def makeWig(fN, assembly, format=None, name=None):
'''format assumes bowtie
suitible for medium mapped files.
takes longer.'''
#assume bowtie
if not format: format = 'Bowtie'
parserFunction = returnParserFunction(format)
if not name: name = cg.getBaseFileName(fN, naked=True)
lDict = cg.returnChromLengthDict(assembly)
for chrom in lDict:
if not chrom in cg.acceptableChroms: continue
for strand in ['1', '-1']:
f = open(fN, 'r')
#create hitmap of chrom and strand
print chrom, strand, 'hitmap'
hitDict = {}
for line in f:
lChrom, lStrand, start, end = cg.tccSplit(parserFunction(line))
lStrand = str(lStrand)
start = int(start)
end = int(end)
if chrom == lChrom and strand == lStrand:
for i in range(start, end + 1):
try:
hitDict[i] += 1
except KeyError:
hitDict[i] = 1
#write results to wig file
writeWigFromHitDict(hitDict, assembly)
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 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 makeWig(fN, assembly, format = None, name = None):
'''format assumes bowtie
suitible for medium mapped files.
takes longer.'''
#assume bowtie
if not format: format = 'Bowtie'
parserFunction = returnParserFunction(format)
if not name: name = cg.getBaseFileName(fN, naked = True)
lDict = cg.returnChromLengthDict(assembly)
for chrom in lDict:
if not chrom in cg.acceptableChroms: continue
for strand in ['1', '-1']:
f = open(fN, 'r')
#create hitmap of chrom and strand
print chrom, strand, 'hitmap'
hitDict = {}
for line in f:
lChrom, lStrand, start, end = cg.tccSplit(parserFunction(line))
lStrand = str(lStrand)
start = int(start)
end = int(end)
if chrom == lChrom and strand == lStrand:
for i in range(start, end + 1):
try:
hitDict[i] += 1
except KeyError:
hitDict[i] = 1
#write results to wig file
writeWigFromHitDict(hitDict, assembly)
def updateWigLength(fN, assembly):
chromLengths = cg.returnChromLengthDict(assembly)
f = open(fN, 'r')
header = f.readline() #header
lineDict = {} # chr : []
for line in f:
chrom = line.split('\t')[0]
if chrom not in lineDict:
lineDict[chrom] = []
lineDict[chrom].append(line)
else:
lineDict[chrom].append(line)
f.close()
for chrom in lineDict:
print 'extending', chrom
chromLength = chromLengths[chrom]
lastChromLine = lineDict[chrom][-1]
lastValue = int(lastChromLine.split('\t')[2])
print lastValue
if lastValue < chromLength:
print ' ', lastValue, chromLength
lineDict[chrom].append('%s\t%s\t%s\t0.000000\n' % (chrom, lastValue, chromLength))
print ' updated'
f = open(fN, 'w')
f.write(header)
for chrom in lineDict:
f.writelines(lineDict[chrom])
f.close()
def writeWigDictToWig(wigDict, chrom, strand, assembly, name, outDir, blankValue = 0):
'''hopefully the coords are in zero based. And this script will convert it to 0,1'''
#init
coords = sorted(wigDict.keys())
lDict = bioLibCG.returnChromLengthDict('hg19')
chromEnd = lDict[chrom]
outFN = outDir + '/%s.%s.%s.wig' % (name, chrom, strand)
f = open(outFN, 'w')
#write first blank line
f.write('%s\t%s\t%s\t%s\n' % (chrom, 0, coords[0], blankValue))
#p.tell(' 'beginning block', coords[0]
prevCoord = coords[0]
prevValue = wigDict[coords[0]]
blockStart = prevCoord
coords = coords[1:]
for coord in coords:
currValue = wigDict[coord]
if coord - 1 == prevCoord:
#Does the value differ?
