def getGlobalString(signal):
signal = smoothSignal(signal, smoothParam)
signalShift, signalScale = computeNorm(signal, 0, len(signal))
signalString = computeString(signal,
0,
len(signal),
signalShift,
signalScale,
levels,
overflow=overflow)
return signalString
def getLevelStr(signal, l):
currSignal = np.array(copy.deepcopy(signal), float)
currSignal = smoothSignal(currSignal, smoothParam)
currSignalShift, currSignalScale = computeNorm(currSignal, 0,
len(currSignal))
currString = computeString(
currSignal,
0,
len(currSignal),
currSignalShift,
currSignalScale,
l,
overflow=overflow,
)
return currString
def getDictFromSequence(signal, refWindowSize, refWindowJump):
dic = {}
for winBeg in range(0, len(signal) - refWindowSize + 1, refWindowJump):
winEnd = winBeg + refWindowSize
currSignal = np.array(copy.deepcopy(signal[winBeg:winEnd]), float)
currSignal = smoothSignal(currSignal, smoothParam)
currSignalShift, currSignalScale = computeNorm(currSignal, 0,
refWindowSize)
currString = computeString(currSignal,
0,
refWindowSize,
currSignalShift,
currSignalScale,
levels,
overflow=overflow)
dic.update(buildDictionary(currString, kmerLength))
return dic
def getDictFromSequence(signal, l=False):
dic = {}
currSignal = np.array(copy.deepcopy(signal), float)
currSignal = smoothSignal(currSignal, smoothParam)
currSignalShift, currSignalScale = computeNorm(currSignal, 0,
len(currSignal))
currString = computeString(
currSignal,
0,
len(currSignal),
currSignalShift,
currSignalScale,
levels,
overflow=overflow,
)
if l == True:
return currString
return buildDictionarySpecial(currString, kmerLength)
fakeSignal = []
fakeIndex = -1
while len(fakeSignal) <= toSignal:
fakeIndex = random.randint(0, len(negReadsPaths) - 1)
fakeSignal = np.array(getSignalFromRead(negReadsPaths[fakeIndex]),
dtype=float)
fakeSignal = fakeSignal[fromSignal:toSignal]
readSignal = readSignal[:workingLen]
refSignal = refSignal[:workingLen]
fakeSignal = fakeSignal[:workingLen]
readSignal = smoothSignal(readSignal, smoothParam)
refSignal = smoothSignal(refSignal, smoothParam)
fakeSignal = smoothSignal(fakeSignal, smoothParam)
readShift, readScale = computeNorm(readSignal, 0, len(readSignal))
refShift, refScale = computeNorm(refSignal, 0, len(refSignal))
fakeShift, fakeScale = computeNorm(fakeSignal, 0, len(fakeSignal))
readStrings, refStrings, fakeStrings = {}, {}, {}
for l in levels:
readStrings[l] = computeString(
readSignal,
0,
len(readSignal),
readShift,
readScale,
l,
overflow=0.30,
)
#plt.plot(y)
#plt.xticks(y_pos, x, color='orange', rotation=45, fontweight='bold', horizontalalignment='right')
#plt.tick_params(labelbottom='off')
refSignal = stringToSignal(refSeq, mod, repeatSignal=repeatSignal)
refSeqHelper = []
for i in refSeq:
refSeqHelper.append(i)
for k in range(repeatSignal - 1):
refSeqHelper.append("_")
y = smoothSignal(y, 5)
refSignal = smoothSignal(refSignal, 5)
ySignalShift, ySignalScale = computeNorm(y, 0, len(y))
#y -= ySignalShift
#y /= ySignalScale
refSignalShift, refSignalScale = computeNorm(refSignal, 0, len(refSignal))
#refSignal -= refSignalShift
#refSignal /= refSignalScale
readString = computeString(y,
0,
len(y),
ySignalShift,
ySignalScale,
levels,
overflow=0.30)
refString = computeString(refSignal,
refSignal = np.array(
stringToSignal(refSeq, mod, repeatSignal=repeatSignal), float)
# fakeSignal = np.array(stringToSignal(fakeSeq, mod, repeatSignal = repeatSignal),
# float)
fakeSignal = []
