def makeBins5(fN, fOut, typeFilter):
fOut = open(fOut, 'w')
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
id = ls[0]
type, tcc = ls[1:3]
chrom, strand, st, en = bioLibCG.tccSplit(tcc)
#only take seqs that are long enough
if en - st < 100: continue
if not typeFilter in type: continue
tccBins = [] #0 is the first nt from the 3' end
if strand == '1':
for i in range(0, 100):
s, e = st + i, st + i
tccBins.append(bioLibCG.makeTcc(chrom,strand,s,e))
elif strand == '-1':
for i in range(0, 100):
s, e = en - i, en - i
tccBins.append(bioLibCG.makeTcc(chrom,strand,s,e))
pString = [id] + tccBins
fOut.write('\t'.join([str(x) for x in pString]) + '\n')
def collectIntronSeqs(fN, outFN, assembly, amount=100, prime3=True):
myG = gf.GenomeFetch(assembly)
fOut = open(outFN, 'w')
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
if not 'C_INTRON' in ls[1]: continue
chrom, strand, start, end = bioLibCG.tccSplit(ls[2])
if end - start < amount: continue
if strand == '1':
if prime3:
tcc = bioLibCG.makeTcc(chrom, strand, end - amount, end)
seq = myG.getSequence(tcc)
fOut.write('%s\n' % seq[::-1])
else:
if prime3:
tcc = bioLibCG.makeTcc(chrom, strand, start, start + amount)
seq = myG.getSequence(tcc)
fOut.write('%s\n' % seq[::-1])
f.close()
fOut.close()
def collectIntronSeqs(fN, outFN, assembly, amount = 100, prime3 = True):
myG = gf.GenomeFetch(assembly)
fOut = open(outFN, 'w')
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
if not 'C_INTRON' in ls[1]: continue
chrom, strand, start, end = bioLibCG.tccSplit(ls[2])
if end - start < amount: continue
if strand == '1':
if prime3:
tcc = bioLibCG.makeTcc(chrom, strand, end - amount, end)
seq = myG.getSequence(tcc)
fOut.write('%s\n' % seq[::-1])
else:
if prime3:
tcc = bioLibCG.makeTcc(chrom, strand, start, start + amount)
seq = myG.getSequence(tcc)
fOut.write('%s\n' % seq[::-1])
f.close()
fOut.close()
def get3UTRFromTranscriptome(tranFN, outFN, wholeGene = False ):
fOut = open(outFN, 'w')
f = open(tranFN, 'r')
for i, line in enumerate(f):
ls = line.strip().split('\t')
tChrom, tStrand = ls[1], bioLibCG.switchStrandFormat(ls[2])
tStart, tEnd = int(ls[3]), int(ls[4]) - 1
cStart, cEnd = int(ls[5]), int(ls[6]) - 1
if wholeGene:
utrTcc = bioLibCG.makeTcc(tChrom, tStrand, tStart, tEnd)
fOut.write('%s\n' % utrTcc)
continue
#5UTR
if tStrand == '1':
range5 = (tStart, cStart - 1)
else:
range5 = (cEnd + 1, tEnd)
#3UTR
if tStrand == '1':
range3 = (cEnd + 1, tEnd)
else:
range3 = (tStart, cStart - 1)
utrTcc = bioLibCG.makeTcc(tChrom, tStrand, range3[0], range3[1])
fOut.write('%s\n' % utrTcc)
f.close()
fOut.close()
def getTccs(fN):
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom, strand, start, end = ls[0], bioLibCG.switchStrandFormat(ls[5]), ls[1], ls[2]
print bioLibCG.makeTcc(chrom, strand, start, end)
def transcriptSetOverlapTargets(aDir):
geneSetFN = '/home/chrisgre/dataSources/known/Human/geneSets/ensemblAllTranscripts.tsv'
allExons = cgGenes.createGeneSetFromFile(geneSetFN)
#get degradome TCCS
#note that you need to test the AS peaks, this is the location of the targetted transcript
aDC = cgDB.dataController(aDir, cgAlignment.cgAlignment)
id_alignment = aDC.load()
#create list of unique tccs.
uniqTccs = []
for alignment in id_alignment.values():
chrom, strand, start, end = cg.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
tcc = cg.makeTcc(chrom, strand, start, end)
if tcc not in uniqTccs: uniqTccs.append(tcc)
degTccs = [cg.convertToAS(x) for x in uniqTccs]
#find all overlapping exons/transcripts, then all results sequences that overlap exons
overlappingExons = allExons.transcriptOverlaps(degTccs)
overlappingExonTccs = [x.tcc for x in overlappingExons]
overlappingDegTccs = compare.compareTwoTcc(degTccs, overlappingExonTccs, 1)
#update
for obj in id_alignment.values():
chrom, strand, start, end = cg.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
tcc = cg.makeTcc(chrom, strand, start, end)
degTcc = cg.convertToAS(tcc)
if degTcc in overlappingDegTccs:
obj.transcriptOverlap = True
else:
obj.transcriptOverlap = False
aDC.commit(id_alignment)
def transcriptSetOverlapTargets(aDir):
geneSetFN = '/home/chrisgre/dataSources/known/Human/geneSets/ensemblAllTranscripts.tsv'
allExons = cgGenes.createGeneSetFromFile(geneSetFN)
#get degradome TCCS
#note that you need to test the AS peaks, this is the location of the targetted transcript
aDC = cgDB.dataController(aDir, cgAlignment.cgAlignment)
id_alignment = aDC.load()
#create list of unique tccs.
