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 sortResults(cName = None):
#INIT
conf = cgConfig.getConfig(cName)
pFileName = conf.conf['results']
minDensity = 4
minSmallHits = 1
pFile = open(pFileName, 'r')
fileLines = [] #This will hold the lists to be sorted...
for line in pFile:
fileLines.append(line.strip().split('\t'))
pFile.close()
#highest prediction density
densityDict = {} #CID: highest density --> used to sort out clusters without proper density
for line in fileLines:
CID = line[7]
pDensity = int(line[5])
if CID in densityDict:
if pDensity > densityDict[CID]:
densityDict[CID] = pDensity
else:
densityDict[CID] = pDensity
#take out clusters that didn't make the cut
CIDpassed = []
keptLines = []
for line in fileLines:
CID = line[7]
#smallClusterHits = int(line[10]) Not using this metric anymore...
if (densityDict[CID] >= minDensity): #Density
if line[8] == '0': #doesn't overlap with anything known (the cluster doesn't...that is)
keptLines.append(line)
if CID not in CIDpassed:
CIDpassed.append(CID)
#remake keptLines with integer in field ten
#at this point just sort by cluster density...
sID = 5
for line in keptLines:
line[sID] = int(line[sID])
sortedData = sorted(keptLines, key=itemgetter(sID), reverse = True) #sort by small RNA hits
for line in sortedData:
line[sID] = str(line[sID])
#output
sortedFile = open(conf.conf['results'] + '.sorted', 'w')
for line in sortedData:
sortedFile.write('\t'.join(line) + '\n')
sortedFile.close()
#Now output stats
statFile = open('statFile.data', 'w')
statFile.write('Total Clusters: %s\n' % len(densityDict))
statFile.write('Passed: %s\n' % len(CIDpassed))
def updateOverlaps(cName=None):
#init
conf = cgConfig.getConfig(cName)
pFileName = conf.conf['results']
overlapsFileName = conf.conf['matureOverlaps']
#put overlapping sequences in list:
overlaps = []
overlapFile = open(overlapsFileName, 'r')
for tcc in overlapFile:
overlaps.append(tcc.strip())
overlapFile.close()
#check each line of pred file for overlap, add 1 for overlap and 0 for non
predFile = open(pFileName, 'r')
newFileLines = []
for line in predFile:
mTcc = line.strip().split('\t')[1]
if mTcc in overlaps:
newLine = line.strip().split('\t')
newLine[6] = str(1)
newLine = '\t'.join(newLine) + '\n'
newFileLines.append(newLine)
else:
newLine = line.strip().split('\t')
newLine[6] = str(0)
newLine = '\t'.join(newLine) + '\n'
newFileLines.append(newLine)
predFile.close()
#write new File
newFile = open(pFileName, 'w')
for line in newFileLines:
newFile.write(line)
def returnChromLengthDict(assembly):
mConf = c.getConfig('Main.conf')
lenFileName = mConf.conf['chromosomeLengths'] + '/' + assembly
f = open(lenFileName, 'r')
lenDict = {}
for line in f:
lenDict[line.split('\t')[0]] = int(line.split('\t')[1])
return lenDict
def returnChromLengthDict(assembly):
mConf = c.getConfig('Main.conf')
lenFileName = mConf.conf['chromosomeLengths'] + '/' + assembly
f = open(lenFileName, 'r')
lenDict = {}
for line in f:
lenDict[line.split('\t')[0]] = int(line.split('\t')[1])
return lenDict
def scanVectorsOrganism(tccList, config=None):
'''Given tcc list --> scan Organism wig files and coord:value...
'''
config = c.getConfig(config)
coordDict = {} # tcc: [list values]
for tcc in tccList:
chrom, strand, tccStart, tccEnd = cg.tccSplit(tcc)
#print 'Checking Tcc'
org = config.conf['organism']
mConf = c.getConfig('Main.conf')
wigDir = mConf.conf['wig%s' % org]
fN = wigDir + '/Merge.%s.%s.wig.%s.wig' % (org.lower(), strand, chrom)
#print 'Checking Index'
#goto correct line in index
fIndex = cgIndex.lineIndex(
fN, header=True
) #!!!there actually is a header...have to deal with this...
fIndex.passCheckFunction(cgIndex.wigCheckFunction)
fIndex.binarySearch(
tcc) #places file pointer at beginning of tcc as beginning
stop = False
for line in fIndex.file:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1])
lEnd = int(cg.ss(line)[2])
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd):
coordDict[i] = lValue
if stop: break
return coordDict
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 updateClusterInfo(cName = None):
#init
conf = cgConfig.getConfig(cName)
pFileName = conf.conf['results']
sortedClustersFileName = conf.conf['sortedClusters']
#add cluster IDs to tccs
sortedFile = open(sortedClustersFileName, 'r')
idDict = {} # format: tcc : clusterID
i = 0
for line in sortedFile:
clusterID = str(i)
i = i + 1
#put cluster into list
cluster = line.strip().split(',')
cluster.remove('') #pesky last comma
for tcc in cluster:
idDict[tcc] = clusterID
sortedFile.close()
#get tccs that are overlapping with known sequences
overDict = {}
predFile = open(pFileName, 'r')
for line in predFile:
if line.strip().split('\t')[6] == '1':
clusterID = idDict[line.strip().split('\t')[1]]
overDict[clusterID] = 1
predFile.close()
#now remake file
newLines = []
predFile = open(pFileName, 'r')
for line in predFile:
clusterID = idDict[line.strip().split('\t')[1]]
if clusterID in overDict:
newLine = line.strip().split('\t')
newLine[7] = str(clusterID)
newLine[8] = str(1)
newLine = '\t'.join(newLine) + '\n'
newLines.append(newLine)
else:
newLine = line.strip().split('\t')
newLine[7] = str(clusterID)
newLine[8] = str(0)
newLine = '\t'.join(newLine) + '\n'
newLines.append(newLine)
predFile.close()
#write new File
newFile = open(pFileName, 'w')
for line in newLines:
newFile.write(line)
def scanVectorsOrganism(tccList, config = None):
'''Given tcc list --> scan Organism wig files and coord:value...
'''
config = c.getConfig(config)
coordDict = {} # tcc: [list values]
for tcc in tccList:
chrom, strand, tccStart, tccEnd = cg.tccSplit(tcc)
#print 'Checking Tcc'
org = config.conf['organism']
mConf = c.getConfig('Main.conf')
wigDir = mConf.conf['wig%s' % org]
fN = wigDir + '/Merge.%s.%s.wig.%s.wig' % (org.lower(),strand,chrom)
#print 'Checking Index'
#goto correct line in index
fIndex = cgIndex.lineIndex(fN, header = True) #!!!there actually is a header...have to deal with this...
fIndex.passCheckFunction(cgIndex.wigCheckFunction)
fIndex.binarySearch(tcc) #places file pointer at beginning of tcc as beginning
stop = False
for line in fIndex.file:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1])
lEnd = int(cg.ss(line)[2])
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd):
coordDict[i] = lValue
if stop: break
return coordDict
def svCoord(tccList, config = None):
'''Given tcc list --> scan Organism wig files and coord:value...
