = sequence.reverseComplement(self.sequence)
if preserveHeader:
return FastaRecord(self.header, rcSequence)
else:
rcName = '{0} [revcomp]'.format(self.header.strip())
return FastaRecord(rcName, rcSequence)
def reference(self, aligned=True, orientation="native"):
if not (orientation == "native" or orientation == "genomic"):
raise ValueError, "Bad `orientation` value"
tSeq = self.bam.referenceFasta[self.referenceName].sequence[self.tStart : self.tEnd]
shouldRC = orientation == "native" and self.isReverseStrand
tSeqOriented = reverseComplement(tSeq) if shouldRC else tSeq
if aligned:
x = np.fromstring(tSeqOriented, dtype=np.int8)
y = self._gapifyRef(x, orientation)
return y.tostring()
else:
return tSeqOriented
def reference(self, aligned=True, orientation="native"):
if not (orientation == "native" or orientation == "genomic"):
raise ValueError, "Bad `orientation` value"
tSeq = self.bam.referenceFasta[self.referenceName].sequence[self.tStart:self.tEnd]
shouldRC = orientation == "native" and self.isReverseStrand
tSeqOriented = reverseComplement(tSeq) if shouldRC else tSeq
if aligned:
x = np.fromstring(tSeqOriented, dtype=np.int8)
y = self._gapifyRef(x, orientation)
return y.tostring()
else:
return tSeqOriented
def test_reverse_complement_error(self):
sequence.reverseComplement(self.bad_sequence)
def test_iupac(self):
assert_equal(self.iupac_complement,
sequence.complement(self.iupac_sequence))
assert_equal(self.iupac_reverse_complement,
sequence.reverseComplement(self.iupac_sequence))
def test_reverseComplement(self):
assert_equal(self.reverse_complement,
sequence.reverseComplement(self.sequence))
assert_equal(self.sequence,
sequence.reverseComplement(self.reverse_complement))
def test_reverse_complement_error(self):
sequence.reverseComplement(self.bad_sequence)
def test_iupac(self):
assert_equal(self.iupac_complement,
sequence.complement(self.iupac_sequence))
assert_equal(self.iupac_reverse_complement,
sequence.reverseComplement(self.iupac_sequence))
def test_reverseComplement(self):
assert_equal(self.reverse_complement,
sequence.reverseComplement(self.sequence))
assert_equal(self.sequence,
sequence.reverseComplement(self.reverse_complement))
def test_reverse_complement_error(self):
with pytest.raises(ValueError):
sequence.reverseComplement(self.BAD_SEQUENCE)
def test_iupac(self):
assert self.IUPAC_COMPLEMENT == sequence.complement(
self.IUPAC_SEQUENCE)
assert self.IUPAC_REVERSE_COMPLEMENT == sequence.reverseComplement(
self.IUPAC_SEQUENCE)
def test_reverseComplement(self):
assert self.REVERSE_COMPLEMENT == sequence.reverseComplement(
self.SEQUENCE)
assert self.SEQUENCE == sequence.reverseComplement(
self.REVERSE_COMPLEMENT)
def extractIPDonlyTgenomic(cbamfile, alncbamfile, sbamfile, zmwqueue, outqueue): #filebase, nit):
with pb.IndexedBamReader(alncbamfile) as alncbam, pb.IndexedBamReader(sbamfile) as sbam,\
pb.IndexedBamReader(cbamfile) as cbam:
zmw = zmwqueue.get()
while zmw is not None:
cc = cbam.readsByHoleNumber(zmw)[0]
ccread = cc.read(aligned=False, orientation='native')
subrs = sbam.readsByHoleNumber(zmw)
ccalns = alncbam.readsByHoleNumber(zmw)
zmres = {}
if len(ccalns) > 0:
alnlen = np.array([ccal.readLength for ccal in ccalns])
usealn = np.where(alnlen == np.max(alnlen))[0][0]
elif len(ccalns) == 1:
usealn = 0
elif len(ccalns) == 0:
usealn = None
allipds = np.empty((len(subrs), len(ccread)), dtype='float32')
allipds.fill(np.nan)
subrOrient = np.empty(len(subrs), dtype='bool')
for index, sr in enumerate(subrs):
# Test if this subread aligns to the forward or reverse of the CCS
forwardSread = sr.read(aligned=False, orientation='native')