uniqTccs = []
for alignment in id_alignment.values():
chrom, strand, start, end = cg.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
tcc = cg.makeTcc(chrom, strand, start, end)
if tcc not in uniqTccs: uniqTccs.append(tcc)
degTccs = [cg.convertToAS(x) for x in uniqTccs]
#find all overlapping exons/transcripts, then all results sequences that overlap exons
overlappingExons = allExons.transcriptOverlaps(degTccs)
overlappingExonTccs = [x.tcc for x in overlappingExons]
overlappingDegTccs = compare.compareTwoTcc(degTccs, overlappingExonTccs, 1)
#update
for obj in id_alignment.values():
chrom, strand, start, end = cg.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
tcc = cg.makeTcc(chrom, strand, start, end)
degTcc = cg.convertToAS(tcc)
if degTcc in overlappingDegTccs:
obj.transcriptOverlap = True
else:
obj.transcriptOverlap = False
aDC.commit(id_alignment)
def getTccs(fN):
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom, strand, start, end = ls[0], bioLibCG.switchStrandFormat(
ls[5]), ls[1], ls[2]
print bioLibCG.makeTcc(chrom, strand, start, end)
def addOne(fN):
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom, strand, start, end = bioLibCG.tccSplit(ls[0])
start += 1
end += 1
print bioLibCG.makeTcc(chrom,strand,start,end)
def loadEditingSites(fN, nt = 'A'):
'''Using our labs format, load the editing site into a list'''
cBases = {'A':'T', 'T':'A', 'G':'C', 'C':'G'}
f = open(fN, 'r')
f.readline() #header
eList = []
for line in f:
ls = line.strip().split('\t')
e = EditingSite()
e.chromosome = ls[0]
e.coordinate = int(ls[1])
e.gene = ls[3]
e.eRatio = ls[6]
refBase = ls[2]
cBase = cBases[nt]
if refBase == cBase:
e.strand = '-1'
else:
e.strand = '1'
e.tcc = bioLibCG.makeTcc(e.chromosome, e.strand, e.coordinate, e.coordinate)
e.ID = int(ls[13])
eList.append(e)
f.close()
return eList
def markCenterExpression(aFN, wigDir, rn = None, tn = None):
extend = 25
timer = bioLibCG.cgTimer()
timer.start()
aNX = cgNexusFlat.Nexus(aFN, cgAlignmentFlat.cgAlignment)
aNX.load(['centerExpression', 'tTcc', 'tStart', 'sLength', 'tELevel'], [rn, tn])
#load expression of degradome
wigDict = cgWig.loadWigDict(wigDir)
for aID in aNX.centerExpression:
aNX.centerExpression[aID] = [0.0, 0.0, 0.0]
chrom, strand, start, end = bioLibCG.tccSplit(aNX.tTcc[aID])
offset = aNX.tStart[aID]
sLen = aNX.sLength[aID]
if strand == '1':
start = start - extend + offset
end = start + sLen
else:
end = end + extend - offset
start = end - sLen
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgWig.getExpressionProfile(scanRange, wigDict)
#make sure peak is in the small range
peakLevel = aNX.tELevel[aID]
peakInRange = (peakLevel in stretch.values())
expressionSum = sum(stretch.values())
sortedKeys = stretch.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
if expressionSum != 0 and peakInRange:
sumE = 0.0
for key in sortedKeys[8:12]:
sumE += stretch[key]
aNX.centerExpression[aID][0] = sumE/expressionSum
sumE = 0.0
for key in sortedKeys[7:13]:
sumE += stretch[key]
aNX.centerExpression[aID][1] = sumE/expressionSum
sumE = 0.0
for key in sortedKeys[6:14]:
sumE += stretch[key]
aNX.centerExpression[aID][2] = sumE/expressionSum
aNX.save()
def truncate5(fN):
tCount = 0
shortCount = 0
fOut = open(fN + '.trun5', 'w')
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
type, tcc = ls[1], ls[2]
tCount += 1
c, s, st, en = bioLibCG.tccSplit(tcc)
cLen = en - st
if cLen < 50:
shortCount += 1
continue
if s == '1':
st = st + 50
elif s == '-1':
en = en - 50
else:
print 'error'
return 1
ls[2] = bioLibCG.makeTcc(c, s, st, en)
line = '\t'.join(str(x) for x in [ls[0], ls[1], ls[2]]) + '\n'
fOut.write(line)
print shortCount, tCount
def makeTranscriptome(tranFN, outFN):
p = bioLibCG.cgPrint()
p.show = False
gf = GenomeFetch.GenomeFetch('hg19')
fOut = open(outFN, 'w')
f = open(tranFN, 'r')
for line in f:
ls = line.strip().split('\t')
tChrom, tStrand = ls[1], bioLibCG.switchStrandFormat(ls[2])
exonStarts = [int(x) + 1 for x in ls[8][:-1].split(',')]
exonEnds = [int(x) for x in ls[9][:-1].split(',')]
exonPairs = zip(exonStarts, exonEnds)
tID = ls[0]
gID = ls[10]
seqList = []
for eStart, eEnd in exonPairs:
tcc = bioLibCG.makeTcc(tChrom, tStrand, eStart, eEnd)
seqList.append(gf.getSequence(tcc))
mRNA = ''.join(seqList)