if currValue == prevValue:
#keep extending block
prevCoord = coord
prevValue = currValue
else:
#p.tell(' 'writing last equal block', blockStart, prevCoord, prevValue
#finish last block, with NO blank block
f.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, prevCoord + 1, prevValue))
#init next block
prevCoord = coord
prevValue = currValue
blockStart = coord
else: #finish last block, write zero block, start another block
#last
#p.tell(' 'finishing last block', blockStart, prevCoord
f.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, prevCoord + 1, prevValue))
#zero
#p.tell(' 'zero After:', prevCoord, coord
f.write('%s\t%s\t%s\t%s\n' % (chrom, prevCoord + 1, coord, blankValue))
#init next block
prevCoord = coord
blockStart = coord
prevValue = currValue
#write last block and last blank block line
f.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, coord + 1, prevValue))
f.write('%s\t%s\t%s\t%s\n' % (chrom, coord + 1, chromEnd, blankValue))
f.close()
def writeWigFromHitDict(hitDict, assembly, name, directory=None):
mConf = c.getConfig('Main.conf')
if not directory: directory = mConf.conf['wigs']
if not name: name = cg.getBaseFileName(name, naked=True)
lDict = cg.returnChromLengthDict(assembly)
cg.clearDirectory(directory, overwrite=False)
#write results to wig file
for chrom in hitDict:
for strand in hitDict[chrom]:
oF = open(directory + '/%s.%s.%s.wig' % (name, chrom, strand), 'w')
oF.write('track type=bedGraph name=%s.%s.%s\n' %
(name, chrom, strand))
#print ' sorting'
#print hitDict[chrom]
chromEnd = lDict[chrom] #
hitDict[chrom][strand][chromEnd] = 0
keys = hitDict[chrom][strand].keys()
keys.sort()
#print ' writing blocks'
prevVal = 0
prevCoord = 0
blockStart = 0
blockEnd = 1
for key in keys:
val = hitDict[chrom][strand][key]
if prevCoord == key - 1:
if val == prevVal: #should be combined
blockEnd = key + 1
else: #no zero block
#write old block
oF.write('%s\t%s\t%s\t%s\n' %
(chrom, blockStart, blockEnd,
prevVal)) #!make it a float value?
#start new block
blockStart = key
blockEnd = key + 1
else:
#write old block
oF.write('%s\t%s\t%s\t%s\n' %
(chrom, blockStart, blockEnd, prevVal))
#write zero block
oF.write('%s\t%s\t%s\t%s\n' % (chrom, blockEnd, key, 0))
#start new block
blockStart = key
blockEnd = key + 1
prevVal = val
prevCoord = key
oF.close()
def makeWigMem(fN, assembly, format=None, name=None, directory=None):
'''format assumes bowtie
suitible for small mapped files.'''
if not name: name = cg.getBaseFileName(fN, naked=True)
if not format: format = 'Bowtie'
parserFunction = returnParserFunction(format)
lDict = cg.returnChromLengthDict(assembly)
f = open(fN, 'r')
f.readline() #header...file might not have one but its one read...
#create hitmap of chrom and strand
hitDict = {} #format = chr: { strand : { coord : value
for line in f:
try:
lChrom, lStrand, start, end = cg.tccSplit(parserFunction(line))
except AttributeError:
continue
lStrand = str(lStrand)
start = int(start)
end = int(end)
if lChrom in cg.acceptableChroms:
#wig for degradome
if lStrand == '1':
i = start + 20
else:
i = start
try:
hitDict[lChrom][lStrand][i] += 1
except KeyError:
if lChrom not in hitDict:
hitDict[lChrom] = {}
if lStrand not in hitDict[lChrom]:
hitDict[lChrom][lStrand] = {}
hitDict[lChrom][lStrand][i] = 1
'''
for i in range(start, end):
try:
hitDict[lChrom][lStrand][i] += 1
except KeyError:
if lChrom not in hitDict:
hitDict[lChrom] = {}
if lStrand not in hitDict[lChrom]:
hitDict[lChrom][lStrand] = {}
hitDict[lChrom][lStrand][i] = 1
'''
f.close()
#write results to wig file
writeWigFromHitDict(hitDict, assembly, name, directory)
def makeWigMem(fN, assembly, format = None, name = None, directory = None):
'''format assumes bowtie
suitible for small mapped files.'''
if not name: name = cg.getBaseFileName(fN, naked = True)
if not format: format = 'Bowtie'
parserFunction = returnParserFunction(format)
lDict = cg.returnChromLengthDict(assembly)
f = open(fN, 'r')
f.readline() #header...file might not have one but its one read...