'''
#init
config = c.getConfig(config)
org = config.conf['organism']
wigDir = config.conf['wigSetDir']
wigSetName = config.conf['wigSetName']
splitIntoChroms = config.conf['wigChromSplit']
if splitIntoChroms == 'True':
splitIntoChroms = True
else:
splitIntoChroms = False
coordDict = {} # tcc: [list values]
for tcc in tccList:
chrom, strand, tccStart, tccEnd = cg.tccSplit(tcc)
if splitIntoChroms:
fN = wigDir + '/%s.%s.%s.wig' % (wigSetName, chrom, strand)
else:
fN = wigDir + '/Merge.%s.%s.wig' % (org.lower(), strand)
fIndex = cgIndex.lineIndex(fN, header = True)
fIndex.passCheckFunction(cgIndex.wigCheckFunction)
fIndex.binarySearch(tcc) #places file pointer at beginning of tcc as beginning
stop = False
for line in fIndex.file:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1]) + 1
#print 'lBeg', lBeg
lEnd = int(cg.ss(line)[2])
#print 'lEnd', lEnd
#print '--'
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd + 1):
coordDict[i] = lValue
if stop: break
fIndex.close() #close the file and the index after use...
return coordDict
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 filterOut(cName = None): #Init conf = cgConfig.getConfig(cName) predictionList = compare.tccFileToList(conf.conf['resultsRaw'], 1) #predictionList = compare.tccFileToList(conf.conf['resultsRaw'], 0) overlapped = compare.filterOutTccs(predictionList, conf.conf['knownDirectory'], True) #True gives me the filtered out ones instead of the list without filtered out matureOverlaps = open(conf.conf['matureOverlaps'], 'w') for tcc in overlapped: matureOverlaps.write(tcc + '\n')
def filterOut(cName=None):
# Init
conf = cgConfig.getConfig(cName)
predictionList = compare.tccFileToList(conf.conf["resultsRaw"], 1)
# predictionList = compare.tccFileToList(conf.conf['resultsRaw'], 0)
overlapped = compare.filterOutTccs(
predictionList, conf.conf["knownDirectory"], True
) # True gives me the filtered out ones instead of the list without filtered out
matureOverlaps = open(conf.conf["matureOverlaps"], "w")
for tcc in overlapped:
matureOverlaps.write(tcc + "\n")
def addPeriods(cName = None):
#init
conf = cgConfig.getConfig(cName) #gets the current configuration instructions
pFileName = conf.conf['resultsRaw']
nFileName = conf.conf['results']
pFile = open(pFileName, 'r')
nFile = open(nFileName, 'w')
newLines = []
for line in pFile:
newLine = '\t'.join(line.strip().split('\t')[0:5]) + '\t.\t.\t.\t.\t.\t.\n'
nFile.write(newLine)
def splitExonsIntrons(cName = None):
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
#init
organism = conf.conf['organism']
minOverlap = 50
cHairs = getHairpins.getHairpins() #CID: HAIRPIN
exonList = compare.tccFileToList('%sExons.tcc' % organism, 0)
hairpins = []
for CID in cHairs:
hairpins.append(cHairs[CID])
print 'checking overlaps'
#check which hairpins overlap exons and by how much
exonOverlapped = compare.compareTwoTcc(hairpins, exonList, 1, amount = True)
print ' ', len(exonOverlapped)
print 'removing partial introns'
#remove the ones that didn't overlap more than X:
remList = []
for tcc, oAmount in exonOverlapped:
if oAmount < minOverlap:
remList.append([tcc, oAmount])
for item in remList:
exonOverlapped.remove(item)
print ' ', len(exonOverlapped), 'out of', len(cHairs.keys())
#get CIDs of exons
exonCIDs = []
for tcc, oAmount in exonOverlapped:
for CID in cHairs:
if cHairs[CID] == tcc:
exonCIDs.append(str(CID))
#Open sorted predictions and write lines with CIDs to respective files
predFile = open(conf.conf['resultsSorted'], 'r')
exonFile = open(conf.conf['resultsSorted'] + '.exons', 'w')
intronFile = open(conf.conf['resultsSorted'] + '.introns', 'w')
for line in predFile:
if line.split('\t')[7] in exonCIDs:
exonFile.write(line)
else:
intronFile.write(line)
predFile.close()
exonFile.close()
intronFile.close()
def updateDensity(cName=None):
#Create hitmap for blocks, cValdict for block
conf = cgConfig.getConfig(cName)
blockFileName = conf.conf[
'hitsPerFrame'] # created in defineCluster script folder
blockFile = open(blockFileName, 'r')
blocksList = []
cValBlockDict = {}
for line in blockFile:
blocksList.append(line.strip().split('\t')[0])
cValBlockDict[line.strip().split('\t')[0]] = int(
line.strip().split('\t')[1])
blockFile.close()
blockHitmap = bioLibCG.createHitMap(blocksList)
#Now append the cVal for each predicted line:
predictedFileName = conf.conf['results']
predictedFile = open(predictedFileName, 'r')
newFileList = []
counter = 0
for line in predictedFile:
counter = counter + 1
#print counter
cVal = 0
#what blocks does this prediction overlap?
tccPrediction = line.strip().split('\t')[1] #This should be mature?
coordsPrediction = bioLibCG.stripTripleColon(tccPrediction)
for i in range(int(coordsPrediction['start']),
int(coordsPrediction['end'])):
if i in blockHitmap:
for block in blockHitmap[i]:
if bioLibCG.tccOverlap(tccPrediction, block):
if cValBlockDict[block] > cVal:
cVal = cValBlockDict[block]
newLine = line.strip().split('\t')
newLine[5] = str(cVal)
newLine = '\t'.join(newLine) + '\n'
newFileList.append(newLine)
predictedFile.close()
newFileName = conf.conf['results']
newFile = open(newFileName, 'w')
for line in newFileList:
newFile.write(line)
newFile.close()
def splitExonsIntrons(cName=None):
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
#init
organism = conf.conf['organism']
minOverlap = 50
cHairs = getHairpins.getHairpins() #CID: HAIRPIN
exonList = compare.tccFileToList('%sExons.tcc' % organism, 0)
hairpins = []
for CID in cHairs:
hairpins.append(cHairs[CID])
print 'checking overlaps'
#check which hairpins overlap exons and by how much
exonOverlapped = compare.compareTwoTcc(hairpins, exonList, 1, amount=True)
print ' ', len(exonOverlapped)
print 'removing partial introns'
#remove the ones that didn't overlap more than X:
remList = []
for tcc, oAmount in exonOverlapped:
if oAmount < minOverlap:
remList.append([tcc, oAmount])
for item in remList:
exonOverlapped.remove(item)
print ' ', len(exonOverlapped), 'out of', len(cHairs.keys())
#get CIDs of exons
exonCIDs = []
for tcc, oAmount in exonOverlapped:
for CID in cHairs:
if cHairs[CID] == tcc:
exonCIDs.append(str(CID))
#Open sorted predictions and write lines with CIDs to respective files
predFile = open(conf.conf['resultsSorted'], 'r')
exonFile = open(conf.conf['resultsSorted'] + '.exons', 'w')
intronFile = open(conf.conf['resultsSorted'] + '.introns', 'w')
for line in predFile:
if line.split('\t')[7] in exonCIDs:
exonFile.write(line)
else:
intronFile.write(line)
predFile.close()
exonFile.close()
intronFile.close()
def finalSort(pType, cName = None):
#INIT
conf = cgConfig.getConfig(cName)
if pType == "E":
pFileName = conf.conf['resultsExons']
else:
pFileName = conf.conf['resultsIntrons']
minDensity = 4
maxPVal = float(.05)
pFile = open(pFileName, 'r')
fileLines = [] #This will hold the lists to be sorted...
for line in pFile:
fileLines.append(line.strip().split('\t'))
pFile.close()
keptLines = []
for line in fileLines:
CID = line[7]
cDensity = int(line[12])
pVal = float(line[15])
#print pVal, cDensity
if cDensity > 5 and pVal < maxPVal:
#print ' kept'
keptLines.append(line)
print len(keptLines)
#remake keptLines with float(pVal)
i = 12
for line in keptLines:
line[i] = float(line[i])
sortedData = sorted(keptLines, key=itemgetter(i), reverse = True) #sort by i
for line in sortedData:
line[i] = str(line[i])
#print len(sortedData)
#output
sortedFile = open(pFileName + '.sorted', 'w')
for line in sortedData:
sortedFile.write('\t'.join(line) + '\n')
sortedFile.close()
def addPeriods(cName=None):
#init
conf = cgConfig.getConfig(
cName) #gets the current configuration instructions
pFileName = conf.conf['resultsRaw']
nFileName = conf.conf['results']
pFile = open(pFileName, 'r')
nFile = open(nFileName, 'w')
newLines = []
for line in pFile:
newLine = '\t'.join(
line.strip().split('\t')[0:5]) + '\t.\t.\t.\t.\t.\t.\n'
nFile.write(newLine)
def finalSort(pType, cName=None):
#INIT
conf = cgConfig.getConfig(cName)
if pType == "E":
pFileName = conf.conf['resultsExons']
else:
pFileName = conf.conf['resultsIntrons']
minDensity = 4
maxPVal = float(.05)
pFile = open(pFileName, 'r')
fileLines = [] #This will hold the lists to be sorted...