#reverse direction if negative strand
if tStrand == '-1':
mRNA = mRNA[::-1]
fOut.write('> %s:%s:%s\n' % (tID, gID, len(mRNA)))
fOut.write(mRNA + '\n\n')
fOut.close()
f.close()
def truncate(fN):
tCount = 0
shortCount = 0
fOut = open(fN + '.trun', 'w')
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
type, tcc = ls[1], ls[2]
tCount += 1
c, s, st, en = bioLibCG.tccSplit(tcc)
cLen = en - st
if cLen < 50:
shortCount += 1
continue
if s == '1':
en = en - 50
elif s == '-1':
st = st + 50
else:
print 'error'
return 1
ls[2] = bioLibCG.makeTcc(c, s, st, en)
line = '\t'.join(str(x) for x in [ls[0], ls[1], ls[2]]) + '\n'
fOut.write(line)
print shortCount, tCount
def makePeakInputQ(cName, minExpression = 2000):
'''Uses shell script and qsub to get peaks quickly'''
mConf = c.getConfig('Main.conf')
conf = c.getConfig(cName)
assembly = conf.conf['assembly']
tccList = []
chromLens = cg.returnChromLengthDict(assembly)
for chrom in chromLens:
if chrom not in cg.acceptableChroms: continue
for strand in ['1','-1']:
print 'Getting Peaks for ', chrom, strand
prevI = 0
for i in rangePoints(1, chromLens[chrom], 30):
if i == 1:
prevI = i
continue
start = prevI
end = i
prevI = i
tcc = cg.makeTcc(chrom, strand, start, end)
log = 'logs/o-' + str(start)
elog = 'logs/e-%s-%s-%s-%s' % (chrom, strand, start, end)
subprocess.Popen(['qsub', '-V', '-cwd', '-e', elog, '-o', log, '-l', 'mem=3G', '-l', 'rt=3600', 'q.sh', tcc, cName, str(minExpression)]).wait()
def makeTranscriptome(tranFN, outFN):
p = bioLibCG.cgPrint()
p.show = False
gf = GenomeFetch.GenomeFetch('hg19')
fOut = open(outFN, 'w')
f = open(tranFN, 'r')
for line in f:
ls = line.strip().split('\t')
tChrom, tStrand = ls[1], bioLibCG.switchStrandFormat(ls[2])
exonStarts = [int(x) + 1 for x in ls[8][:-1].split(',')]
exonEnds = [int(x) for x in ls[9][:-1].split(',')]
exonPairs = zip(exonStarts, exonEnds)
tID = ls[0]
gID = ls[10]
seqList = []
for eStart, eEnd in exonPairs:
tcc = bioLibCG.makeTcc(tChrom, tStrand, eStart, eEnd)
seqList.append(gf.getSequence(tcc))
mRNA = ''.join(seqList)
#reverse direction if negative strand
if tStrand == '-1':
mRNA = mRNA[::-1]
fOut.write('> %s:%s:%s\n' % (tID, gID, len(mRNA)))
fOut.write(mRNA + '\n\n')
fOut.close()
f.close()
def profileAroundPoint(zeroPoint, span, cName, ratio=False, ratioCoord=None):
'''span is +/- that number... if you put in 30, you'll get
-29 to 29 around 0
zeroPoint must be in tcc format with point at start
ratio will return all points around zero point as a ratio of zero point
so zeroPoint = 1.0 and the rest will be fractions of that...'''
chrom, strand, zPoint, end = cg.tccSplit(zeroPoint)
rStart = zPoint - span
rEnd = zPoint + span
rTcc = cg.makeTcc(chrom, strand, rStart, rEnd)
scanDict = svCoord([rTcc], cName)
#reorient so that zero is at zero
#find position of zero
returnDict = {}
if not ratio:
for i in range(1 - span, span):
returnDict[i] = scanDict[zPoint + i]
else:
if ratioCoord:
zeroVal = scanDict[ratioCoord]
else:
zeroVal = scanDict[zPoint]
if zeroVal == 0: zeroVal = 1
for i in range(1 - span, span):
r = float(scanDict[zPoint + i]) / float(zeroVal)
returnDict[i] = r
return returnDict
def makePeakInputQ(cName, minExpression=2000):
'''Uses shell script and qsub to get peaks quickly'''
mConf = c.getConfig('Main.conf')
conf = c.getConfig(cName)
assembly = conf.conf['assembly']
tccList = []
chromLens = cg.returnChromLengthDict(assembly)
for chrom in chromLens:
if chrom not in cg.acceptableChroms: continue
for strand in ['1', '-1']:
print 'Getting Peaks for ', chrom, strand
prevI = 0
for i in rangePoints(1, chromLens[chrom], 30):
if i == 1:
prevI = i
continue
start = prevI
end = i
prevI = i
tcc = cg.makeTcc(chrom, strand, start, end)
log = 'logs/o-' + str(start)
elog = 'logs/e-%s-%s-%s-%s' % (chrom, strand, start, end)
subprocess.Popen([
'qsub', '-V', '-cwd', '-e', elog, '-o', log, '-l',
'mem=3G', '-l', 'rt=3600', 'q.sh', tcc, cName,
str(minExpression)
]).wait()
def profileAroundPoint(zeroPoint, span, cName, ratio = False, ratioCoord = None):
'''span is +/- that number... if you put in 30, you'll get
-29 to 29 around 0
zeroPoint must be in tcc format with point at start
ratio will return all points around zero point as a ratio of zero point
so zeroPoint = 1.0 and the rest will be fractions of that...'''
