def SC_intuit(self, seq): #def SC_intuit(self,seq,minSizePerc=1.0):
"""
This method REQUIRES that len(seq) == self.n
Right now we consider only sequences with length higher or equal to the length of the motif"""
if not self.validMotif:
# hacer que se informe de que el motivo no mola y no se va a tener en cuenta. La funcion debe devolver un valor que informe del hecho. Por ejemplo se me ocurre None
pass
if (len(seq) != self.n):
raise AttributeError("SC_intuit REQUIRES that len(seq) == self.n")
#if not self._isValidSeq(seq,minSizePerc):
# self.SC_intuit_forSmallerSeqs(seq) # TODO! YOu would need to program this method
else:
drow = {}
drevrow = {}
conti = 0
for i in ['A','C','G','T']:
for j in ['A','C','G','T']:
drow[str(i)+str(j)] = conti
drevrow[conti] = str(i)+str(j)
conti += 1
dcol = {}
drevcol = {}
conti = 0
for i in range(self.n):
for j in range(i+1,self.n):
dcol[(i,j)] = conti
drevcol[conti] = (i,j)
conti += 1
A = [self.intuitM.transpose(),self.intuitV.transpose()]
simsNorm = []
for indexi,posi in enumerate(seq[0:-1]):
for indexj,posj in enumerate(seq[indexi+1:]):
maxSim,minSim = utils.computeMaxMin(A,dcol[(indexi,indexj+ indexi +1)])
mu = A[0][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]]
maxV = max(A[1][dcol[(indexi,indexj+ indexi +1)]])
v = maxV - A[1][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]]
curSim = mu * v
curSimNorm = (curSim - minSim) / (maxSim -minSim)
simsNorm.append(curSimNorm)
return average(simsNorm)
def SC_intuit_forSmallerSeqs(self, smallSeq, A):
#pass
# TODO!
# El procedimiento seria:
# values = []
# para cada submotivo necesario:
# currSubMotivo = self._createSubmotif(ini,fin)
# values.append(currSubMotivo.Sc_intuit(seq))
# return values
drow = {}
drevrow = {}
conti = 0
for i in ['A','C','G','T']:
for j in ['A','C','G','T']:
drow[str(i)+str(j)] = conti
drevrow[conti] = str(i)+str(j)
conti += 1
dcol = {}
drevcol = {}
conti = 0
for i in range(len(smallSeq)):
for j in range(i+1,len(smallSeq)):
dcol[(i,j)] = conti
drevcol[conti] = (i,j)
conti += 1
#print shape(A[0]), A[0]
#print shape(A[1]), A[1]
#print self.validSeqs
simsNorm = []
for indexi,posi in enumerate(smallSeq[0:-1]):
for indexj,posj in enumerate(smallSeq[indexi+1:]):
maxSim,minSim = utils.computeMaxMin(A,dcol[(indexi,indexj+ indexi +1)])
mu = A[0][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]]
maxV = max(A[1][dcol[(indexi,indexj+ indexi +1)]])
v = maxV - A[1][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]] #If there are many 'N', same values can be in A[1][dcol[(indexi,indexj+ indexi +1)]] and if this value is tmaximum -> v = 0 -> curSim = 0 maxSim=0 and minSim=0 (see utils.computeMaxMin) -> curSimNorm = nan
curSim = mu * v
curSimNorm = (curSim - minSim) / (maxSim -minSim)
#print A[1][dcol[(indexi,indexj+ indexi +1)]]
#print [drow[posi+posj]]
#print A[1][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]]
#print mu, maxV, curSim, maxSim, minSim, (maxSim -minSim), curSimNorm
#raw_input("Press Enter to continue...")
simsNorm.append(curSimNorm)
return average(simsNorm)
def SC_intuit(self,seq):
self.computeIntuitionisticMotif()
drow = {}
drevrow = {}
conti = 0
for i in ['A','C','G','T']:
for j in ['A','C','G','T']:
drow[str(i)+str(j)] = conti
drevrow[conti] = str(i)+str(j)
conti += 1
dcol = {}
drevcol = {}
conti = 0
for i in range(self.n):
for j in range(i+1,self.n):
dcol[(i,j)] = conti
drevcol[conti] = (i,j)
conti += 1
A = [self.intuitM.transpose(),self.intuitV.transpose()]
simsNorm = []
for indexi,posi in enumerate(seq[0:-1]):
for indexj,posj in enumerate(seq[indexi+1:]):
maxSim,minSim = utils.computeMaxMin(A,dcol[(indexi,indexj+ indexi +1)])
mu = A[0][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]]
maxV = max(A[1][dcol[(indexi,indexj+ indexi +1)]])
v = maxV - A[1][dcol[(indexi,indexj+ indexi +1)]][drow[posi+posj]]
curSim = mu * v
curSimNorm = (curSim - minSim) / (maxSim -minSim)
simsNorm.append(curSimNorm)
return average(simsNorm)