#create hitmap of chrom and strand
hitDict = {} #format = chr: { strand : { coord : value
for line in f:
try:
lChrom, lStrand, start, end = cg.tccSplit(parserFunction(line))
except AttributeError:
continue
lStrand = str(lStrand)
start = int(start)
end = int(end)
if lChrom in cg.acceptableChroms:
#wig for degradome
if lStrand == '1':
i = start + 20
else:
i = start
try:
hitDict[lChrom][lStrand][i] += 1
except KeyError:
if lChrom not in hitDict:
hitDict[lChrom] = {}
if lStrand not in hitDict[lChrom]:
hitDict[lChrom][lStrand] = {}
hitDict[lChrom][lStrand][i] = 1
'''
for i in range(start, end):
try:
hitDict[lChrom][lStrand][i] += 1
except KeyError:
if lChrom not in hitDict:
hitDict[lChrom] = {}
if lStrand not in hitDict[lChrom]:
hitDict[lChrom][lStrand] = {}
hitDict[lChrom][lStrand][i] = 1
'''
f.close()
#write results to wig file
writeWigFromHitDict(hitDict, assembly, name, directory)
def writeWigFromHitDict(hitDict, assembly, name, directory = None):
mConf = c.getConfig('Main.conf')
if not directory: directory = mConf.conf['wigs']
if not name: name = cg.getBaseFileName(name, naked = True)
lDict = cg.returnChromLengthDict(assembly)
cg.clearDirectory(directory, overwrite = False)
#write results to wig file
for chrom in hitDict:
for strand in hitDict[chrom]:
oF = open(directory + '/%s.%s.%s.wig' % (name, chrom, strand), 'w')
oF.write('track type=bedGraph name=%s.%s.%s\n' % (name, chrom, strand))
#print ' sorting'
#print hitDict[chrom]
chromEnd = lDict[chrom] #
hitDict[chrom][strand][chromEnd] = 0
keys = hitDict[chrom][strand].keys()
keys.sort()
#print ' writing blocks'
prevVal = 0
prevCoord = 0
blockStart = 0
blockEnd = 1
for key in keys:
val = hitDict[chrom][strand][key]
if prevCoord == key - 1:
if val == prevVal:#should be combined
blockEnd = key + 1
else: #no zero block
#write old block
oF.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, blockEnd, prevVal)) #!make it a float value?
#start new block
blockStart = key
blockEnd = key + 1
else:
#write old block
oF.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, blockEnd, prevVal))
#write zero block
oF.write('%s\t%s\t%s\t%s\n' % (chrom, blockEnd, key, 0))
#start new block
blockStart = key
blockEnd = key + 1
prevVal = val
prevCoord = key
oF.close()
def writeSetToWig(wigSet, chrom, strand, assembly, name, outDir):
print 'if TP in set', (208148750 in wigSet)
#init
coords = sorted(wigSet)
lDict = bioLibCG.returnChromLengthDict('hg19')
chromEnd = lDict[chrom]
outFN = outDir + '/%s.%s.%s.wig' % (name, chrom, strand)
f = open(outFN, 'w')
#write first 0 line
f.write('%s\t%s\t%s\t%s\n' % (chrom, 0, coords[0], 0))
prevCoord = coords[0]
blockStart = prevCoord
coords = coords[1:]
for coord in coords:
if coord - 1 == prevCoord:
#keep extending block
prevCoord = coord
else: #finish last block, write zero block, start another block
#last
f.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, prevCoord + 1, 1))
#zero
f.write('%s\t%s\t%s\t%s\n' % (chrom, prevCoord + 1, coord, 0))
#init next block
prevCoord = coord
blockStart = coord
#write last block and last 0 block line
f.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, coord + 1, 1))
f.write('%s\t%s\t%s\t%s\n' % (chrom, coord + 1, chromEnd, 0))
f.close()
def spacerDistData(tranFN, outFN):
'''chr strand tranStart tranEnd'''
chrom_length = bioLibCG.returnChromLengthDict('hg19')
chrom_strand_iSet = {}
for chrom in chrom_length:
for strand in ('+', '-'):
chrom_strand_iSet.setdefault(chrom,
{}).setdefault(strand, IntervalSet())
print 'making intervals'
f = open(tranFN, 'r')
for line in f:
ls = line.strip().split('\t')
tranStart, tranEnd = int(ls[3]), int(ls[4])
strand = ls[2]
chrom = ls[1]
chrom_strand_iSet[chrom][strand].add(Interval(tranStart, tranEnd))
f.close()
spacerData = []
print 'creating spacer data'
for chrom in chrom_strand_iSet:
for strand in chrom_strand_iSet[chrom]:
iSet = chrom_strand_iSet[chrom][strand]
for i, interv in enumerate(iSet):
if interv == iSet[-1]: break
nextInterv = iSet[i + 1]
seperation = nextInterv.lower_bound - interv.upper_bound
spacerData.append(seperation)
f = open(outFN, 'w')
outLines = [str(x) + '\n' for x in spacerData]
f.writelines(outLines)
f.close()
def spacerDistData(tranFN, outFN):
'''chr strand tranStart tranEnd'''
chrom_length = bioLibCG.returnChromLengthDict('hg19')
chrom_strand_iSet = {}
for chrom in chrom_length:
for strand in ('+', '-'):
chrom_strand_iSet.setdefault(chrom, {}).setdefault(strand, IntervalSet())
print 'making intervals'
f = open(tranFN, 'r')
for line in f:
ls = line.strip().split('\t')
tranStart, tranEnd = int(ls[3]), int(ls[4])
strand = ls[2]
chrom = ls[1]
chrom_strand_iSet[chrom][strand].add(Interval(tranStart, tranEnd))
f.close()
spacerData = []
print 'creating spacer data'
for chrom in chrom_strand_iSet:
for strand in chrom_strand_iSet[chrom]:
iSet = chrom_strand_iSet[chrom][strand]
for i, interv in enumerate(iSet):
if interv == iSet[-1]: break
nextInterv = iSet[i + 1]
seperation = nextInterv.lower_bound - interv.upper_bound
spacerData.append(seperation)
f = open(outFN, 'w')
outLines = [str(x) + '\n' for x in spacerData]
f.writelines(outLines)
f.close()
def makePeakInput(cName, minExpression=2000):
mConf = c.getConfig('Main.conf')
conf = c.getConfig(cName)
assembly = conf.conf['assembly']
tccList = []
chromLens = cg.returnChromLengthDict(assembly)
f = open('peakData.%s' % minExpression, 'w')
for chrom in chromLens:
if chrom not in cg.acceptableChroms: continue
for strand in ['1', '-1']:
print 'Getting Peaks for ', chrom, strand
prevI = 0
endCheck = 0
for i in rangePoints(1, chromLens[chrom], 1000):
if i == 1:
prevI = i
continue
start = prevI
end = i
prevI = i
tcc = cg.makeTcc(chrom, strand, start, end)
#print 'scanning range', tcc
peaks = cgPeaks.stretch(tcc, cName)
peaks.createPeaks(span=3, minVal=minExpression)
for x in peaks.peaks:
if x < endCheck:
continue
#scan a 30 bp range around this point and find the best roof...
pRange = 30
rTcc = cg.makeTcc(chrom, strand, x, x + 1)
#now make profile for roof...
cProfile = stepVectorScan.profileAroundPoint(rTcc,
pRange,
cName,
ratio=True)
#now get highest stretch length and the rNext coord.
minVal = .80
highest = 0
stretch = 0
startCurrent = None
startFinal = None
endFinal = None
for i in range(1 - pRange, pRange):
if cProfile[i] > minVal:
stretch += 1
if startCurrent == None:
startCurrent = i
else:
if stretch > 0:
if stretch > highest: #stretch ended and was higher than previous
highest = stretch
endFinal = i - 1
startFinal = startCurrent
startCurrent = None
else:
startCurrent = None
stretch = 0
#get +/- 4 value...
val = [1.0, 1.0]
if (startFinal) and (endFinal):
low = startFinal - 4
high = endFinal + 4
if low > (1 - pRange) and high < pRange:
val[0] = float(cProfile[startFinal - 4])
val[1] = float(cProfile[endFinal + 4])
else:
continue
else:
continue
endCheck = x + high
#filter out peaks that look a certain way.
if 14 < highest < 26: #rooflength
if val[0] < 0.2 and val[1] < .2: #drop values
goodTcc = cg.makeTcc(chrom, strand, x + low,
x + high)
#print goodTcc
f.write('%s\n' % goodTcc)
f.close()
def makeWigMem(fN, assembly, format = None, name = None, directory = None, degWig = False, switchStrand = True, normalized = False):
'''format assumes bowtie
suitible for small mapped files.