chrom, strand, zPoint, end = cg.tccSplit(zeroPoint)
rStart = zPoint - span
rEnd = zPoint + span
rTcc = cg.makeTcc(chrom, strand, rStart, rEnd)
scanDict = svCoord([rTcc], cName)
#reorient so that zero is at zero
#find position of zero
returnDict = {}
if not ratio:
for i in range(1-span, span):
returnDict[i] = scanDict[zPoint + i]
else:
if ratioCoord:
zeroVal = scanDict[ratioCoord]
else:
zeroVal = scanDict[zPoint]
if zeroVal == 0: zeroVal = 1
for i in range(1-span, span):
r = float(scanDict[zPoint + i])/float(zeroVal)
returnDict[i] = r
return returnDict
def markCenterExpression(aFN, wigDir, rn=None, tn=None):
extend = 25
timer = bioLibCG.cgTimer()
timer.start()
aNX = cgNexusFlat.Nexus(aFN, cgAlignmentFlat.cgAlignment)
aNX.load(['centerExpression', 'tTcc', 'tStart', 'sLength', 'tELevel'],
[rn, tn])
#load expression of degradome
wigDict = cgWig.loadWigDict(wigDir)
for aID in aNX.centerExpression:
aNX.centerExpression[aID] = [0.0, 0.0, 0.0]
chrom, strand, start, end = bioLibCG.tccSplit(aNX.tTcc[aID])
offset = aNX.tStart[aID]
sLen = aNX.sLength[aID]
if strand == '1':
start = start - extend + offset
end = start + sLen
else:
end = end + extend - offset
start = end - sLen
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgWig.getExpressionProfile(scanRange, wigDict)
#make sure peak is in the small range
peakLevel = aNX.tELevel[aID]
peakInRange = (peakLevel in stretch.values())
expressionSum = sum(stretch.values())
sortedKeys = stretch.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
if expressionSum != 0 and peakInRange:
sumE = 0.0
for key in sortedKeys[8:12]:
sumE += stretch[key]
aNX.centerExpression[aID][0] = sumE / expressionSum
sumE = 0.0
for key in sortedKeys[7:13]:
sumE += stretch[key]
aNX.centerExpression[aID][1] = sumE / expressionSum
sumE = 0.0
for key in sortedKeys[6:14]:
sumE += stretch[key]
aNX.centerExpression[aID][2] = sumE / expressionSum
aNX.save()
def extendPeakTest(tcc, pRange, minVal, maxAvgNoise, minPeakLength, maxPeakLength, cName):
chrom, strand, peakPosition, end = cg.tccSplit(tcc)
cProfile = stepVectorScan.profileAroundPoint(tcc, pRange, cName, ratio = True)
#extend this peak left and right
leftRange = range(1-pRange, 0)
rightRange = range(1, pRange)
leftRange.reverse() #going from the middle outward
#left
startFinal = leftRange[-1]
for i in leftRange:
if cProfile[i] > minVal:
print ' extending stretch'
else:
print ' end of stretch L'
startFinal = i + 1
break
#right
endFinal = rightRange[-1]
for i in rightRange:
if cProfile[i] > minVal:
print ' extending stretch'
else:
print ' end of stretch R'
endFinal = i - 1
break
peakLength = endFinal - startFinal + 1
#avg expression around peak check...
#get total expression before peak
low = startFinal
high = endFinal
noiseExpression = 0
lowRange = range(1 - pRange, low)
highRange = range(high + 1, pRange)
totalLength = len(lowRange) + len(highRange)
print totalLength, pRange, low, high, lowRange, highRange
for i in lowRange:
noiseExpression += cProfile[i]
for i in highRange:
noiseExpression += cProfile[i]
try:
avgNoise = noiseExpression/float(totalLength)
except:
return False
#filter out peaks that look a certain way.
if (minPeakLength < peakLength < maxPeakLength) and (avgNoise < maxAvgNoise):
goodTcc = cg.makeTcc(chrom, strand, peakPosition + startFinal, peakPosition + endFinal)
print '*KEEPER'
return goodTcc
else:
return False
def getOverlappingElements(self, tcc):
'''Given region, Which element (INTRON, EXON, 5UTR, 3UTR)'''
overlappingElements = []
try:
for utrSegment in self.utr5:
utr5Tcc = bioLibCG.makeTcc(self.chromosome, self.strand,
utrSegment[0], utrSegment[1])
if bioLibCG.tccOverlap(utr5Tcc, tcc):
overlappingElements.append([utrSegment, '5UTR'])
except IndexError:
pass
for exon in self.exonList:
exonTcc = bioLibCG.makeTcc(self.chromosome, self.strand, exon[0],
exon[1])
#print '@ ', exonTcc, tcc, 'EXON'
if bioLibCG.tccOverlap(exonTcc, tcc):
overlappingElements.append([exon, 'EXON'])
for intron in self.intronList:
intronTcc = bioLibCG.makeTcc(self.chromosome, self.strand,
intron[0], intron[1])
#print '@ ', intronTcc, tcc, 'INTRON'
if bioLibCG.tccOverlap(intronTcc, tcc):
overlappingElements.append([intron, 'INTRON'])
try:
for utrSegment in self.utr3:
utr3Tcc = bioLibCG.makeTcc(self.chromosome, self.strand,
utrSegment[0], utrSegment[1])
if bioLibCG.tccOverlap(utr3Tcc, tcc):
overlappingElements.append([utrSegment, '3UTR'])
except IndexError:
pass
#!!!Eventually add a way to find if overlapping EXON_UTR as well
if 'EXON' in overlappingElements and 'INTRON' in overlappingElements:
overlappingElements.append('EXON_INTRON')
overlappingElements.remove('EXON')
overlappingElements.remove('INTRON')
return overlappingElements
def getTccFromBowtieLine(line):
chrom = line.strip().split('\t')[2]
strand = line.strip().split('\t')[1]
if strand == '+':
strand = '1'
else:
strand = '-1'
start = int(line.strip().split('\t')[3])
end = start + len(line.strip().split('\t')[4])
return cg.makeTcc(chrom, strand, start, end)
def getTccFromBedLine(line):
chrom = line.strip().split('\t')[0]
strand = line.strip().split('\t')[5]
if strand == '+':
strand = '1'
else:
strand = '-1'
start = int(line.strip().split('\t')[1])
end = int(line.strip().split('\t')[2])
return cg.makeTcc(chrom, strand, start, end)
def getTccFromBowtieLine(line):
chrom = line.strip().split('\t')[2]
strand = line.strip().split('\t')[1]
if strand == '+':
strand = '1'
else:
strand = '-1'
start = int(line.strip().split('\t')[3])
end = start + len(line.strip().split('\t')[4])
return cg.makeTcc(chrom, strand, start, end)
def getTccFromBedLine(line):
chrom = line.strip().split('\t')[0]
strand = line.strip().split('\t')[5]
if strand == '+':
strand = '1'
else:
strand = '-1'
start = int(line.strip().split('\t')[1])
end = int(line.strip().split('\t')[2])
return cg.makeTcc(chrom, strand, start, end)
def toTcc(fN, strand):
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom = ls[0].split(':')[0]
start = ls[0].split(':')[1].split('-')[0]
end = ls[0].split(':')[1].split('-')[1]
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
print newTcc
def toTcc(fN, strand):
f = open(fN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom = ls[0].split(':')[0]
start = ls[0].split(':')[1].split('-')[0]
end = ls[0].split(':')[1].split('-')[1]
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
print newTcc
def getTccFromUCSCLine(line):
'''format may change...'''