for line in pFile:
fileLines.append(line.strip().split('\t'))
pFile.close()
keptLines = []
for line in fileLines:
CID = line[7]
cDensity = int(line[12])
pVal = float(line[15])
#print pVal, cDensity
if cDensity > 5 and pVal < maxPVal:
#print ' kept'
keptLines.append(line)
print len(keptLines)
#remake keptLines with float(pVal)
i = 12
for line in keptLines:
line[i] = float(line[i])
sortedData = sorted(keptLines, key=itemgetter(i), reverse=True) #sort by i
for line in sortedData:
line[i] = str(line[i])
#print len(sortedData)
#output
sortedFile = open(pFileName + '.sorted', 'w')
for line in sortedData:
sortedFile.write('\t'.join(line) + '\n')
sortedFile.close()
def updateDensity(cName=None):
# Create hitmap for blocks, cValdict for block
conf = cgConfig.getConfig(cName)
blockFileName = conf.conf["hitsPerFrame"] # created in defineCluster script folder
blockFile = open(blockFileName, "r")
blocksList = []
cValBlockDict = {}
for line in blockFile:
blocksList.append(line.strip().split("\t")[0])
cValBlockDict[line.strip().split("\t")[0]] = int(line.strip().split("\t")[1])
blockFile.close()
blockHitmap = bioLibCG.createHitMap(blocksList)
# Now append the cVal for each predicted line:
predictedFileName = conf.conf["results"]
predictedFile = open(predictedFileName, "r")
newFileList = []
counter = 0
for line in predictedFile:
counter = counter + 1
# print counter
cVal = 0
# what blocks does this prediction overlap?
tccPrediction = line.strip().split("\t")[1] # This should be mature?
coordsPrediction = bioLibCG.stripTripleColon(tccPrediction)
for i in range(int(coordsPrediction["start"]), int(coordsPrediction["end"])):
if i in blockHitmap:
for block in blockHitmap[i]:
if bioLibCG.tccOverlap(tccPrediction, block):
if cValBlockDict[block] > cVal:
cVal = cValBlockDict[block]
newLine = line.strip().split("\t")
newLine[5] = str(cVal)
newLine = "\t".join(newLine) + "\n"
newFileList.append(newLine)
predictedFile.close()
newFileName = conf.conf["results"]
newFile = open(newFileName, "w")
for line in newFileList:
newFile.write(line)
newFile.close()
def mergeInputs(cName, eLevel):
conf = c.getConfig(cName)
assembly = conf.conf['assembly']
ending = '%s.%s' % (eLevel, assembly)
print 'merging all files with ending', ending
newLines = []
for fN in cg.recurseDir('out', end = ending):
print os.getcwd(), fN
fN = os.getcwd() + '/' + fN
f = open(fN, 'r')
newLines.extend(f.readlines())
f.close()
f = open('peakData.%s.%s' % (eLevel, assembly), 'w')
f.writelines(newLines)
f.close()
def mergeInputs(cName, eLevel):
conf = c.getConfig(cName)
assembly = conf.conf['assembly']
ending = '%s.%s' % (eLevel, assembly)
print 'merging all files with ending', ending
newLines = []
for fN in cg.recurseDir('out', end=ending):
print os.getcwd(), fN
fN = os.getcwd() + '/' + fN
f = open(fN, 'r')
newLines.extend(f.readlines())
f.close()
f = open('peakData.%s.%s' % (eLevel, assembly), 'w')
f.writelines(newLines)
f.close()
def parallelMakePeaks(tcc, cName, minExpression):
conf = c.getConfig(cName)
f = open('out/peakData.%s.%s.%s' % (tcc, minExpression, conf.conf['assembly']), 'w')
chrom, strand, start, end = cg.tccSplit(tcc)
peaks = cgPeaks.stretch(tcc, cName)
print 'getting peaks'
peaks.createPeaks(span = 1, minVal = int(minExpression))
for x in peaks.peaks:
print x
newTcc = cg.makeTcc(chrom, strand, x, x + 1)
testedPeak = extendPeakTest(newTcc, 20, .2, .05, 0, 6, cName)
#testedPeak = roofPeakTest(newTcc, 30, .85, .9, .2, 6, 17, 24, cName)
if testedPeak:
f.write('%s\n' % testedPeak)
f.close()
def parallelMakePeaks(tcc, cName, minExpression):
conf = c.getConfig(cName)
f = open('out/peakData.%s.%s.%s' % (tcc, minExpression, conf.conf['assembly']), 'w')
chrom, strand, start, end = cg.tccSplit(tcc)
peaks = cgPeaks.stretch(tcc, cName)
print 'getting peaks'
peaks.createPeaks(span = 1, minVal = int(minExpression))
for x in peaks.peaks:
print ""
print chrom, strand, x,
newTcc = cg.makeTcc(chrom, strand, x, x + 1)
testedPeak = extendPeakTest(newTcc, 20, .2, .05, 0, 6, cName)
#testedPeak = roofPeakTest(newTcc, 30, .85, .9, .2, 8, 16, 25, cName)
if testedPeak:
f.write('%s\n' % testedPeak)
f.close()
def updateOverlaps(cName = None):
#init
conf = cgConfig.getConfig(cName)
pFileName = conf.conf['results']
overlapsFileName = conf.conf['matureOverlaps']
#put overlapping sequences in list:
overlaps = []
overlapFile = open(overlapsFileName, 'r')
for tcc in overlapFile:
overlaps.append(tcc.strip())
overlapFile.close()
#check each line of pred file for overlap, add 1 for overlap and 0 for non
predFile = open(pFileName, 'r')
newFileLines = []
for line in predFile:
mTcc = line.strip().split('\t')[1]
if mTcc in overlaps:
newLine = line.strip().split('\t')
newLine[6] = str(1)
newLine = '\t'.join(newLine) + '\n'
newFileLines.append(newLine)
else:
newLine = line.strip().split('\t')
newLine[6] = str(0)
newLine = '\t'.join(newLine) + '\n'
newFileLines.append(newLine)
predFile.close()
#write new File
newFile = open(pFileName, 'w')
for line in newFileLines:
newFile.write(line)
def intronNoisy(cName=None):
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
#init
cHairs = getHairpins.getHairpins(
conf.conf['resultsIntrons']) #CID: HAIRPIN
organism = conf.conf['organism']
exonList = compare.tccFileToList('%sExons.tcc' % organism, 0)
slide = 1000
#make prediction overlap hitmap
predMap = {}
predList = []
for CID in cHairs:
hPin = cHairs[CID]
predList.append(hPin)
#collapse Overlaps
print ' collapsing predictions'
predList = compare.collapseOverlaps(predList)
print ' collapsing exons'
exonList = compare.collapseOverlaps(exonList)
#collect levels for each hairpin region
cidLevels = {}
for CID in cHairs:
print CID
hPin = cHairs[CID]
chrom = ss(hPin, ':')[0]
strand = ss(hPin, ':')[1]
start = int(ss(hPin, ':')[2])
end = int(ss(hPin, ':')[3])
scanStart = start - slide
scanEnd = end + slide
scanRange = []
scanRange.append('%s:%s:%s:%s' % (chrom, strand, scanStart, start))
scanRange.append('%s:%s:%s:%s' % (chrom, strand, end, scanEnd))
print scanRange
scanRange = compare.subtractTwoTccLists(scanRange, predList)
scanRange = compare.subtractTwoTccLists(scanRange, exonList)
levels = []
print ' Retrieving Expression levels:', cg.getTccListTotalLength(
scanRange)
levels = []
hPinLevels = stepVectorScan.scanVectorsHist(scanRange, cName)
for hPin in hPinLevels:
levels.extend(hPinLevels[hPin])
cidLevels[CID] = levels
#output levels to file
#find longest
longest = 0
for CID in cidLevels:
length = len(cidLevels[CID])
if length > longest:
longest = length
sortedKeys = cidLevels.keys()
sortedKeys.sort()
newLines = []
for j in range(0, longest): #how many lines are there
newLine = []
for CID in sortedKeys:
if len(cidLevels[CID]) > j: # add it
newLine.append(str(cidLevels[CID][j]))
else:
newLine.append('NA')
newLines.append('\t'.join(newLine) + '\n')
outFileN = conf.conf['intronNoiseData']
outFile = open(outFileN, 'w')
outFile.write('\t'.join(sortedKeys) + '\n')
outFile.writelines(newLines)
outFile.close()
def parallelMakePeaks(tcc, cName, minExpression):
conf = c.getConfig(cName)
f = open('out/peakData.%s.%s.%s' % (tcc, minExpression, conf.conf['assembly']), 'w')
print 'scanning range', tcc
chrom, strand, start, end = cg.tccSplit(tcc)
peaks = cgPeaks.stretch(tcc, cName)
#print 'getting peaks'
peaks.createPeaks(span = 1, minVal = int(minExpression))
print 'len peaks', len(peaks.peaks)
endCheck = 0
for x in peaks.peaks:
print x, endCheck
'''
if x < endCheck:
print 'endChecked'
continue
'''
#scan a 30 bp range around this point and find the best roof...
