def p_g_1D(k,g,l):
if k > l:
print >> sys.stderr, "Warning: there are not %s tbs in the window of size %s" % (k,l)
return None
w = (k-1)*g+k
num = max(0,l-w+1)*math.pow((g+1),k-1)+d_0(k,g,min(l,w-1))
den = combinations(l,k)
return float(num)/den
def p_d(k,la,lb,nab,na,nb):
if nab > min(na,nb):
print >> sys.stderr, "Warning: during the computation of p_d(k=%s,la=%s,lb=%s,nab=%s,na=%s,nb=%s), there was too many dispersed paralogies in the MHP, thus for the calculation we chose nab=min(na,nb)=%s" % (k,la,lb,nab,na,nb,min(na,nb))
nab = min(na,nb)
elif k > nab:
print >> sys.stderr, "Warning: not able to compute p_d(k=%s,la=%s,lb=%s,nab=%s,na=%s,nb=%s) because there are not %s hps in the MHP" % (k,la,lb,nab,na,nb,k)
return None
elif k > min([la, lb]):
print >> sys.stderr, "Warning: not able to compute p_d(k=%s,la=%s,lb=%s,nab=%s,na=%s,nb=%s) because there are too many dispersed paralogies in the window" % (k,la,lb,nab,na,nb)
return None
limSum = min(la-k,nab-k)
sum_ = 0
for i in range(0,limSum+1):
try:
foo = combinations(nab-k,i)*combinations(na-nab,la-(k+i))*combinations(nb-(k+i),lb-k)
except Exception as e:
print >> sys.stderr, "Warning: not able to compute p_d(k=%s,la=%s,lb=%s,nab=%s,na=%s,nb=%s) because %s" % (k,la,lb,nab,na,nb,e)
foo = 0
sum_ += foo
num = combinations(nab,k)*sum_
den = combinations(na,la)*combinations(nb,lb)
if den == float('inf') and num != float('inf'):
return 0.0
elif den == float('inf') and num == float('inf'):
return None
elif den == 0.0 and num != float('inf'):
return float('inf')
elif den == 0.0 and num == float('inf'):
return None
else :
return float(num)/den
num = combinations(nab,k)*sum_
den = combinations(na,la)*combinations(nb,lb)
if den == float('inf') and num != float('inf'):
return 0.0
elif den == float('inf') and num == float('inf'):
return None
elif den == 0.0 and num != float('inf'):
return float('inf')
elif den == 0.0 and num == float('inf'):
return None
else :
return float(num)/den
d_0 = lambda k,g,l : sum([((-1)**i)*combinations(k-1,i)*combinations(l-i*(g+1),k) for i in range(0,int(math.floor(float(l-k)/float(g+1))+1))])
# probability that k marked tbs form a maxgap(g)-cluster in a sequence of l tbs
def p_g_1D(k,g,l):
if k > l:
print >> sys.stderr, "Warning: there are not %s tbs in the window of size %s" % (k,l)
return None
w = (k-1)*g+k
num = max(0,l-w+1)*math.pow((g+1),k-1)+d_0(k,g,min(l,w-1))
den = combinations(l,k)
return float(num)/den
# probability that k marked hps form a maxgap(g)-cluster in both sequences of lengths la and lb
def p_g_2D(k,g,la,lb):
if k > min(la,lb):
print >> sys.stderr, "Warning: there are not %s tbs in the window of size %sx%s" % (k,la,lb)