chrom = line.strip().split('\t')[1]
strand = line.strip().split('\t')[6]
if strand == '+':
strand = '1'
else:
strand = '-1'
start = int(line.strip().split('\t')[2])
end = int(line.strip().split('\t')[3])
return cg.makeTcc(chrom, strand, start, end)
def getTccFromUCSCLine(line):
'''format may change...'''
chrom = line.strip().split('\t')[1]
strand = line.strip().split('\t')[6]
if strand == '+':
strand = '1'
else:
strand = '-1'
start = int(line.strip().split('\t')[2])
end = int(line.strip().split('\t')[3])
return cg.makeTcc(chrom, strand, start, end)
def peakToSeq(peakFN, extend, outFN, assembly):
#extend is +25 for degradome and -6/-4 for oRNA
extend = int(extend)
gf = GenomeFetch.GenomeFetch(assembly)
outF = open(outFN, 'w')
f = open(peakFN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom, strand, start, end = bioLibCG.tccSplit(ls[0])
start, end = start - extend, end + extend
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
outF.write(gf.getSequence(newTcc) + '\n')
def plotResults(rFN, smallCName, degCName):
f = open(rFN, 'r')
i = 1
for line in f:
chrom, strand, start, end = bioLibCG.tccSplit(line.strip().split('\t')[0])
start = start - 30
end = end + 30
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
cgPlot.plotSmallDeg(newTcc, smallCName, degCName, 'newResults', line.strip(), str(i))
i += 1
def peakToSeq(peakFN, extend, outFN):
#extend is +25 for degradome and -6/-4 for oRNA
extend = int(extend)
gf = GenomeFetch.GenomeFetch('hg19')
outF = open(outFN, 'w')
f = open(peakFN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom, strand, start, end = bioLibCG.tccSplit(ls[0])
start, end = start - extend, end + extend
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
outF.write(gf.getSequence(newTcc) + '\n')
def plotPairs(oDir, aDir, cName):
oDC = cgDB.dataController(oDir, cgOriginRNA.OriginRNA)
id_oRNA = oDC.load()
aDC = cgDB.dataController(aDir, cgAlignment.cgAlignment)
id_alignment = aDC.load()
for oID, oRNA in id_oRNA.items():
if not oRNA.passedFilter:
continue
for aID in oRNA.filteredTargets:
alignment = id_alignment[aID]
chrom, strand, start, end = bioLibCG.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
print sLen
print oRNA.sequence
print oRNA.tcc
print alignment.tTcc
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
print chrom, strand, start, end
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgPeaks.stretch(scanRange, cName)
sortedKeys = stretch.profile.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
xVals = range(1, sLen + 2)
xVals = sortedKeys
yVals = [stretch.profile[x] for x in sortedKeys]
print xVals, len(xVals)
print yVals, len(yVals)
plt.plot(xVals, yVals)
plt.show()
return 0
def getOverlappingElements(self, tcc):
"""Given region, Which element (INTRON, EXON, 5UTR, 3UTR)"""
overlappingElements = []
try:
for utrSegment in self.utr5:
utr5Tcc = bioLibCG.makeTcc(self.chromosome, self.strand, utrSegment[0], utrSegment[1])
if bioLibCG.tccOverlap(utr5Tcc, tcc):
overlappingElements.append([utrSegment, "5UTR"])
except IndexError:
pass
for exon in self.exonList:
exonTcc = bioLibCG.makeTcc(self.chromosome, self.strand, exon[0], exon[1])
# print '@ ', exonTcc, tcc, 'EXON'
if bioLibCG.tccOverlap(exonTcc, tcc):
overlappingElements.append([exon, "EXON"])
for intron in self.intronList:
intronTcc = bioLibCG.makeTcc(self.chromosome, self.strand, intron[0], intron[1])
# print '@ ', intronTcc, tcc, 'INTRON'
if bioLibCG.tccOverlap(intronTcc, tcc):
overlappingElements.append([intron, "INTRON"])
try:
for utrSegment in self.utr3:
utr3Tcc = bioLibCG.makeTcc(self.chromosome, self.strand, utrSegment[0], utrSegment[1])
if bioLibCG.tccOverlap(utr3Tcc, tcc):
overlappingElements.append([utrSegment, "3UTR"])
except IndexError:
pass
#!!!Eventually add a way to find if overlapping EXON_UTR as well
if "EXON" in overlappingElements and "INTRON" in overlappingElements:
overlappingElements.append("EXON_INTRON")
overlappingElements.remove("EXON")
overlappingElements.remove("INTRON")
return overlappingElements
def plotASProfile(tcc, cName, directory = None, min = 0, extra = "0"):
if not directory:
fN = extra + '.' + tcc + '.png'
else:
fN = directory + '/' + extra + '.' + tcc + '.png'
#Get S Profile
tccStretch = cgPeaks.stretch(tcc, cName)
highest = tccStretch.getHighestLevel()
if highest < min:
return 0
sortedX = tccStretch.profile.keys()
sortedX.sort()
sortedY = []
for X in sortedX:
sortedY.append(tccStretch.profile[X])
#Get AS Profile
chr, strand, start, end = tcc.strip().split(':')
if strand == '1':
strand = '-1'
else:
strand = '1'
tcc = cg.makeTcc(chr, strand, start, end)
tccStretchAS = cgPeaks.stretch(tcc, cName)
highest = tccStretchAS.getHighestLevel()
if highest < min:
return 0 #AS can have minimum I guess...
