def conv_align2GRmat(aliDfile, pScoreCutoff, aliFormat):
in1 = open(aliDfile, 'r')
U = {}
NU = {}
h_readId = {}
h_refId = {}
genomes = []
read = []
gCnt = 0
rCnt = 0
maxScore = None
minScore = None
for ln in in1:
if (ln[0] == '@' or ln[0] == '#'):
continue
l = ln.split('\t')
readId = l[0]
if (aliFormat == 0 or aliFormat == 1): # gnu-sam or sam
if int(l[1]) & 0x4 == 4: # bitwise FLAG - 0x4 : segment unmapped
continue
refId = l[2]
elif (aliFormat == 2): # bl8
refId = l[1]
if refId == '*':
continue
#refId=refId.split("ti:")[-1]
mObj = re.search(r'ti\|(\d+)\|org\|([^|]+)\|', refId)
if mObj:
refId = "ti|" + mObj.group(1) + "|org|" + mObj.group(2)
else:
mObj = re.search(r'ti\|(\d+)\|', refId)
if mObj and mObj.group(1) != "-1":
refId = "ti|" + mObj.group(1)
(pScore, skipFlag) = find_entry_score(ln, l, aliFormat, pScoreCutoff)
if skipFlag:
continue
if ((maxScore == None) or (pScore > maxScore)):
maxScore = pScore
if ((minScore == None) or (pScore < minScore)):
minScore = pScore
gIdx = h_refId.get(refId, -1)
if gIdx == -1:
gIdx = gCnt
h_refId[refId] = gIdx
genomes.append(refId)
gCnt += 1
rIdx = h_readId.get(readId, -1)
if rIdx == -1:
#hold on this new read
#first, wrap previous read profile and see if any previous read has a same profile with that!
rIdx = rCnt
h_readId[readId] = rIdx
read.append(readId)
rCnt += 1
U[rIdx] = [[gIdx], [pScore], [float(pScore)], pScore]
else:
if (rIdx in U):
if gIdx in U[rIdx][0]:
continue
NU[rIdx] = U[rIdx]
del U[rIdx]
if gIdx in NU[rIdx][0]:
continue
NU[rIdx][0].append(gIdx)
NU[rIdx][1].append(pScore)
if pScore > NU[rIdx][3]:
NU[rIdx][3] = pScore
# length = len(NU[rIdx][1])
# NU[rIdx][2] = [1.0/length]*length
in1.close()
if (aliFormat == 1): # sam
(U, NU) = samUtils.rescale_samscore(U, NU, maxScore, minScore)
del h_refId, h_readId
for rIdx in U:
U[rIdx] = [U[rIdx][0][0], U[rIdx][1][0]] #keep gIdx and score only
for rIdx in NU:
pScoreSum = sum(NU[rIdx][1])
NU[rIdx][2] = [k / pScoreSum for k in NU[rIdx][1]] #Normalizing pScore
return U, NU, genomes, read
def conv_align2GRmat(aliDfile,pScoreCutoff,aliFormat):
in1 = open(aliDfile,'r')
U = {}
NU = {}
h_readId = {}
h_refId = {}
genomes = []
read =[]
gCnt = 0
rCnt = 0
maxScore = None
minScore = None
for ln in in1:
if (ln[0] == '@' or ln[0] == '#'):
continue
l = ln.split('\t')
readId=l[0]
if (aliFormat == 0 or aliFormat == 1): # gnu-sam or sam
if int(l[1])&0x4 == 4: # bitwise FLAG - 0x4 : segment unmapped
continue
refId=l[2]
elif (aliFormat == 2): # bl8
refId=l[1]
if refId == '*':
continue
#refId=refId.split("ti:")[-1]
mObj=re.search(r'ti\|(\d+)\|org\|([^|]+)\|gi',refId)
if mObj:
refId = "ti|"+mObj.group(1)+"|org|"+mObj.group(2)
else:
mObj=re.search(r'ti\|(\d+)\|gi',refId)
if mObj and mObj.group(1)!="-1":
refId = "ti|"+mObj.group(1)
(pScore, skipFlag) = find_entry_score(ln, l, aliFormat, pScoreCutoff)
if skipFlag:
continue
if ((maxScore == None) or (pScore > maxScore)):
maxScore = pScore
if ((minScore == None) or (pScore < minScore)):
minScore = pScore
gIdx = h_refId.get(refId,-1)
if gIdx == -1:
gIdx = gCnt
h_refId[refId] = gIdx
genomes.append(refId)
gCnt += 1
rIdx = h_readId.get(readId,-1)
if rIdx == -1:
#hold on this new read
#first, wrap previous read profile and see if any previous read has a same profile with that!
rIdx = rCnt
h_readId[readId] = rIdx
read.append(readId)
rCnt += 1
U[rIdx] = [[gIdx], [pScore], [float(pScore)], pScore]
else:
if (rIdx in U):
if gIdx in U[rIdx][0]:
continue
NU[rIdx] = U[rIdx]
del U[rIdx]
if gIdx in NU[rIdx][0]:
continue
NU[rIdx][0].append(gIdx)
NU[rIdx][1].append(pScore)
if pScore > NU[rIdx][3]:
NU[rIdx][3] = pScore
# length = len(NU[rIdx][1])
# NU[rIdx][2] = [1.0/length]*length
in1.close()
if (aliFormat == 1): # sam
(U, NU) = samUtils.rescale_samscore(U, NU, maxScore, minScore)
del h_refId, h_readId
for rIdx in U:
U[rIdx] = [U[rIdx][0][0], U[rIdx][1][0]] #keep gIdx and score only
for rIdx in NU:
pScoreSum = sum(NU[rIdx][1])
NU[rIdx][2] = [k/pScoreSum for k in NU[rIdx][1]] #Normalizing pScore
return U, NU, genomes, read