def callBinPeak(self,pvalue=1e-05):
self.pvalue=pvalue
self.binpeak=[]
threshold=1
while 1:
if prob.poisson_cdf(threshold,self.lam,False) < pvalue: break
threshold+=1
for i in range(len(self.bins)):
print >>sys.stderr,"Call Bins Pvalue in",self.chrs[i],"\r",
self.binpeak.append([False for row in range(len(self.bins[i]))])
for j in range(len(self.bins[i])):
if self.bins[i][j]>=threshold:
# if prob.poisson_cdf(self.bins[i][j],self.lam,False)<pvalue:
# if prob.poisson_cdf(self.bins[i][j],self.lam,False)<pvalue:
self.binpeak[i][j]=True
print >>sys.stderr,"Call Bins Pvalue Done "
def callBinPeak(self,pvalue=1e-05):
self.pvalue=pvalue
self.binpeak=[]
threshold=1
while 1:
if prob.poisson_cdf(threshold,self.lam,False) < pvalue: break
threshold+=1
for i in range(len(self.bins)):
print >>sys.stderr,"Call Bins Pvalue in",self.chrs[i],"\r",
self.binpeak.append([False for row in range(len(self.bins[i]))])
for j in range(len(self.bins[i])):
if self.bins[i][j]>=threshold:
# if prob.poisson_cdf(self.bins[i][j],self.lam,False)<pvalue:
# if prob.poisson_cdf(self.bins[i][j],self.lam,False)<pvalue:
self.binpeak[i][j]=True
print >>sys.stderr,"Call Bins Pvalue Done "
def iterPeaks(self):
self.peaks = []
print >> sys.stderr, "Reading ", self.bamfilename, " HMM Segments"
k = 0
for i in self.parse_segment():
k += 1
if k % 10000 == 0:
print >> sys.stderr, "parsed ", k, " segments\r",
(chr, start, stop) = i
a = self.samfile.fetch(chr, start, stop)
refine_start = None
refine_stop = None
reads_num = 0
for j in a:
reads_num += 1
if refine_start == None:
refine_start = j.pos
if refine_stop == None:
refine_stop = j.pos + j.qlen
if j.pos < refine_start: refine_start = j.pos
if j.pos + j.qlen > refine_stop: refine_stop = j.pos + j.qlen
peak_intensity = 0
peak_pos = 0
coverage = 0.0
for pileupcolumn in self.samfile.pileup(chr, refine_start,
refine_stop):
if pileupcolumn.n > peak_intensity:
peak_intensity = pileupcolumn.n
peak_pos = pileupcolumn.pos
coverage += pileupcolumn.n
coverage /= (refine_stop - refine_start)
lam = float(self.mapped) / (self.total_length /
(refine_stop - refine_start))
pvalue = prob.poisson_cdf(reads_num, lam, False)
yield Peak([
chr, refine_start, refine_stop, reads_num, pvalue, coverage,
peak_pos, peak_intensity
])
print >> sys.stderr, "Reading HMM Segments Done! "
def iterPeaks(self):
self.peaks=[]
print >>sys.stderr,"Reading ",self.bamfilename," HMM Segments"
k=0
for i in self.parse_segment():
k+=1
if k%10000==0 :
print >>sys.stderr,"parsed ",k," segments\r",
(chr,start,stop)=i
a=self.samfile.fetch(chr,start,stop)
refine_start=None
refine_stop=None
reads_num=0
for j in a:
reads_num+=1
if refine_start==None:
refine_start=j.pos
if refine_stop==None:
refine_stop=j.pos+j.qlen
if j.pos < refine_start: refine_start=j.pos
if j.pos+j.qlen > refine_stop: refine_stop=j.pos+j.qlen
peak_intensity=0
peak_pos=0
coverage=0.0
for pileupcolumn in self.samfile.pileup(chr,refine_start,refine_stop):
if pileupcolumn.n > peak_intensity:
peak_intensity=pileupcolumn.n
peak_pos=pileupcolumn.pos
coverage+=pileupcolumn.n
coverage/=(refine_stop-refine_start)
lam=float(self.mapped)/(self.total_length/(refine_stop-refine_start))
pvalue=prob.poisson_cdf(reads_num,lam,False)
yield (chr,refine_start,refine_stop,reads_num,pvalue,coverage,peak_pos,peak_intensity)
print >>sys.stderr,"Reading HMM Segments Done! "