sortedXAS = tccStretchAS.profile.keys()
sortedXAS.sort()
sortedYAS = []
for X in sortedXAS:
sortedYAS.append(tccStretchAS.profile[X])
#Plot them
gDevice = importr('grDevices')
gDevice.png(file=fN, width=1680, height=1050)
r('split.screen(c(2,1))')
r('screen(1)')
r.plot(sortedX, sortedY, xlab = "Coordinates", ylab = "(Syn) Expression Level" )
r.lines(sortedX, sortedY, type = "b")
r('screen(2)')
r.plot(sortedXAS, sortedYAS, xlab = "Coordinates", ylab = "(Anti) Expression Level")
r.lines(sortedXAS, sortedYAS, type = "b")
gDevice.dev_off()
def plotPairs(oDir, aDir, cName):
oDC = cgDB.dataController(oDir, cgOriginRNA.OriginRNA)
id_oRNA = oDC.load()
aDC = cgDB.dataController(aDir, cgAlignment.cgAlignment)
id_alignment = aDC.load()
for oID, oRNA in id_oRNA.items():
if not oRNA.passedFilter:
continue
for aID in oRNA.filteredTargets:
alignment = id_alignment[aID]
chrom, strand, start, end = bioLibCG.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
print sLen
print oRNA.sequence
print oRNA.tcc
print alignment.tTcc
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
print chrom, strand, start, end
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgPeaks.stretch(scanRange, cName)
sortedKeys = stretch.profile.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
xVals = range(1, sLen + 2)
xVals = sortedKeys
yVals = [stretch.profile[x] for x in sortedKeys]
print xVals, len(xVals)
print yVals, len(yVals)
plt.plot(xVals, yVals)
plt.show()
return 0
def updateSequence(oFN, oFF, extend, assembly):
NX = Nexus(oFN, oFF)
NX.load(['sequence', 'tcc'])
gf = GenomeFetch.GenomeFetch(assembly)
while NX.nextID():
chrom, strand, start, end = bioLibCG.tccSplit(NX.tcc)
start, end = start - extend, end + extend
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
NX.sequence = gf.getSequence(newTcc)
NX.save()
def plotResults(rFN, smallCName, degCName):
f = open(rFN, 'r')
i = 1
for line in f:
chrom, strand, start, end = bioLibCG.tccSplit(
line.strip().split('\t')[0])
start = start - 30
end = end + 30
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
cgPlot.plotSmallDeg(newTcc, smallCName, degCName, 'newResults',
line.strip(), str(i))
i += 1
def updateSequence(oFN, oFF, extend, assembly):
NX = Nexus(oFN, oFF)
NX.load(['sequence', 'tcc'])
gf = GenomeFetch.GenomeFetch(assembly)
while NX.nextID():
chrom, strand, start, end = bioLibCG.tccSplit(NX.tcc)
start, end = start - extend, end + extend
newTcc = bioLibCG.makeTcc(chrom, strand, start, end)
NX.sequence = gf.getSequence(newTcc)
NX.save()
def countWithBinsSet(dFN, binDir, type = 'INTRON'):
dNX = cgNexusFlat.Nexus(dFN, cgDegPeak.Peak)
dNX.load(['tcc'])
numBins = 1
c_s_bin_set = {}
for chrom in bioLibCG.humanChromosomes:
for strand in ('1', '-1'):
#initialize data structure
for i in range(0, numBins):
c_s_bin_set.setdefault(chrom, {}).setdefault(strand, {})[i] = set()
f = open(binDir + '/%s.%s.%s.bins' % (type, chrom, strand), 'r')
for line in f:
ls = line.strip().split('\t')
tccs = ls[1:numBins + 1]
for i in range(0,numBins):
ch, st, sta, end = bioLibCG.tccSplit(tccs[i])
for j in range(sta, end + 1):
c_s_bin_set[chrom][strand][i].add(j)
#collect dTtcs in list
dTccs = []
for dID in dNX.tcc:
tcc = dNX.tcc[dID]
c, s, st, en = bioLibCG.tccSplit(tcc)
if s == '1':
s = '-1'
en = st
else:
s = '1'
st = en
dTccs.append(bioLibCG.makeTcc(c,s,st,en))
#make bCounts
binCounts = [0] * numBins
#count for each bin
for i in range(0, numBins):
for dTcc in dTccs:
c, s, st, en = bioLibCG.tccSplit(dTcc)
for j in range(st, en + 1):
if j in c_s_bin_set[c][s][i]:
binCounts[i] += 1
print '%s\t%s' % (i, binCounts[i])
def countWithBinsSet(dFN, binDir, type='INTRON'):
dNX = cgNexusFlat.Nexus(dFN, cgDegPeak.Peak)
dNX.load(['tcc'])
numBins = 1
c_s_bin_set = {}
for chrom in bioLibCG.humanChromosomes:
for strand in ('1', '-1'):
#initialize data structure
for i in range(0, numBins):
c_s_bin_set.setdefault(chrom, {}).setdefault(strand,
{})[i] = set()
f = open(binDir + '/%s.%s.%s.bins' % (type, chrom, strand), 'r')
for line in f:
ls = line.strip().split('\t')
tccs = ls[1:numBins + 1]
for i in range(0, numBins):
ch, st, sta, end = bioLibCG.tccSplit(tccs[i])
for j in range(sta, end + 1):
c_s_bin_set[chrom][strand][i].add(j)
#collect dTtcs in list
dTccs = []
for dID in dNX.tcc:
tcc = dNX.tcc[dID]
c, s, st, en = bioLibCG.tccSplit(tcc)
if s == '1':
s = '-1'
en = st
else:
s = '1'
st = en