pRange = 40
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 = .70
highest = 0
stretch = 0
startCurrent = None
startFinal = None
endFinal = None
for i in range(1 - pRange, pRange):
print ' ', x + i, cProfile[i]
if cProfile[i] > minVal:
print ' extending stretch'
stretch += 1
if startCurrent == None:
startCurrent = i
else:
if stretch > 0:
print 'end of stretch'
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 +/- extend value...
val = [1.0, 1.0]
extend = 1
if (startFinal) and (endFinal):
low = startFinal - extend
high = endFinal + extend
if low > (1 - pRange) and high < pRange:
val[0] = float(cProfile[startFinal - extend])
val[1] = float(cProfile[endFinal + extend])
else:
print 'out of range'
continue
else:
print 'no start and end of peak'
continue
print low, high, x, endFinal
endCheck = x + endFinal
#avg expression around peak check...
#get total expression before peak
noiseExpression = 0
lowRange = range(1 - pRange, low)
highRange = range(high + 1, pRange)
totalLength = len(lowRange) + len(highRange)
for i in lowRange:
noiseExpression += cProfile[i]
for i in highRange:
noiseExpression += cProfile[i]
avgNoise = noiseExpression/float(totalLength)
#filter out peaks that look a certain way.
print highest, val[0], val[1], avgNoise
if 0 < highest < 5: #rooflength 14/26
if val[0] < 0.20 and val[1] < .20: #drop values
if avgNoise < .3:
goodTcc = cg.makeTcc(chrom, strand, x + low, x + high)
print '*KEEPER'
f.write('%s\n' % goodTcc)
f.close()
print 'DONE', tcc
import getHairpins
import cgGenes
import cgConfig as c
import bioLibCG as cg
mConf = c.getConfig('Main.conf')
geneSetFolder = mConf.conf['geneSetsHuman']
fN = '/home/chrisgre/projects/NoncodingHuman/results/NChuman-s3k8b17.results.sorted.introns.sorted'
cHairs = getHairpins.getHairpins(fN)
ensGenes = cgGenes.createGeneSetFromFile(geneSetFolder + '/ensemblAllTranscripts.tsv')
cDesc = {} #CID:gDesc
for CID in cHairs:
tcc = cHairs[CID]
cDesc[CID] = "NONE"
overlappingGenes = ensGenes.geneOverlaps([tcc])
if len(overlappingGenes) > 0:
print overlappingGenes[0].type
cDesc[CID] = overlappingGenes[0].type
f = open(fN, 'r')
newLines = []
for line in f:
CID = line.strip().split('\t')[7]
newLines.append(cg.appendToLine(line, cDesc[CID], 16))
f.close()
import cgGenes
import compareData as compare
import cgConfig as c
cName = 'mm9.conf'
mConf = c.getConfig('Main.conf')
conf = c.getConfig(cName)
organism = conf.conf['organism']
geneSetFolder = mConf.conf['geneSets%s' % organism]
genes = cgGenes.createGeneSetFromFile(geneSetFolder + '/allTransciptsType.tsv')
peakTccs = compare.tccFileToList('peakData.500.mm9', 0)
tOverlaps = genes.transcriptOverlaps(peakTccs)
typeDict = {}
for transcript in tOverlaps:
if transcript.type not in typeDict:
typeDict[transcript.type] = 1
else:
typeDict[transcript.type] += 1
#count the amounts of each type for each transcript
amount = {}
for gene in genes.genes:
for t in gene.transcripts:
if t.type in amount:
amount[t.type] += 1
else:
amount[t.type] = 1
print 'Total Peaks:', len(peakTccs)
def intronNoisy(cName = None):
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
#init
cHairs = getHairpins.getHairpins(conf.conf['resultsIntrons']) #CID: HAIRPIN
organism = conf.conf['organism']
exonList = compare.tccFileToList('%sExons.tcc' % organism, 0)
slide = 1000
#make prediction overlap hitmap
predMap = {}
predList = []
for CID in cHairs:
hPin = cHairs[CID]
predList.append(hPin)
#collapse Overlaps
print ' collapsing predictions'
predList = compare.collapseOverlaps(predList)
print ' collapsing exons'
exonList = compare.collapseOverlaps(exonList)
#collect levels for each hairpin region
cidLevels = {}
for CID in cHairs:
print CID
hPin = cHairs[CID]
chrom = ss(hPin, ':')[0]
strand = ss(hPin, ':')[1]
start = int(ss(hPin, ':')[2])
end = int(ss(hPin, ':')[3])
scanStart = start - slide
scanEnd = end + slide
scanRange = []
scanRange.append('%s:%s:%s:%s' % (chrom, strand, scanStart, start))
scanRange.append('%s:%s:%s:%s' % (chrom, strand, end, scanEnd))
print scanRange
scanRange = compare.subtractTwoTccLists(scanRange, predList)
scanRange = compare.subtractTwoTccLists(scanRange, exonList)
levels = []
print ' Retrieving Expression levels:', cg.getTccListTotalLength(scanRange)
levels = []
hPinLevels = stepVectorScan.scanVectorsHist(scanRange, cName)
for hPin in hPinLevels:
levels.extend(hPinLevels[hPin])
cidLevels[CID] = levels
#output levels to file
#find longest
longest = 0
for CID in cidLevels:
length = len(cidLevels[CID])
if length > longest:
longest = length
sortedKeys = cidLevels.keys()
sortedKeys.sort()
newLines = []
for j in range(0, longest): #how many lines are there
newLine = []
for CID in sortedKeys:
if len(cidLevels[CID]) > j:# add it
newLine.append(str(cidLevels[CID][j]))
else:
newLine.append('NA')
newLines.append('\t'.join(newLine) + '\n')
outFileN = conf.conf['intronNoiseData']
outFile = open(outFileN, 'w')
outFile.write('\t'.join(sortedKeys) + '\n')
outFile.writelines(newLines)
outFile.close()
def scanVectorsSingleCoord(tccList, cName):
'''Given tcc list --> scan wig files and coord:value...
'''
conf = c.getConfig(cName)
org = conf.conf['organism']
mConf = c.getConfig('Main.conf')
wigDir = mConf.conf['wig%s' % org]
timer = cg.cgTimer()
timer.start()
coordDict = {} # tcc: [list values]
for tcc in tccList:
theSplit = ss(tcc, ':')
chrom, strand, tccStart, tccEnd = theSplit[0], theSplit[1], int(
theSplit[2]), int(theSplit[3])
#goto correct fild, correct line in index
fN = wigDir + '/Merge.%s.%s.wig.%s.wig' % (org.lower(), strand, chrom)
fNindex = wigDir + '/Merge.%s.%s.wig.%s.wig.index' % (org.lower(),
strand, chrom)
#print timer.split()
#get line in index file
iFile = open(fNindex, 'r')
startByte = 'None'
for line in iFile:
beg = int(cg.ss(line)[1])
end = int(cg.ss(line)[2])
if beg <= tccStart < end:
startByte = int(cg.ss(line)[0])
#print 'INDEX', line.strip()
break
iFile.close()
#print timer.split()
#grab value
f = open(fN, 'r')
f.seek(startByte, 0)
stop = False
for line in f:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1])
lEnd = int(cg.ss(line)[2])
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd):
coordDict[i] = lValue
if stop: break
f.close()
return coordDict
def exonNoisy(cName = None):
#init
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
cHairs = getHairpins.getHairpins(conf.conf['resultsExons']) #CID: HAIRPIN
organism = conf.conf['organism']
geneSetFolder = mConf.conf['geneSets%s' % organism]
#make prediction overlap hitmap
print 'Making prediction list'
predList = []
for CID in cHairs:
hPin = cHairs[CID]
predList.append(hPin)
if compare.checkIfOverlaps(predList):
predList = compare.collapseOverlaps(predList)
#make genes for Ensemble/make list of tccs for exons.