switch strand does not switch the strands, it just makes sure if the data is backwards (HeLa) that it will
put the peak in the right spot'''
print 'degWig Value', degWig
print 'switch strands?', switchStrand
if not name: name = cg.getBaseFileName(fN, naked = True)
if not format: format = 'Bowtie'
parserFunction = returnParserFunction(format)
lDict = cg.returnChromLengthDict(assembly)
f = open(fN, 'r')
f.readline() #header...file might not have one but its one read...
#create hitmap of chrom and strand
hitDict = {} #format = chr: { strand : { coord : value
for line in f:
lChrom, lStrand, start, end = cg.tccSplit(parserFunction(line))
lStrand = str(lStrand)
start = int(start)
end = int(end)
numPlacesMapped = int(line.strip().split('\t')[6])
numPlacesMapped += 1
readCount = 1
if normalized:
readCount = float(readCount)/numPlacesMapped
if lChrom in cg.acceptableChroms:
if degWig:
#wig for degradome NOTE:!!! change lStrand == '1' to '-1' for Bracken!
if switchStrand:
if lStrand == '1':
i = start + (end - start)
else:
i = start + 1
else:
if lStrand == '-1':
i = start + (end - start)
else:
i = start + 1
hitDict.setdefault(lChrom, {}).setdefault(lStrand, {})
hitDict[lChrom][lStrand][i] = hitDict[lChrom][lStrand].get(i, 0) + readCount
else:
#wig for regular
for i in range(start, end):
try:
hitDict[lChrom][lStrand][i] += readCount
except KeyError:
if lChrom not in hitDict:
hitDict[lChrom] = {}
if lStrand not in hitDict[lChrom]:
hitDict[lChrom][lStrand] = {}
hitDict[lChrom][lStrand][i] = readCount
f.close()
#write results to wig file
writeWigFromHitDict(hitDict, assembly, name, directory)
def writeWigDictToWig(wigDict,
chrom,
strand,
assembly,
name,
outDir,
blankValue=0):
'''hopefully the coords are in zero based. And this script will convert it to 0,1'''
#init
coords = sorted(wigDict.keys())
lDict = bioLibCG.returnChromLengthDict('hg19')
chromEnd = lDict[chrom]
outFN = outDir + '/%s.%s.%s.wig' % (name, chrom, strand)
f = open(outFN, 'w')
#write first blank line
f.write('%s\t%s\t%s\t%s\n' % (chrom, 0, coords[0], blankValue))
#p.tell(' 'beginning block', coords[0]
prevCoord = coords[0]
prevValue = wigDict[coords[0]]
blockStart = prevCoord
coords = coords[1:]
for coord in coords:
currValue = wigDict[coord]
if coord - 1 == prevCoord:
#Does the value differ?
if currValue == prevValue:
#keep extending block
prevCoord = coord
prevValue = currValue
else:
#p.tell(' 'writing last equal block', blockStart, prevCoord, prevValue
#finish last block, with NO blank block
f.write('%s\t%s\t%s\t%s\n' %
(chrom, blockStart, prevCoord + 1, prevValue))
#init next block
prevCoord = coord
prevValue = currValue
blockStart = coord
else: #finish last block, write zero block, start another block
#last
#p.tell(' 'finishing last block', blockStart, prevCoord
f.write('%s\t%s\t%s\t%s\n' %
(chrom, blockStart, prevCoord + 1, prevValue))
#zero
#p.tell(' 'zero After:', prevCoord, coord
f.write('%s\t%s\t%s\t%s\n' %
(chrom, prevCoord + 1, coord, blankValue))
#init next block
prevCoord = coord
blockStart = coord
prevValue = currValue
#write last block and last blank block line
f.write('%s\t%s\t%s\t%s\n' % (chrom, blockStart, coord + 1, prevValue))
f.write('%s\t%s\t%s\t%s\n' % (chrom, coord + 1, chromEnd, blankValue))
f.close()
def makeWigMem(fN,
assembly,
format=None,
name=None,
directory=None,
degWig=False,
switchStrand=True,
normalized=False):
'''format assumes bowtie
suitible for small mapped files.
switch strand does not switch the strands, it just makes sure if the data is backwards (HeLa) that it will
put the peak in the right spot'''
print 'degWig Value', degWig
print 'switch strands?', switchStrand
if not name: name = cg.getBaseFileName(fN, naked=True)
if not format: format = 'Bowtie'
parserFunction = returnParserFunction(format)
lDict = cg.returnChromLengthDict(assembly)
f = open(fN, 'r')
f.readline() #header...file might not have one but its one read...