dTccs.append(bioLibCG.makeTcc(c, s, st, en))
#make bCounts
binCounts = [0] * numBins
#count for each bin
for i in range(0, numBins):
for dTcc in dTccs:
c, s, st, en = bioLibCG.tccSplit(dTcc)
for j in range(st, en + 1):
if j in c_s_bin_set[c][s][i]:
binCounts[i] += 1
print '%s\t%s' % (i, binCounts[i])
def markCenterExpression(aDir, cName):
aDC = cgDB.dataController(aDir, cgAlignment.cgAlignment)
id_alignment = aDC.load()
for alignment in id_alignment.values():
alignment.centerExpression = [0.0, 0.0, 0.0]
chrom, strand, start, end = bioLibCG.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgPeaks.stretch(scanRange, cName)
expressionSum = stretch.getSumOfLevels()
sortedKeys = stretch.profile.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
if expressionSum != 0:
sum = 0.0
for key in sortedKeys[8:12]:
sum += stretch.profile[key]
alignment.centerExpression[0] = sum/expressionSum
sum = 0.0
for key in sortedKeys[7:13]:
sum += stretch.profile[key]
alignment.centerExpression[1] = sum/expressionSum
sum = 0.0
for key in sortedKeys[6:14]:
sum += stretch.profile[key]
alignment.centerExpression[2] = sum/expressionSum
aDC.commit(id_alignment)
def markCenterExpression(aDir, cName):
aDC = cgDB.dataController(aDir, cgAlignment.cgAlignment)
id_alignment = aDC.load()
for alignment in id_alignment.values():
alignment.centerExpression = [0.0, 0.0, 0.0]
chrom, strand, start, end = bioLibCG.tccSplit(alignment.tTcc)
offset = alignment.tStart
sLen = alignment.sLength
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgPeaks.stretch(scanRange, cName)
expressionSum = stretch.getSumOfLevels()
sortedKeys = stretch.profile.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
if expressionSum != 0:
sum = 0.0
for key in sortedKeys[8:12]:
sum += stretch.profile[key]
alignment.centerExpression[0] = sum / expressionSum
sum = 0.0
for key in sortedKeys[7:13]:
sum += stretch.profile[key]
alignment.centerExpression[1] = sum / expressionSum
sum = 0.0
for key in sortedKeys[6:14]:
sum += stretch.profile[key]
alignment.centerExpression[2] = sum / expressionSum
aDC.commit(id_alignment)
def markCenterExpression(aFN, cName, rn=None, tn=None):
aNX = cgNexusFlat.Nexus(aFN, cgAlignmentFlat.cgAlignment)
aNX.load(['centerExpression', 'tTcc', 'tStart', 'sLength'], [rn, tn])
for aID in aNX.centerExpression:
aNX.centerExpression[aID] = [0.0, 0.0, 0.0]
chrom, strand, start, end = bioLibCG.tccSplit(aNX.tTcc[aID])
offset = aNX.tStart[aID]
sLen = aNX.sLength[aID]
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgPeaks.stretch(scanRange, cName)
expressionSum = stretch.getSumOfLevels()
sortedKeys = stretch.profile.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
if expressionSum != 0:
sum = 0.0
for key in sortedKeys[8:12]:
sum += stretch.profile[key]
aNX.centerExpression[aID][0] = sum / expressionSum
sum = 0.0
for key in sortedKeys[7:13]:
sum += stretch.profile[key]
aNX.centerExpression[aID][1] = sum / expressionSum
sum = 0.0
for key in sortedKeys[6:14]:
sum += stretch.profile[key]
aNX.centerExpression[aID][2] = sum / expressionSum
aNX.save()
def profileTargetsHistoAS(tccList, cName, name='boxplot'):
range = 50
histDict = {} # {coord: []}
histDictAS = {}
for tcc in tccList:
chrom, strand, start, end = cg.tccSplit(tcc)
#Get highest peak (sense)
tccStretch = cgPeaks.stretch(tcc, cName)
tccStretch.createPeaks(span=2)
highestCoord = tccStretch.getHighestPeak()
if highestCoord == None: continue
#AS
tccAS = cg.convertToAS(tcc)
tccStretch = cgPeaks.stretch(tccAS, cName)
tccStretch.createPeaks(span=2)
highestCoordAS = tccStretch.getHighestPeak()
if highestCoordAS == None: continue
#profile around point (Sense)
zPoint = cg.makeTcc(chrom, strand, highestCoord, end)
cProfile = svs.profileAroundPoint(zPoint, range, cName, ratio=True)
for coord in cProfile:
try:
histDict[coord].append(cProfile[coord])
except: #quicker way to initialize
histDict[coord] = [cProfile[coord]]
#profile around point (AS)
zPoint = cg.convertToAS(zPoint)
cProfile = svs.profileAroundPoint(zPoint,
range,
cName,
ratio=True,
ratioCoord=highestCoordAS)
for coord in cProfile:
try:
histDictAS[coord].append(cProfile[coord])
except: #quicker way to initialize
histDictAS[coord] = [cProfile[coord]]
plot.boxPlotHistoAS(histDict, histDictAS, name=name)
def markCenterExpression(aFN, cName, rn = None, tn = None):
aNX = cgNexusFlat.Nexus(aFN, cgAlignmentFlat.cgAlignment)