print 'Creating gene set'
ensGenes = cgGenes.createGeneSetFromFile(geneSetFolder + '/ensemblAllExons.tsv')
print ' loaded # genes:', len(ensGenes.set)
#collect levels for each haipin region
print '[Checking all levels]'
cidLevels = {}
for CID in cHairs:
print CID
hPin = cHairs[CID]
#for each hairpin, --> find overlapping transcripts in same gene
overlappingGenes = ensGenes.geneOverlaps([hPin])
if len(overlappingGenes) > 0:
gIDs = [gene.id for gene in overlappingGenes]
allTccs = ensGenes.getTccsFromGIDs(gIDs)
if compare.checkIfOverlaps:
print ' Overlaps...collapsing'
allTccs = compare.collapseOverlaps(allTccs)
else:
print 'NO GENE OVERLAPS!!!!!', CID, hPin
#filter out my predictions.
print ' Filtering out predictions'
checkList = compare.subtractTwoTccLists(allTccs, predList)
#Get Expression level for gene.
print ' Retrieving Expression levels:', cg.getTccListTotalLength(checkList)
levels = []
hPinLevels = stepVectorScan.scanVectorsHist(checkList, cName)
for hPin in hPinLevels:
levels.extend(hPinLevels[hPin])
cidLevels[CID] = levels
#output levels to file
print 'Outputting to file'
#find longest
longest = 0
for CID in cidLevels:
length = len(cidLevels[CID])
if length > longest:
longest = length
sortedKeys = cidLevels.keys()
sortedKeys.sort()
#print sortedKeys
newLines = []
for j in range(0, longest): #how many lines are there
newLine = []
for CID in sortedKeys:
if len(cidLevels[CID]) > j:# add it
newLine.append(str(cidLevels[CID][j]))
else:
newLine.append('NA')
newLines.append('\t'.join(newLine) + '\n')
outFileN = conf.conf['exonNoiseData']
outFile = open(outFileN, 'w')
outFile.write('\t'.join(sortedKeys) + '\n')
outFile.writelines(newLines)
outFile.close()
def findPeaks(pType, cName = None):
#init
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
if pType == 'E':
predName = conf.conf['resultsExonsSorted']
else:
predName = conf.conf['resultsIntronsSorted']
print predName
#make CID:hairpin:peak dictionary
cHairs = getHairpins.getHairpins(predName)
peakDict = {}
for CID in cHairs:
peakDict[CID] = [cHairs[CID],'None']
timer = cg.cgTimer()
timer.start()
#put peaks in memory
print 'Creating peak data'
peaks = {} # chr:peak:value
for CID in cHairs:
chrom, strand, start, end = cg.tccSplit(cHairs[CID])
tcc = cHairs[CID]
#init dictionary
if chrom not in peaks:
peaks[chrom] = {}
if strand not in peaks[chrom]:
peaks[chrom][strand] = {}
#create peaks for tcc and add to peak dictionary
stretch = cgPeaks.stretch(tcc, cName)
stretch.createPeaks()
for peakCoord in stretch.peaks:
peaks[chrom][strand][peakCoord] = 0
print timer.split()
print 'finding best combos'
bestCombos = []
aPass = 0
bPass = 0
cPass = 0
numT = 0
for CID in peakDict:
cgFlag = False
if CID == '538':cgFlag = True
tcc = peakDict[CID][0]
#print tcc
tccPeaks = []
chrom = cg.ss(tcc, ':')[0]
strand = cg.ss(tcc, ':')[1]
start = int(cg.ss(tcc, ':')[2])
end = int(cg.ss(tcc, ':')[3])
#get all peaks
for i in range(start, end + 1):
if i in peaks[chrom][strand]:
#print ' peak added', i
tccPeaks.append(i)
#Calculate parameters...
pairStrings = [] #used to check if pair already added
peakCombos = []
for x in tccPeaks:
#scan a 30 bp range around this point and find the best roof...
pRange = 30
rTcc = cg.makeTcc(chrom, strand, x, x + 1)
#quickly get max value...kinda a long way to do it but whatever
cProfile = stepVectorScan.profileAroundPoint(rTcc, 1, cName, ratio = False)
xval = cProfile[0]
max = xval
highestValueCoord = x
#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):
if high < pRange:
val[0] = float(cProfile[startFinal - 4])
val[1] = float(cProfile[endFinal + 4])
#fill in other details...
y = 'S'
dist = 'S'
ratio = 'S'
peakCombos.append([tcc,x,y,dist,ratio,max,highest,val])
#print ' ', peakCombos[-1]
#find best combo...
topCombo = None
for combo in peakCombos:
roofLength = combo[6]
dropValue = combo[7][0]
if combo[7][1] > dropValue:
dropValue = combo[7][1]
#print roofLength, dropValue
if 14 < roofLength < 26:
if 0.0 < dropValue < 0.2:
#pick one with rooflength nearest 20:
if topCombo:
if (math.fabs(22 - roofLength)) < (math.fabs(22 - topCombo[6])):
topCombo = combo
else:
topCombo = combo
if topCombo:
peakDict[CID][1] = topCombo
bestCombos.append(topCombo)
print bestCombos[-1]
else:
#print 'None'
pass
print timer.split()
#now update predFile (SLOT 13)
predFile = open(predName, 'r')
newLines = []
for line in predFile:
CID = cg.ss(line)[7]
if peakDict[CID][1] == 'None':
peakInfo = 'None'
else:
peakInfo = '%s:%s:%s:%s:%s:%s' % (str(peakDict[CID][1][1])[-3:], 'S', str(peakDict[CID][1][4]).split('.')[0], peakDict[CID][1][5],peakDict[CID][1][6], peakDict[CID][1][7])
newLines.append(cg.appendToLine(line, peakInfo, 13))
predFile.close()
predFile = open(predName, 'w')
predFile.writelines(newLines)
predFile.close()
def updateReadDensity(tType, cName):
#go through wig each chromosome and check the mature seqs
mainConf = cgConfig.cgConfig('Main.conf')
conf = cgConfig.getConfig(cName)
organism = conf.conf['organism']
wigFolder = mainConf.conf['wig%s' % organism]
newLines = []
#Differentiate between exon or intron...
if tType == 'E':
pFileName = conf.conf['resultsExons']
elif tType == 'I':
pFileName = conf.conf['resultsIntrons']
else:
print 'READ UPDATE FAIL'
print ' Updating Read Density:', tType
#get read density for each line...
print ' calculating hits for mature seqs'
#calculate total hits per mature
mirFile = open(pFileName, 'r')
for line in mirFile:
mTcc = line.strip().split('\t')[1]
mirID = line.strip().split('\t')[0]
tccStretch = cgPeaks.stretch(mTcc, cName)
highestHit = 0
for i in range(int(mTcc.split(':')[2]), int(mTcc.split(':')[3])):
if i in tccStretch.profile:
if tccStretch.profile[i] > highestHit:
highestHit = tccStretch.profile[i]
newLines.append(cg.appendToLine(line, str(highestHit), 11))
mirFile.close()
print 'Writing New File'
#write new results file
outFile = open(pFileName, 'w')
for line in newLines:
outFile.write(line)
outFile.close()
####NOW UPDATE HIGHEST HIT PER CLUSTER####
clusterCount = {}
pFile = open(pFileName, 'r')
for line in pFile:
predictionCount = int(line.strip().split('\t')[11])
CID = line.strip().split('\t')[7]
if CID in clusterCount:
if clusterCount[CID] < predictionCount:
clusterCount[CID] = predictionCount
else:
clusterCount[CID] = predictionCount
pFile.close()
#update the file --> cluster small count
newLines = []
predFile = open(pFileName, 'r')
for line in predFile:
CID = line.strip().split('\t')[7]
numMax = clusterCount[CID]
newLines.append(cg.appendToLine(line, str(numMax), 12))
predFile.close()
#sort newLines by clusterID
sortDict = {}
CIDs = []
for line in newLines:
CID = int(line.strip().split('\t')[7])
if CID not in CIDs:
CIDs.append(CID)
if CID in sortDict:
sortDict[CID].append(line)
else:
sortDict[CID] = [line]
CIDs.sort()
newLines = []
for CID in CIDs:
for line in sortDict[CID]:
newLines.append(line)
#write new File
newFile = open(pFileName, 'w')
for line in newLines:
newFile.write(line)
newFile.close()
def updateNoise(pType, cName=None):
#init
mainConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
if pType == 'E':
predName = conf.conf['resultsExons']
else:
predName = conf.conf['resultsIntrons']
#populate cid: exon dist
print 'Populating CID/INtron/exon distribution data'
if pType == 'E':
noiseFN = conf.conf['exonNoiseData']
f = open(noiseFN, 'r')
else:
noiseFN = conf.conf['intronNoiseData']
f = open(noiseFN, 'r')
exonDists = {} #cid: [exon dist]
header = f.readline()
order = {} # num:CID
for i, CID in enumerate(header.strip().split('\t')):
order[i] = CID
exonDists[CID] = []
for line in f:
data = line.strip().split('\t')
for i, dataPoint in enumerate(data):
if dataPoint == 'NA' or dataPoint == '':
continue
else:
dataPoint = float(dataPoint)
CID = order[i]
exonDists[CID].append(dataPoint)
#get highest expression level for each cluster
print 'Populating highest expression levels'
predExpression = {} # CID; highest level
exonFile = open(predName, 'r')
for line in exonFile:
CID = line.strip().split('\t')[7]
hDensity = line.strip().split('\t')[12]
predExpression[CID] = hDensity
#get pVals for each CID
print 'Getting pvals for each cluster'
pVals = {} # CID; [lam,pVal]
for CID in exonDists:
if not len(exonDists[CID]) > 0: #no data in 2kb range.