reverseSread = reverseComplement(forwardSread)
faln = edlib.align(forwardSread, ccread, mode='NW', task='path')
raln = edlib.align(reverseSread, ccread, mode='NW', task='path')
if faln['editDistance'] < raln['editDistance']:
subrOrient[index] = True
alndir = faln
useread = forwardSread
else:
subrOrient[index] = False
alndir = raln
useread = reverseSread
# Use the alignment information to extract IPD at each base that aligns to the CCS
origb = np.empty(len(useread), dtype=np.int16 )
origb.fill(np.nan)
ccb = np.empty(len(useread), dtype=np.int16)
ccb.fill(np.nan)
subI = 0
ccI = 0
for m in p.finditer(alndir['cigar']):
lg = int(m.group()[-len(m.group()):-1])
mtype = m.group()[-1]
if mtype == '=':
origb[subI:(subI + lg)] = range(subI, subI + lg)
ccb[subI:(subI + lg)] = range(ccI, ccI + lg)
subI += lg
ccI += lg
elif mtype == 'X':
subI += lg
ccI += lg
elif mtype == 'I':
subI += lg
elif mtype == 'D':
ccI += lg
ccb = ccb[~np.isnan(ccb)]
origb = origb[~np.isnan(origb)]
if not subrOrient[index]:
for i in range(len(origb)):
origb[i] = -1 - origb[i]
ipds = sr.baseFeature('Ipd',aligned=False, orientation="native")
allipds[index, ccb] = ipds[origb]
# Normalize the IPD values from each subread
allipds = allipds / np.mean(np.percentile(allipds[~np.isnan(allipds)],usepercentiles))
readisb = {refb:np.where([b == refb for b in ccread])[0] for refb in ['A','C','G','T']}
# Take the mean IPD at each position
with warnings.catch_warnings(): # ignoring warnings from taking the mean of columns that are all NaN
warnings.simplefilter("ignore", category=RuntimeWarning)
forwardMean = np.nanmean(allipds[subrOrient == True,:], axis=0)
reverseMean = np.nanmean(allipds[subrOrient == False,:], axis=0)
# get the mean at just Ts
tonlyMean = np.empty(len(ccread), dtype='float32')
tonlyMean.fill(np.nan)
tonlyMean[readisb['T']] = forwardMean[readisb['T']]
tonlyMean[readisb['A']] = reverseMean[readisb['A']]
# Save useful information about this molecule
zmres['zmw'] = zmw
zmres['cclen'] = len(ccread)
zmres['nsubr'] = len(subrs)
zmres['naln'] = len(ccalns)
if usealn is not None:
zmres['chr'] = ccalns[usealn].referenceName
zmres['refStart'] = ccalns[usealn].referenceStart
zmres['refEnd'] = ccalns[usealn].referenceEnd
zmres['alnStart'] = ccalns[usealn].aStart
zmres['alnEnd'] = ccalns[usealn].aEnd
else:
zmres['chr'] = "noAlignment"
zmres['refStart'] = -1
zmres['refEnd'] = -1
zmres['alnStart'] = -1
zmres['alnEnd'] = -1
with warnings.catch_warnings(): # ignoring warnings from taking the mean of all NaN
warnings.simplefilter("ignore", category=RuntimeWarning)
# Something is wrong with the read if these throw a warning, but still no need to print on command line
# The stored value will be NaN
zmres['basemeanA'] = np.nanmean(np.concatenate([forwardMean[readisb['A']], reverseMean[readisb['T']]],axis=None))
zmres['basemeanC'] = np.nanmean(np.concatenate([forwardMean[readisb['C']], reverseMean[readisb['G']]],axis=None))
zmres['basemeanG'] = np.nanmean(np.concatenate([forwardMean[readisb['G']], reverseMean[readisb['C']]],axis=None))
zmres['basemeanT'] = np.nanmean(np.concatenate([forwardMean[readisb['T']], reverseMean[readisb['A']]],axis=None))
zmres['onlyt'] = tonlyMean
# binarize the onlyT ipds
respf, gmmeans, gmweights, aic, convinf = fitGaussian(tonlyMean, initmean=True, zmw=zmw)
zmres['bingmm'] = respf
zmres['gmmlowmean'] = gmmeans[0]
zmres['gmmlowweight'] = gmweights[0]
zmres['gmmhighmean'] = gmmeans[1]
zmres['gmmhighweight'] = gmweights[1]
zmres['aicDiff'] = aic
zmres['convergeInf'] = convinf
outqueue.put(zmres) # put the results in the output queue
zmwqueue.task_done()
zmw = zmwqueue.get()
zmwqueue.task_done()
outqueue.put(None)
