def __init__(self,filename,sample=0):
assert sample.lower() in ["l","r","left","right", "absdiff", "reldiff", "folddiff"]
### treatment is THE TREE FOR CpGs - allows rapid searching.
self.treatment = AggregateTree(filename)
### What is sample lower?
self.sample = sample.lower()
self.valuesBehaviour = missingValuesDontCount
class CpGMethPercentDifference(DataBehaviour):
def __init__(self,filename):
self.treatment = AggregateTree(filename)
self.valuesBehaviour = missingValuesDontCount
def methPercentage(self,values):
lmethTotal = 0
lunmethTotal = 0
rmethTotal = 0
runmethTotal = 0
for (lmeth,lunmeth,rmeth,runmeth) in values:
lmethTotal += lmeth
lunmethTotal += lunmeth
rmethTotal += rmeth
runmethTotal += runmeth
lmeth = float(lmethTotal)
lunmeth = float(lunmethTotal)
# arbitrary cutoff for regions with very few reads
if lmeth + lunmeth <= 10.0:
return []
rmeth = float(rmethTotal)
runmeth = float(runmethTotal)
# arbitrary cutoff for regions with very few reads
if rmeth + runmeth <= 10.0:
return []
lpercentage = lmeth / (lmeth+lunmeth)
rpercentage = rmeth / (rmeth+runmeth)
return [rpercentage - lpercentage]
# get the cpg meth percentage for a given range (chrm,start,stop)
def getValues(self,chrm, start, stop):
return self.methPercentage(self.treatment.getValuesInRange(chrm,start,stop))
elif assembly == "hg19":
cpgIslands = ExtendedBed(
os.path.expanduser(
"~/mount/publicdata/hg19/CpGIslands/cpgislands.bed"))
lINEs = ExtendedBed(
os.path.expanduser(
"~/mount/publicdata/hg19/Repeats/UCSC_HG19_LINEs.bed"),
defaultkeys=["chrom", "chromStart", "chromEnd", "name", "strand"])
sINEs = ExtendedBed(
os.path.expanduser(
"~/mount/publicdata/hg19/Repeats/UCSC_HG19_SINEs.bed"),
defaultkeys=["chrom", "chromStart", "chromEnd", "name", "strand"])
else:
assert False, "Unknown genome build for cpgislands / lines / sines"
methdata = AggregateTree(methdatafile, reverse=reverse)
def methPercentageDiff(values):
assert values != None
lmethTotal = 0
lunmethTotal = 0
rmethTotal = 0
runmethTotal = 0
for (lmeth, lunmeth, rmeth, runmeth) in values:
lmethTotal += lmeth
lunmethTotal += lunmeth
rmethTotal += rmeth
runmethTotal += runmeth
def __init__(self,filename):
self.treatment = AggregateTree(filename)
self.valuesBehaviour = missingValuesDontCount
class CpGMethPercent(DataBehaviour):
#Takes the CpG format input, plus left/right
def __init__(self,filename,sample=0):
assert sample.lower() in ["l","r","left","right", "absdiff", "reldiff", "folddiff"]
### treatment is THE TREE FOR CpGs - allows rapid searching.
self.treatment = AggregateTree(filename)
### What is sample lower?
self.sample = sample.lower()
self.valuesBehaviour = missingValuesDontCount
# For a window?
def methPercentage(self,values):
lmethTotal = 0
lunmethTotal = 0
rmethTotal = 0
runmethTotal = 0
for (lmeth,lunmeth,rmeth,runmeth) in values:
lmethTotal += lmeth
lunmethTotal += lunmeth
rmethTotal += rmeth
runmethTotal += runmeth
lmeth = float(lmethTotal)
lunmeth = float(lunmethTotal)
# arbitrary cutoff for regions with very few reads
if lmeth + lunmeth <= 10.0:
return []
rmeth = float(rmethTotal)
runmeth = float(runmethTotal)
# arbitrary cutoff for regions with very few reads
if rmeth + runmeth <= 10.0:
return []
#Adds an ~0 value to each total (removes a division by 0 error)
lmeth += 0.0000000001
lunmeth += 0.0000000001
rmeth += 0.0000000001
runmeth += 0.0000000001
# Calculates the percentages and differences
lpercentage = lmeth / (lmeth+lunmeth)
rpercentage = rmeth / (rmeth+runmeth)
absDiff = rpercentage - lpercentage
relDiff = absDiff/lpercentage
foldDiff = math.log(rpercentage,2) - math.log(lpercentage,2)
if relDiff > 5:
relDiff = 5
elif relDiff < -5:
relDiff = -5
if foldDiff > 5:
foldDiff = 5
elif foldDiff < -5:
foldDiff = -5
if self.sample.startswith("l"):
return [lpercentage]
elif self.sample.startswith("rel"):
return [relDiff]
elif self.sample.startswith("a"):
return [absDiff]
elif self.sample.startswith("f"):
return [absDiff]
else:
return [rpercentage]
# get the cpg meth percentage for a given range (chrm,start,stop)
def getValues(self,chrm, start, stop):
return self.methPercentage(self.treatment.getValuesInRange(chrm,start,stop))
# Tests for an appropriate R data range
def heatmapHasNegativeValues(self):
if self.sample=="absdiff":
return True
elif self.sample=="reldiff":
return True
else:
return False
def heatmapUpperLowerBounds(self):
if self.sample=="absdiff":
return -1,1
elif self.sample=="reldiff":
return -5,5
else:
return None, None
assert methdatafile != None
assert regionsfile != None
assert outputfile != None
assert genomeBuild != None
if printThreshold == None:
printThreshold = 10
if isinstance(printThreshold, int) is False:
printThreshold = 10
if genomeBuild not in ("hg19", "hg18", "mm9"):
genomeBuild = "hg19"
print "Genome build type unacceptable. Defaulting to genome hg19..."
regions = SimpleBed(regionsfile)
methdata = AggregateTree(methdatafile)
genome = Genome(genomeBuild)
def methChiSquared(pooledMeth):
chi2, p = None, None
try:
chi2, p, dof, expected = scipy.stats.chi2_contingency(pooledMeth)
except ValueError:
p = 1.0
return chi2, p
def methTotals(values):
lmethTotal = 0
lunmethTotal = 0
rmethTotal = 0
for affy in affyCSV:
ensembls = affyannotation.getValues(affy, "Ensembl")
if len(ensembls) == 1:
affyFC = float(affyCSV[affy][affyfccol])
affylogFC = math.log(affyFC) if affyFC > 0.0 else math.log(
abs(affyFC)) * -1.0
affyEnsemblLogFCs[ensembls[0]].append(affylogFC)
affyexpr = float(affyCSV[affy][affyexprcol])
affyEnsemblExprs[ensembls[0]].append(affyexpr)
affyp = float(affyCSV[affy][affypcol])
affyEnsemblPvalues[ensembls[0]].append(affyp)
methdata = AggregateTree(methdatafile)
def methPercentageDiff(values):
lmethTotal = 0
lunmethTotal = 0
rmethTotal = 0
runmethTotal = 0
for (lmeth, lunmeth, rmeth, runmeth) in values:
lmethTotal += lmeth
lunmethTotal += lunmeth
rmethTotal += rmeth
runmethTotal += runmeth
lmeth = float(lmethTotal)
lunmeth = float(lunmethTotal)