aNX.load(['centerExpression', 'tTcc', 'tStart', 'sLength'], [rn, tn])
for aID in aNX.centerExpression:
aNX.centerExpression[aID] = [0.0, 0.0, 0.0]
chrom, strand, start, end = bioLibCG.tccSplit(aNX.tTcc[aID])
offset = aNX.tStart[aID]
sLen = aNX.sLength[aID]
if strand == '1':
start = start - 19 + offset
end = start + sLen
else:
end = end + 19 - offset
start = end - sLen
scanRange = bioLibCG.makeTcc(chrom, strand, start, end)
stretch = cgPeaks.stretch(scanRange, cName)
expressionSum = stretch.getSumOfLevels()
sortedKeys = stretch.profile.keys()
sortedKeys.sort()
if strand == '-1':
sortedKeys.reverse()
if expressionSum != 0:
sum = 0.0
for key in sortedKeys[8:12]:
sum += stretch.profile[key]
aNX.centerExpression[aID][0] = sum/expressionSum
sum = 0.0
for key in sortedKeys[7:13]:
sum += stretch.profile[key]
aNX.centerExpression[aID][1] = sum/expressionSum
sum = 0.0
for key in sortedKeys[6:14]:
sum += stretch.profile[key]
aNX.centerExpression[aID][2] = sum/expressionSum
aNX.save()
def getTccFromSamLine(line):
'''SAM has odd formatting at top'''
try:
lineSplit = line.strip().split('\t')
chrom = lineSplit[2]
strand = lineSplit[1]
if strand == '16':
strand = '-1'
else:
strand = '1'
start = int(lineSplit[3])
end = start + len(lineSplit[9])
return cg.makeTcc(chrom, strand, start, end)
except:
print 'Warning: line failed parsing'
print line.strip()
return None
def locateASignals(dataFN, outFN, rn = None, tn = None):
#load data
NX = cgNexusFlat.Nexus(dataFN, ASite)
NX.load(['coord', 'sequence'], [rn, tn])
f = open(outFN, 'w')
for id in NX.ids:
chrom, strand, start, end = bioLibCG.tccSplit(NX.coord[id])
if len(NX.sequence[id]) < 10: continue
print NX.sequence[id], '\n'
checkFrames = bioLibCG.returnFrames(NX.sequence[id], 6)
for i, frame in enumerate(checkFrames):
if frame == 'AATAAA':
#assume 0-based...?
siteStart, siteEnd = start + i, start + i + 5
f.write('%s\n' % bioLibCG.makeTcc(chrom, strand, siteStart, siteEnd))
f.close()
def overlapWithDegradome(dFN, eFN):
eSites = cgEdit.loadEditingSites(eFN)
degTccs = []
f = open(dFN, 'r')
for line in f:
ls = line.strip().split('\t')
chrom, strand, start, end = bioLibCG.tccSplit(ls[1])
start = start - 3
end = end + 3
degTccs.append(bioLibCG.makeTcc(chrom, strand, start, end))
print degTccs[0:5]
eTccs = [eSite.tcc for eSite in eSites]
overlaps = compareData.compareTwoTcc(eTccs, degTccs, 1)
print len(overlaps)
def overlapWithDegradome(dFN, eFN):
eSites = cgEdit.loadEditingSites(eFN)
degTccs = []
f = open(dFN, "r")
for line in f:
ls = line.strip().split("\t")
chrom, strand, start, end = bioLibCG.tccSplit(ls[1])
start = start - 3
end = end + 3
degTccs.append(bioLibCG.makeTcc(chrom, strand, start, end))
print degTccs[0:5]
eTccs = [eSite.tcc for eSite in eSites]
overlaps = compareData.compareTwoTcc(eTccs, degTccs, 1)
print len(overlaps)
def getTccFromSamLine(line):
'''SAM has odd formatting at top'''
try:
lineSplit = line.strip().split('\t')
chrom = lineSplit[2]
strand = lineSplit[1]
if strand == '16':
strand = '-1'
else:
strand = '1'
start = int(lineSplit[3])
end = start + len(lineSplit[9])
return cg.makeTcc(chrom, strand, start, end)
except:
print 'Warning: line failed parsing'
print line.strip()
return None
def profileTargetsHistoAS(tccList, cName, name = 'boxplot'):
range = 50
histDict = {} # {coord: []}
histDictAS = {}
for tcc in tccList:
chrom, strand, start, end = cg.tccSplit(tcc)
#Get highest peak (sense)
tccStretch = cgPeaks.stretch(tcc, cName)
tccStretch.createPeaks(span = 2)
highestCoord = tccStretch.getHighestPeak()
if highestCoord == None: continue
#AS
tccAS = cg.convertToAS(tcc)
tccStretch = cgPeaks.stretch(tccAS, cName)
tccStretch.createPeaks(span = 2)
highestCoordAS = tccStretch.getHighestPeak()
if highestCoordAS == None: continue
#profile around point (Sense)
zPoint = cg.makeTcc(chrom, strand, highestCoord, end)
cProfile = svs.profileAroundPoint(zPoint, range, cName, ratio = True)
for coord in cProfile:
try:
histDict[coord].append(cProfile[coord])
except: #quicker way to initialize
histDict[coord] = [cProfile[coord]]
#profile around point (AS)
zPoint = cg.convertToAS(zPoint)
cProfile = svs.profileAroundPoint(zPoint, range, cName, ratio = True, ratioCoord = highestCoordAS)
for coord in cProfile:
try:
histDictAS[coord].append(cProfile[coord])
except: #quicker way to initialize
histDictAS[coord] = [cProfile[coord]]
plot.boxPlotHistoAS(histDict, histDictAS, name = name)