lam = 'NA'
pVal = 'NA'
else:
lam = cgStats.getLam(exonDists[CID])
pVal = cgStats.getPValExp(predExpression[CID], lam)
pVals[CID] = [
lam, pVal
] #lam gives a good approximation of noise levels in region...
print 'Updating the file'
#update file...
predFile = open(predName, 'r')
newLines = []
for line in predFile:
CID = line.split('\t')[7]
newLine = cg.appendToLine(line, pVals[CID][0], 14)
newLine = cg.appendToLine(newLine, pVals[CID][1], 15)
newLines.append(newLine)
predFile.close()
predFile = open(predName, 'w')
predFile.writelines(newLines)
predFile.close()
def sortResults(cName=None):
#INIT
conf = cgConfig.getConfig(cName)
pFileName = conf.conf['results']
minDensity = 4
minSmallHits = 1
pFile = open(pFileName, 'r')
fileLines = [] #This will hold the lists to be sorted...
for line in pFile:
fileLines.append(line.strip().split('\t'))
pFile.close()
#highest prediction density
densityDict = {
} #CID: highest density --> used to sort out clusters without proper density
for line in fileLines:
CID = line[7]
pDensity = int(line[5])
if CID in densityDict:
if pDensity > densityDict[CID]:
densityDict[CID] = pDensity
else:
densityDict[CID] = pDensity
#take out clusters that didn't make the cut
CIDpassed = []
keptLines = []
for line in fileLines:
CID = line[7]
#smallClusterHits = int(line[10]) Not using this metric anymore...
if (densityDict[CID] >= minDensity): #Density
if line[8] == '0': #doesn't overlap with anything known (the cluster doesn't...that is)
keptLines.append(line)
if CID not in CIDpassed:
CIDpassed.append(CID)
#remake keptLines with integer in field ten
#at this point just sort by cluster density...
sID = 5
for line in keptLines:
line[sID] = int(line[sID])
sortedData = sorted(keptLines, key=itemgetter(sID),
reverse=True) #sort by small RNA hits
for line in sortedData:
line[sID] = str(line[sID])
#output
sortedFile = open(conf.conf['results'] + '.sorted', 'w')
for line in sortedData:
sortedFile.write('\t'.join(line) + '\n')
sortedFile.close()
#Now output stats
statFile = open('statFile.data', 'w')
statFile.write('Total Clusters: %s\n' % len(densityDict))
statFile.write('Passed: %s\n' % len(CIDpassed))
def findPeaks(pType, cName=None):
#init
mConf = c.cgConfig('Main.conf')
conf = c.getConfig(cName)
if pType == 'E':
predName = conf.conf['resultsExonsSorted']
else:
predName = conf.conf['resultsIntronsSorted']
print predName
#make CID:hairpin:peak dictionary
cHairs = getHairpins.getHairpins(predName)
peakDict = {}
for CID in cHairs:
peakDict[CID] = [cHairs[CID], 'None']
timer = cg.cgTimer()
timer.start()
#put peaks in memory
print 'Creating peak data'
peaks = {} # chr:peak:value
for CID in cHairs:
chrom, strand, start, end = cg.tccSplit(cHairs[CID])
tcc = cHairs[CID]
#init dictionary
if chrom not in peaks:
peaks[chrom] = {}
if strand not in peaks[chrom]:
peaks[chrom][strand] = {}
#create peaks for tcc and add to peak dictionary
stretch = cgPeaks.stretch(tcc, cName)
stretch.createPeaks()
for peakCoord in stretch.peaks:
peaks[chrom][strand][peakCoord] = 0
print timer.split()
print 'finding best combos'
bestCombos = []
aPass = 0
bPass = 0
cPass = 0
numT = 0
for CID in peakDict:
cgFlag = False
if CID == '538': cgFlag = True
tcc = peakDict[CID][0]
#print tcc
tccPeaks = []
chrom = cg.ss(tcc, ':')[0]
strand = cg.ss(tcc, ':')[1]
start = int(cg.ss(tcc, ':')[2])
end = int(cg.ss(tcc, ':')[3])
#get all peaks
for i in range(start, end + 1):
if i in peaks[chrom][strand]:
#print ' peak added', i
tccPeaks.append(i)
#Calculate parameters...
pairStrings = [] #used to check if pair already added
peakCombos = []
for x in tccPeaks:
#scan a 30 bp range around this point and find the best roof...
pRange = 30
rTcc = cg.makeTcc(chrom, strand, x, x + 1)
#quickly get max value...kinda a long way to do it but whatever
cProfile = stepVectorScan.profileAroundPoint(rTcc,
1,
cName,
ratio=False)
xval = cProfile[0]
max = xval
highestValueCoord = x
#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):
if high < pRange:
val[0] = float(cProfile[startFinal - 4])
val[1] = float(cProfile[endFinal + 4])
#fill in other details...
y = 'S'
dist = 'S'
ratio = 'S'
peakCombos.append([tcc, x, y, dist, ratio, max, highest, val])
#print ' ', peakCombos[-1]
#find best combo...
topCombo = None
for combo in peakCombos:
roofLength = combo[6]
dropValue = combo[7][0]
if combo[7][1] > dropValue:
dropValue = combo[7][1]
#print roofLength, dropValue
if 14 < roofLength < 26:
if 0.0 < dropValue < 0.2:
#pick one with rooflength nearest 20:
if topCombo:
if (math.fabs(22 - roofLength)) < (
math.fabs(22 - topCombo[6])):
topCombo = combo
else:
topCombo = combo
if topCombo:
peakDict[CID][1] = topCombo
bestCombos.append(topCombo)
print bestCombos[-1]
else:
#print 'None'
pass
print timer.split()
#now update predFile (SLOT 13)
predFile = open(predName, 'r')
newLines = []
for line in predFile:
CID = cg.ss(line)[7]
if peakDict[CID][1] == 'None':
peakInfo = 'None'
else:
peakInfo = '%s:%s:%s:%s:%s:%s' % (
str(peakDict[CID][1][1])[-3:], 'S', str(
peakDict[CID][1][4]).split('.')[0], peakDict[CID][1][5],
peakDict[CID][1][6], peakDict[CID][1][7])
newLines.append(cg.appendToLine(line, peakInfo, 13))
predFile.close()
predFile = open(predName, 'w')
predFile.writelines(newLines)
predFile.close()
def scanVectorsSingleCoord(tccList, cName):
'''Given tcc list --> scan wig files and coord:value...