fakeIndex = -1
while len(fakeSignal) <= signalTo:
fakeIndex = random.randint(0, len(negReads) - 1)
fakeSignal = np.array(getSignalFromRead(negReads[fakeIndex]),
dtype=float)
fakeSignal = fakeSignal[signalFrom:signalTo]
readSignalSm = smoothSignal(readSignal, 5)
refSignalSm = smoothSignal(refSignal, 5)
fakeSignalSm = smoothSignal(fakeSignal, 5)
readShiftSm, readScaleSm = computeNorm(readSignalSm, 0, len(readSignalSm))
refShiftSm, refScaleSm = computeNorm(refSignalSm, 0, len(refSignalSm))
fakeShiftSm, fakeScaleSm = computeNorm(fakeSignalSm, 0, len(fakeSignalSm))
readString2Sm = computeString(
readSignalSm,
0,
len(readSignalSm),
readShiftSm,
readScaleSm,
levels,
overflow=0.25,
)
refString2Sm = computeString(refSignalSm,
0,
len(refSignalSm),
refShiftSm,
refSeq = str(Fasta(refFilePath)[ctg][fromRef:toRef])
else:
refSeq = str(-Fasta(refFilePath)[ctg][fromRef:toRef])
refSignal = np.array(
stringToSignal(refSeq, mod, repeatSignal=repeatSignal), float)
readSignal = np.array(getSignalFromRead(posRead), dtype=float)
readSignal = readSignal[fromSignal:toSignal]
readSignal = readSignal[:workingLen]
refSignal = refSignal[:workingLen]
readSignalSm = smoothSignal(readSignal, smoothParam)
refSignalSm = smoothSignal(refSignal, smoothParam)
readShift, readScale = computeNorm(readSignal, 0, len(readSignal))
readShiftSm, readScaleSm = computeNorm(readSignalSm, 0, len(readSignalSm))
refShift, refScale = computeNorm(refSignal, 0, len(refSignal))
refShiftSm, refScaleSm = computeNorm(refSignalSm, 0, len(refSignalSm))
readStrings, readStringsSm, refStrings, refStringsSm = {}, {}, {}, {}
for l in levels:
readStrings[l] = computeString(
readSignal,
0,
len(readSignal),
readShift,
readScale,
l,
overflow=0.30,
ref = Fasta(refFilePath)
mod = KmerModel.load_from_hdf5(kmerModelFilePath)
for contig in ref:
refSeqPos = str(contig[:])
refSeqNeg = str(contig[:].complement)
refSignalPos = np.array(
stringToSignal(refSeqPos, mod, repeatSignal=repeatSignal), float)
refSignalNeg = np.array(
stringToSignal(refSeqNeg, mod, repeatSignal=repeatSignal), float)
refSignalPos = smoothSignal(refSignalPos, smoothParam)
refSignalNeg = smoothSignal(refSignalNeg, smoothParam)
refSignalPosShift, refSignalPosScale = computeNorm(refSignalPos, 0,
len(refSignalPos))
refSignalNegShift, refSignalNegScale = computeNorm(refSignalNeg, 0,
len(refSignalNeg))
for l in levels:
refStringPos = computeString(
refSignalPos,
0,
len(refSignalPos),
refSignalPosShift,
refSignalPosScale,
l,
overflow=overflow,
)
refStringNeg = computeString(
hits = [
aln for aln in referenceIdx.map(readFastq)
if aln.q_en - aln.q_st > 0.95 *
len(readFastq) and aln.strand == 1 and aln.ctg == workingContig
]
if len(hits) != 1:
continue
hit = hits[0]
location = hit.r_st / len(Fasta(refFilePath)[hit.ctg])
print(f"I am in ctg {hit.ctg} in around {location}")
refSeq = str(Fasta(refFilePath)[hit.ctg][hit.r_st:hit.r_en])
refSignal = stringToSignal(refSeq, mod, repeatSignal)
refSignal = smoothSignal(refSignal, smoothParam)
refShift, refScale = computeNorm(refSignal, 0, len(refSignal))
#refShift, refScale = globalNorms[hit.ctg][0], globalNorms[hit.ctg][1]
refString = computeString(refSignal,
0,
len(refSignal),
refShift,
refScale,
level,
overflow=overflow)
refString = refString[10:len(refString) - 10]
contigX = storeContig[hit.ctg]
startInRef = -1
for e in range(len(contigX) - len(refString) + 1):
w = contigX[e:e + len(refString)]
if w == refString:
