class ChIPWithControl():
#Constructor
def __init__(self, ChIPFilename, controlFilename, extend=None):
self.ChIPData = BedFile(ChIPFilename, extend=extend)
self.ControlData = BedFile(controlFilename, extend=extend)
#Gets the tag counts for both ChIP and control datasets
def getTagsInRegion(self, chrm, start, stop):
# store a tuple of sample/cntrl values
return self.ChIPData.getIntervalsInRange(
chrm, start,
stop), self.ControlData.getIntervalsInRange(chrm, start, stop)
#Gets the reads per million for the library:
def getLibraryNormalizationFactor(self, data):
return data.count / 1000000
#Gets normalised tag counts for a region - using a variety of normalisation strategies
def normaliseTags(self, chrm, start, end, chipTagsInRegion,
controlTagsInRegion):
assert end > start, str(end) + " > " + str(start)
chipFractionalTagsInRegion = getFractionalTagsInRegion(
chrm, start, end, chipTagsInRegion)
controlFractionalTagsInRegion = getFractionalTagsInRegion(
chrm, start, end, controlTagsInRegion)
chipTags = math.fsum(chipFractionalTagsInRegion)
chipTagsRegionNormalised = chipTags / float(end - start)
chipTagsLibraryNormalised = chipTags / self.getLibraryNormalizationFactor(
self.ChIPData)
chipTagsRegionLibraryNormalised = chipTagsRegionNormalised / self.getLibraryNormalizationFactor(
self.ChIPData)
controlTags = math.fsum(controlFractionalTagsInRegion)
controlTagsRegionNormalised = controlTags / float(end - start)
controlTagsLibraryNormalised = controlTags / self.getLibraryNormalizationFactor(
self.ControlData)
controlTagsRegionLibraryNormalised = controlTagsRegionNormalised / self.getLibraryNormalizationFactor(
self.ControlData)
diffTags = chipTags - controlTags
diffTagsRegionNormalised = chipTagsRegionNormalised - controlTagsRegionNormalised
diffTagsLibraryNormalised = chipTagsLibraryNormalised - controlTagsLibraryNormalised
diffTagsRegionLibraryNormalised = chipTagsRegionLibraryNormalised - controlTagsRegionLibraryNormalised
if (controlTags > 0):
ratioTags = chipTags / controlTags
ratioTagsRegionNormalised = chipTagsRegionNormalised / controlTagsRegionNormalised
ratioTagsLibraryNormalised = chipTagsLibraryNormalised / controlTagsLibraryNormalised
ratioTagsRegionLibraryNormalised = chipTagsRegionLibraryNormalised / controlTagsRegionLibraryNormalised
else:
ratioTags = 0.0
ratioTagsRegionNormalised = 0.0
ratioTagsLibraryNormalised = 0.0
ratioTagsRegionLibraryNormalised = 0.0
return chipTags, chipTagsRegionNormalised, chipTagsLibraryNormalised, chipTagsRegionLibraryNormalised, controlTags, controlTagsRegionNormalised, controlTagsLibraryNormalised, controlTagsRegionLibraryNormalised, diffTags, diffTagsRegionNormalised, diffTagsLibraryNormalised, diffTagsRegionLibraryNormalised, ratioTags, ratioTagsRegionNormalised, ratioTagsLibraryNormalised, ratioTagsRegionLibraryNormalised
class BedWithControl(Bed):
# sample is a Bed from above
def __init__(self,controlFilename,sample,extend=None,fractions=True,normalizeLibrarySize=True):
# control is already loaded in the sample Bed object
self.sample = sample
self.control = BedFile(controlFilename,extend=extend)
self.fractions = fractions
self.normalizeLibrarySize = normalizeLibrarySize # divide by 1mil to get a normalization factor
assert self.fractions == self.sample.fractions, "Bed and BedWithControl in different formats"
assert self.normalizeLibrarySize == self.sample.normalizeLibrarySize
def getValues(self,chrm, start, stop):
# store a tuple of sample/cntrl values
return self.sample.getValues(chrm,start,stop),self.control.getIntervalsInRange(chrm, start, stop)
def getLibraryNormalizationFactor(self):
return self.control.count / 1000000
def valuesBehaviour(self,chrm,values,start,end):
sampleValues, controlValues = values
if self.fractions:
return [self.sample.fractionallyCountByBP(chrm,sampleValues,start,end) - self.fractionallyCountByBP(chrm,controlValues,start,end)]
else:
return [self.sample.count(chrm,sampleValues,start,end) - self.count(chrm,controlValues,start,end)]
def __init__(self,controlFilename,sample,extend=None,fractions=True,normalizeLibrarySize=True):
# control is already loaded in the sample Bed object
self.sample = sample
self.control = BedFile(controlFilename,extend=extend)
self.fractions = fractions
self.normalizeLibrarySize = normalizeLibrarySize # divide by 1mil to get a normalization factor
assert self.fractions == self.sample.fractions, "Bed and BedWithControl in different formats"
assert self.normalizeLibrarySize == self.sample.normalizeLibrarySize
def __init__(self, ChIPFilename, controlFilename, extend=None):
self.ChIPData = BedFile(ChIPFilename, extend=extend)
self.ControlData = BedFile(controlFilename, extend=extend)
sys.exit(2)
# add executing directory as part of path
sys.path.append(sys.path[0])
for o, a in opts:
if (o == "-d"):
debug = True
for o, a in opts:
if (o == "-d"):
pass # we dealt with this already
elif o == "-b":
treatmentFileName = a
if debug:
replaceTreatment = BedTreatment(treatmentFileName, 50000)
else:
replaceTreatment = BedTreatment(treatmentFileName)
elif o == "-a":
baseTreatmentFileName = a
elif o == "-o":
outputFile = a
else:
assert False, "Unhandled option"
baseTreatment = csv.reader(open(baseTreatmentFileName, "r"),
delimiter='\t')
outputFile = csv.writer(open(outputFile, "w"), delimiter='\t')
for row in baseTreatment:
if len(row) == 0 or row[0].startswith("#"):
# print str(err) # will print something like "option -a not recognized"
usage()
sys.exit(2)
# add executing directory as part of path
sys.path.append(sys.path[0])
treatmentFileName = None
baseTreatmentFileName = None
outputFile = None
pad = 0
for o, a in opts:
if o == "-b":
treatmentFileName = a
subtractTreatment = BedTreatment(treatmentFileName)
elif o == "-a":
baseTreatmentFileName = a
elif o == "-o":
outputFile = a
elif o == "-p":
pad = int(a)
else:
assert False, "Unhandled option"
if treatmentFileName is None or baseTreatmentFileName is None or outputFile is None:
usage()
sys.exit(2)
baseTreatment = csv.reader(open(baseTreatmentFileName, "r"),
delimiter='\t')