'''
conf = c.getConfig(cName)
org = conf.conf['organism']
mConf = c.getConfig('Main.conf')
wigDir = mConf.conf['wig%s' % org]
timer = cg.cgTimer()
timer.start()
coordDict = {} # tcc: [list values]
for tcc in tccList:
theSplit = ss(tcc, ':')
chrom, strand, tccStart, tccEnd = theSplit[0], theSplit[1],int(theSplit[2]),int(theSplit[3])
#goto correct fild, correct line in index
fN = wigDir + '/Merge.%s.%s.wig.%s.wig' % (org.lower(),strand,chrom)
fNindex = wigDir + '/Merge.%s.%s.wig.%s.wig.index' % (org.lower(),strand,chrom)
#print timer.split()
#get line in index file
iFile = open(fNindex, 'r')
startByte = 'None'
for line in iFile:
beg = int(cg.ss(line)[1])
end = int(cg.ss(line)[2])
if beg <= tccStart < end:
startByte = int(cg.ss(line)[0])
#print 'INDEX', line.strip()
break
iFile.close()
#print timer.split()
#grab value
f = open(fN, 'r')
f.seek(startByte, 0)
stop = False
for line in f:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1])
lEnd = int(cg.ss(line)[2])
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd):
coordDict[i] = lValue
if stop: break
f.close()
return coordDict
def scanVectorsHist(tccList, cName):
'''Given tcc list --> scan wig files and get histogram values
can be modified to do single/total values...
THIS USES INDEXES!!! = BAD...'''
conf = c.getConfig(cName)
org = conf.conf['organism']
mConf = c.getConfig('Main.conf')
wigDir = mConf.conf['wig%s' % org]
timer = cg.cgTimer()
timer.start()
histDict = {} # tcc: [list values]
for tcc in tccList:
theSplit = ss(tcc, ':')
chrom, strand, tccStart, tccEnd = theSplit[0], theSplit[1], int(
theSplit[2]), int(theSplit[3])
#goto correct fild, correct line in index
fN = wigDir + '/Merge.%s.%s.wig.%s.wig' % (org.lower(), strand, chrom)
fNindex = wigDir + '/Merge.%s.%s.wig.%s.wig.index' % (org.lower(),
strand, chrom)
#print timer.split()
#get line in index file
iFile = open(fNindex, 'r')
startByte = 'None'
for line in iFile:
beg = int(cg.ss(line)[1])
end = int(cg.ss(line)[2])
if beg <= tccStart < end:
startByte = int(cg.ss(line)[0])
#print 'INDEX', line.strip()
break
iFile.close()
#print timer.split()
#grab value
f = open(fN, 'r')
f.seek(startByte, 0)
stop = False
for line in f:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1])
lEnd = int(cg.ss(line)[2])
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd):
try:
histDict[tcc].append(lValue)
except KeyError: #just for zero...so you don't have to if every time...
histDict[tcc] = [lValue]
if stop: break
f.close()
#print timer.split()
return histDict
def parallelMakePeaks(tcc, cName, minExpression):
conf = c.getConfig(cName)
f = open(
'out/peakData.%s.%s.%s' % (tcc, minExpression, conf.conf['assembly']),
'w')
print 'scanning range', tcc
chrom, strand, start, end = cg.tccSplit(tcc)
peaks = cgPeaks.stretch(tcc, cName)
#print 'getting peaks'
peaks.createPeaks(span=1, minVal=int(minExpression))
print 'len peaks', len(peaks.peaks)
endCheck = 0
for x in peaks.peaks:
print x, endCheck
'''
if x < endCheck:
print 'endChecked'
continue
'''
#scan a 30 bp range around this point and find the best roof...
pRange = 40
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 = .70
highest = 0
stretch = 0
startCurrent = None
startFinal = None
endFinal = None
for i in range(1 - pRange, pRange):
print ' ', x + i, cProfile[i]
if cProfile[i] > minVal:
print ' extending stretch'
stretch += 1
if startCurrent == None:
startCurrent = i
else:
if stretch > 0:
print 'end of stretch'
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 +/- extend value...
val = [1.0, 1.0]
extend = 1
if (startFinal) and (endFinal):
low = startFinal - extend
high = endFinal + extend
if low > (1 - pRange) and high < pRange:
val[0] = float(cProfile[startFinal - extend])
val[1] = float(cProfile[endFinal + extend])
else:
print 'out of range'
continue
else:
print 'no start and end of peak'
continue
print low, high, x, endFinal
endCheck = x + endFinal
#avg expression around peak check...
#get total expression before peak
noiseExpression = 0
lowRange = range(1 - pRange, low)
highRange = range(high + 1, pRange)
totalLength = len(lowRange) + len(highRange)
for i in lowRange:
noiseExpression += cProfile[i]
for i in highRange:
noiseExpression += cProfile[i]
avgNoise = noiseExpression / float(totalLength)
#filter out peaks that look a certain way.
print highest, val[0], val[1], avgNoise
if 0 < highest < 5: #rooflength 14/26
if val[0] < 0.20 and val[1] < .20: #drop values
if avgNoise < .3:
goodTcc = cg.makeTcc(chrom, strand, x + low, x + high)
print '*KEEPER'
f.write('%s\n' % goodTcc)
f.close()
print 'DONE', tcc
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 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 scanVectorsHist(tccList, cName):
'''Given tcc list --> scan wig files and get histogram values
can be modified to do single/total values...
THIS USES INDEXES!!! = BAD...'''
conf = c.getConfig(cName)
org = conf.conf['organism']
mConf = c.getConfig('Main.conf')
wigDir = mConf.conf['wig%s' % org]
timer = cg.cgTimer()
timer.start()
histDict = {} # tcc: [list values]
for tcc in tccList:
theSplit = ss(tcc, ':')
chrom, strand, tccStart, tccEnd = theSplit[0], theSplit[1],int(theSplit[2]),int(theSplit[3])
#goto correct fild, correct line in index
fN = wigDir + '/Merge.%s.%s.wig.%s.wig' % (org.lower(),strand,chrom)
fNindex = wigDir + '/Merge.%s.%s.wig.%s.wig.index' % (org.lower(),strand,chrom)
#print timer.split()
#get line in index file
iFile = open(fNindex, 'r')
startByte = 'None'
for line in iFile:
beg = int(cg.ss(line)[1])
end = int(cg.ss(line)[2])
if beg <= tccStart < end:
startByte = int(cg.ss(line)[0])
#print 'INDEX', line.strip()
break
iFile.close()
#print timer.split()
#grab value
f = open(fN, 'r')
f.seek(startByte, 0)
stop = False
for line in f:
#print 'Line:', line.strip()
lBeg = int(cg.ss(line)[1])
lEnd = int(cg.ss(line)[2])
lValue = int(cg.ss(line)[3].split('.')[0])
if tccStart > lBeg:
lBeg = tccStart
if tccEnd < lEnd:
lEnd = tccEnd
stop = True
#print timer.split()
for i in range(lBeg, lEnd):
try:
histDict[tcc].append(lValue)
except KeyError: #just for zero...so you don't have to if every time...
histDict[tcc] = [lValue]
if stop: break
f.close()
#print timer.split()
return histDict
def updateReadDensity(tType, cName):
#go through wig each chromosome and check the mature seqs
mainConf = cgConfig.cgConfig('Main.conf')
conf = cgConfig.getConfig(cName)
organism = conf.conf['organism']
wigFolder = mainConf.conf['wig%s' % organism]
newLines = []
#Differentiate between exon or intron...
if tType == 'E':
pFileName = conf.conf['resultsExons']
elif tType == 'I':
pFileName = conf.conf['resultsIntrons']
else:
print 'READ UPDATE FAIL'
print ' Updating Read Density:', tType
#get read density for each line...