#create hitmap of chrom and strand
hitDict = {} #format = chr: { strand : { coord : value
for line in f:
lChrom, lStrand, start, end = cg.tccSplit(parserFunction(line))
lStrand = str(lStrand)
start = int(start)
end = int(end)
numPlacesMapped = int(line.strip().split('\t')[6])
numPlacesMapped += 1
readCount = 1
if normalized:
readCount = float(readCount) / numPlacesMapped
if lChrom in cg.acceptableChroms:
if degWig:
#wig for degradome NOTE:!!! change lStrand == '1' to '-1' for Bracken!
if switchStrand:
if lStrand == '1':
i = start + (end - start)
else:
i = start + 1
else:
if lStrand == '-1':
i = start + (end - start)
else:
i = start + 1
hitDict.setdefault(lChrom, {}).setdefault(lStrand, {})
hitDict[lChrom][lStrand][i] = hitDict[lChrom][lStrand].get(
i, 0) + readCount
else:
#wig for regular
for i in range(start, end):
try:
hitDict[lChrom][lStrand][i] += readCount
except KeyError:
if lChrom not in hitDict:
hitDict[lChrom] = {}
if lStrand not in hitDict[lChrom]:
hitDict[lChrom][lStrand] = {}
hitDict[lChrom][lStrand][i] = readCount
f.close()
#write results to wig file
writeWigFromHitDict(hitDict, assembly, name, directory)
def makePeakInput(cName, minExpression = 2000):
mConf = c.getConfig('Main.conf')
conf = c.getConfig(cName)
assembly = conf.conf['assembly']
tccList = []
chromLens = cg.returnChromLengthDict(assembly)
f = open('peakData.%s' % minExpression, 'w')
for chrom in chromLens:
if chrom not in cg.acceptableChroms: continue
for strand in ['1', '-1']:
print 'Getting Peaks for ', chrom, strand
prevI = 0
endCheck = 0
for i in rangePoints(1, chromLens[chrom], 1000):
if i == 1:
prevI = i
continue
start = prevI
end = i
prevI = i
tcc = cg.makeTcc(chrom, strand, start, end)
#print 'scanning range', tcc
peaks = cgPeaks.stretch(tcc, cName)
peaks.createPeaks(span = 3, minVal = minExpression)
for x in peaks.peaks:
if x < endCheck:
continue
#scan a 30 bp range around this point and find the best roof...
pRange = 30
rTcc = cg.makeTcc(chrom, strand, x, x + 1)
#now make profile for roof...
cProfile = stepVectorScan.profileAroundPoint(rTcc, pRange, cName, ratio = True)
#now get highest stretch length and the rNext coord.
minVal = .80
highest = 0
stretch = 0
startCurrent = None
startFinal = None
endFinal = None
for i in range(1 - pRange, pRange):
if cProfile[i] > minVal:
stretch += 1
if startCurrent == None:
startCurrent = i
else:
if stretch > 0:
if stretch > highest: #stretch ended and was higher than previous
highest = stretch
endFinal = i - 1
startFinal = startCurrent
startCurrent = None
else:
startCurrent = None
stretch = 0
#get +/- 4 value...
val = [1.0, 1.0]
if (startFinal) and (endFinal):
low = startFinal - 4
high = endFinal + 4
if low > (1 - pRange) and high < pRange:
val[0] = float(cProfile[startFinal - 4])
val[1] = float(cProfile[endFinal + 4])
else:
continue
else:
continue
endCheck = x + high
#filter out peaks that look a certain way.
if 14 < highest < 26: #rooflength
if val[0] < 0.2 and val[1] < .2: #drop values
goodTcc = cg.makeTcc(chrom, strand, x + low, x + high)
#print goodTcc
f.write('%s\n' % goodTcc)
f.close()