print ' calculating hits for mature seqs'
#calculate total hits per mature
mirFile = open(pFileName, 'r')
for line in mirFile:
mTcc = line.strip().split('\t')[1]
mirID = line.strip().split('\t')[0]
tccStretch = cgPeaks.stretch(mTcc, cName)
highestHit = 0
for i in range(int(mTcc.split(':')[2]), int(mTcc.split(':')[3])):
if i in tccStretch.profile:
if tccStretch.profile[i] > highestHit:
highestHit = tccStretch.profile[i]
newLines.append(cg.appendToLine(line, str(highestHit), 11))
mirFile.close()
print 'Writing New File'
#write new results file
outFile = open(pFileName, 'w')
for line in newLines:
outFile.write(line)
outFile.close()
####NOW UPDATE HIGHEST HIT PER CLUSTER####
clusterCount = {}
pFile = open(pFileName, 'r')
for line in pFile:
predictionCount = int(line.strip().split('\t')[11])
CID = line.strip().split('\t')[7]
if CID in clusterCount:
if clusterCount[CID] < predictionCount:
clusterCount[CID] = predictionCount
else:
clusterCount[CID] = predictionCount
pFile.close()
#update the file --> cluster small count
newLines = []
predFile = open(pFileName, 'r')
for line in predFile:
CID = line.strip().split('\t')[7]
numMax = clusterCount[CID]
newLines.append(cg.appendToLine(line, str(numMax), 12))
predFile.close()
#sort newLines by clusterID
sortDict = {}
CIDs = []
for line in newLines:
CID = int(line.strip().split('\t')[7])
if CID not in CIDs:
CIDs.append(CID)
if CID in sortDict:
sortDict[CID].append(line)
else:
sortDict[CID] = [line]
CIDs.sort()
newLines = []
for CID in CIDs:
for line in sortDict[CID]:
newLines.append(line)
#write new File
newFile = open(pFileName, 'w')
for line in newLines:
newFile.write(line)
newFile.close()
def defineClusters(cName=None):
#Start Timer
timer = cg.cgTimer()
timer.start()
#Get list of mature tccs
conf = cgConfig.getConfig(cName) #passed or default
finalMirFileName = conf.conf['resultsRaw']
matureTccs = compare.tccFileToList(finalMirFileName,
1) # list of all mature micro in tcc
print 'List getting', timer.split()
#make connections dict
matureConnections = compare.makeConnectionsDict(matureTccs)
print 'Make connections:', timer.split()
#Now have to define Clusters...
clusters = []
addedList = []
#I don't think python passes by reference? also I think this function is in the middle because it uses a global variable :P
def createClusters(item=None, mode=None):
if item in addedList:
return 0
elif mode == "top":
clusters.append([item])
addedList.append(
item) ##creates new cluster with the item already stored in it
for connectedItem in matureConnections[item]:
createClusters(connectedItem, "neighbor")
elif mode == "neighbor":
clusters[-1].append(
item) #add this item to the last cluster created
addedList.append(item)
for connectedItem in matureConnections[item]:
createClusters(connectedItem, "neighbor")
for tcc in matureTccs:
createClusters(tcc, "top")
print 'Make Clusters', timer.split()
#Sort Clusters.
sortedClusters = []
for cluster in clusters:
sortedClusters.append(cg.sortTccList(cluster))
print 'Sort Clusters:', timer.split()
#Output sorted cluster file
clusterFileName = conf.conf['sortedClusters']
clusterFile = open(clusterFileName, 'w')
for cluster in sortedClusters:
for hit in cluster:
clusterFile.write('%s,' % hit)
clusterFile.write('\n')
clusterFile.close()
'''
#re-create sortedClusters list:
clusterFileName = 'sortedClusters.data'
clusterFile = open(clusterFileName, 'r')
sortedClusters = []
for line in clusterFile:
sortedClusters.append([])
line = line.strip()[0:-1] #take off last comma ;P
for hit in (line.strip().split(',')):
sortedClusters[-1].append(hit)
'''
print 'Store intermediate data:', timer.split()
#output hitsAround file
outputFile = open(conf.conf['hitsPerFrame'], 'w')
frameLength = 200
frameShift = 1
for cluster in sortedClusters:
#grab first and last coordinate from cluster, for each cluster deduce how many theoretical microRNAs were in hitScope
clusterChrom = cluster[0].split(":")[0]
clusterStrand = cluster[0].split(":")[1]
firstCoord = int(cluster[0].split(":")[2])
#print cluster[-1]
lastCoord = int(cluster[-1].split(":")[3])
startCoord = firstCoord
while startCoord < lastCoord:
#count how many hits there are in this range
rangeStart = startCoord - (frameLength / 2)
rangeEnd = startCoord + (frameLength / 2)
rangeTcc = '%s:%s:%s:%s' % (clusterChrom, clusterStrand,
rangeStart, rangeEnd)
overlappedList = compare.compareTwoTcc([rangeTcc], cluster, 2)
hitCount = len(overlappedList)
#output
outputFile.write('%s\t%s\n' % (rangeTcc, hitCount))
startCoord = startCoord + frameShift #check overlap with range
outputFile.close()
print 'Output Hits per Frame:', timer.split()
print 'Overall Time:', timer.report()
import getHairpins
import cgGenes
import cgConfig as c
import bioLibCG as cg
mConf = c.getConfig('Main.conf')
geneSetFolder = mConf.conf['geneSetsHuman']
fN = '/home/chrisgre/projects/NoncodingHuman/results/NChuman-s3k8b17.results.sorted.introns.sorted'
cHairs = getHairpins.getHairpins(fN)
ensGenes = cgGenes.createGeneSetFromFile(geneSetFolder +
'/ensemblAllTranscripts.tsv')
cDesc = {} #CID:gDesc
for CID in cHairs:
tcc = cHairs[CID]
cDesc[CID] = "NONE"
overlappingGenes = ensGenes.geneOverlaps([tcc])
if len(overlappingGenes) > 0:
print overlappingGenes[0].type
cDesc[CID] = overlappingGenes[0].type
f = open(fN, 'r')
newLines = []
for line in f:
CID = line.strip().split('\t')[7]
newLines.append(cg.appendToLine(line, cDesc[CID], 16))
f.close()
def defineClusters(cName = None):
#Start Timer
timer = cg.cgTimer()
timer.start()
#Get list of mature tccs
conf = cgConfig.getConfig(cName) #passed or default
finalMirFileName = conf.conf['resultsRaw']
matureTccs = compare.tccFileToList(finalMirFileName, 1) # list of all mature micro in tcc
print 'List getting', timer.split()
#make connections dict
matureConnections = compare.makeConnectionsDict(matureTccs)
print 'Make connections:', timer.split()
#Now have to define Clusters...
clusters = []
addedList = []
#I don't think python passes by reference? also I think this function is in the middle because it uses a global variable :P
def createClusters(item = None, mode = None):
if item in addedList:
return 0
elif mode == "top":
clusters.append([item])
addedList.append(item) ##creates new cluster with the item already stored in it
for connectedItem in matureConnections[item]:
createClusters(connectedItem, "neighbor")
elif mode == "neighbor":
clusters[-1].append(item) #add this item to the last cluster created
addedList.append(item)
for connectedItem in matureConnections[item]:
createClusters(connectedItem, "neighbor")
for tcc in matureTccs:
createClusters(tcc, "top")
print 'Make Clusters', timer.split()
#Sort Clusters.
sortedClusters = []
for cluster in clusters:
sortedClusters.append(cg.sortTccList(cluster))
print 'Sort Clusters:', timer.split()
#Output sorted cluster file
clusterFileName = conf.conf['sortedClusters']
clusterFile = open(clusterFileName, 'w')
for cluster in sortedClusters:
for hit in cluster:
clusterFile.write('%s,' % hit)
clusterFile.write('\n')
clusterFile.close()
'''
#re-create sortedClusters list:
clusterFileName = 'sortedClusters.data'
clusterFile = open(clusterFileName, 'r')
sortedClusters = []
for line in clusterFile:
sortedClusters.append([])
line = line.strip()[0:-1] #take off last comma ;P
for hit in (line.strip().split(',')):
sortedClusters[-1].append(hit)
'''
print 'Store intermediate data:', timer.split()
#output hitsAround file
outputFile = open(conf.conf['hitsPerFrame'], 'w')
frameLength = 200
frameShift = 1
for cluster in sortedClusters:
#grab first and last coordinate from cluster, for each cluster deduce how many theoretical microRNAs were in hitScope
clusterChrom = cluster[0].split(":")[0]
clusterStrand = cluster[0].split(":")[1]
firstCoord = int(cluster[0].split(":")[2])
#print cluster[-1]
lastCoord = int(cluster[-1].split(":")[3])
startCoord = firstCoord
while startCoord < lastCoord:
#count how many hits there are in this range
rangeStart = startCoord - (frameLength/2)
rangeEnd = startCoord + (frameLength/2)
rangeTcc = '%s:%s:%s:%s' % (clusterChrom, clusterStrand, rangeStart, rangeEnd)
overlappedList = compare.compareTwoTcc([rangeTcc], cluster, 2)
hitCount = len(overlappedList)
#output
outputFile.write('%s\t%s\n' % (rangeTcc, hitCount))
startCoord = startCoord + frameShift #check overlap with range
outputFile.close()
print 'Output Hits per Frame:', timer.split()
print 'Overall Time:', timer.report()
