def adjListToRepeatList(newAdjacencyList, folderName, repeatFilename):
N1 = len(newAdjacencyList)
G2 = abunGraphLib.seqGraphWt(N1 * 2)
for i in range(N1):
for j in newAdjacencyList[i]:
G2.insertEdge(2 * i, 2 * j + 1, 1)
G2.insertEdge(2 * j + 1, 2 * i, 1)
clusters = G2.findConnectedComponents()
repeatList = []
for eachitem in clusters:
leftList, rightList = [], []
for eachsubitem in eachitem:
if eachsubitem % 2 == 0:
leftList.append(eachsubitem)
else:
rightList.append(eachsubitem)
repeatList.append([
abunHouseKeeper.getDistinct(leftList),
abunHouseKeeper.getDistinct(rightList)
])
with open(folderName + repeatFilename, 'w') as outfile:
json.dump(repeatList, outfile)
json_data = open(folderName + repeatFilename, 'r')
loadData = json.load(json_data)
assert (loadData == repeatList)
def adjListToRepeatList(newAdjacencyList,folderName,repeatFilename):
N1 = len(newAdjacencyList)
G2 = abunGraphLib.seqGraphWt(N1 * 2)
for i in range(N1):
for j in newAdjacencyList[i]:
G2.insertEdge(2 * i, 2 * j + 1, 1)
G2.insertEdge(2 * j + 1, 2 * i, 1)
clusters = G2.findConnectedComponents()
repeatList = []
for eachitem in clusters:
leftList, rightList = [], []
for eachsubitem in eachitem:
if eachsubitem % 2 == 0 :
leftList.append(eachsubitem)
else:
rightList.append(eachsubitem)
repeatList.append([abunHouseKeeper.getDistinct(leftList), abunHouseKeeper.getDistinct(rightList)])
with open(folderName + repeatFilename, 'w') as outfile:
json.dump(repeatList, outfile)
json_data = open(folderName + repeatFilename, 'r')
loadData = json.load(json_data)
assert(loadData == repeatList)
def identifyRepeat(folderName, mummerLink,contigFilename,contigReadGraph, repeatFilename, optionToRun ):
'''
Input : Graph --- phaseStringGraph1
Output: repeat pairs { [ (1,2), (3,4) ] , [(5,6),(7,8)] }
Algorithm:
a) Reachability test on the graph to find the partners
b) Form Bipartite graph
c) Find connected component in the bipartite and define as repeat pairs
'''
# ## (a) reachability test to find partners
G = graphLib.seqGraph(0)
G.loadFromFile(folderName, contigReadGraph)
# G.reportEdge()
lenDicCC = IORobot.obtainLength(folderName, contigFilename+"_Double.fasta")
adjacencyList = [[] for i in range(len(lenDicCC))]
N1 = len(lenDicCC)
# # Debug
# for i in range(14):
# debugGraphPath(i, 2, G, N1)
# # End Debug
for i in range(len(lenDicCC)):
adjacencyList[i] = abunGraphLib.findAllReachable(i, N1, G)
print "i, adjacencyList[i] : ", i , adjacencyList[i]
# ## (b) formation of bipartite graph
if optionToRun == "tandem" :
newAdjacencyList = adjacencyList
elif optionToRun == "xphase":
newAdjacencyList = abunGraphLib.filterEdge(adjacencyList, folderName, contigFilename)
G2 = abunGraphLib.seqGraphWt(N1 * 2)
for i in range(N1):
for j in newAdjacencyList[i]:
G2.insertEdge(2 * i, 2 * j + 1, 1)
G2.insertEdge(2 * j + 1, 2 * i, 1)
clusters = G2.findConnectedComponents()
repeatList = []
for eachitem in clusters:
leftList, rightList = [], []
for eachsubitem in eachitem:
if eachsubitem % 2 == 0 :
leftList.append(eachsubitem)
else:
rightList.append(eachsubitem)
repeatList.append([abunHouseKeeper.getDistinct(leftList), abunHouseKeeper.getDistinct(rightList)])
with open(folderName + repeatFilename, 'w') as outfile:
json.dump(repeatList, outfile)
json_data = open(folderName + repeatFilename, 'r')
loadData = json.load(json_data)
assert(loadData == repeatList)
def identifyRepeat(folderName, mummerLink, contigFilename, contigReadGraph,
repeatFilename, optionToRun):
'''
Input : Graph --- phaseStringGraph1
Output: repeat pairs { [ (1,2), (3,4) ] , [(5,6),(7,8)] }
Algorithm:
a) Reachability test on the graph to find the partners
b) Form Bipartite graph
c) Find connected component in the bipartite and define as repeat pairs
'''
# ## (a) reachability test to find partners
G = graphLib.seqGraph(0)
G.loadFromFile(folderName, contigReadGraph)
# G.reportEdge()
lenDicCC = IORobot.obtainLength(folderName,
contigFilename + "_Double.fasta")
adjacencyList = [[] for i in range(len(lenDicCC))]
N1 = len(lenDicCC)
# # Debug
# for i in range(14):
# debugGraphPath(i, 2, G, N1)
# # End Debug
for i in range(len(lenDicCC)):
adjacencyList[i] = abunGraphLib.findAllReachable(i, N1, G)
print "i, adjacencyList[i] : ", i, adjacencyList[i]
# ## (b) formation of bipartite graph
if optionToRun == "tandem":
newAdjacencyList = adjacencyList
elif optionToRun == "xphase":
newAdjacencyList = abunGraphLib.filterEdge(adjacencyList, folderName,
contigFilename)
G2 = abunGraphLib.seqGraphWt(N1 * 2)
for i in range(N1):
for j in newAdjacencyList[i]:
G2.insertEdge(2 * i, 2 * j + 1, 1)
G2.insertEdge(2 * j + 1, 2 * i, 1)
clusters = G2.findConnectedComponents()
repeatList = []
for eachitem in clusters:
leftList, rightList = [], []
for eachsubitem in eachitem:
if eachsubitem % 2 == 0:
leftList.append(eachsubitem)
else:
rightList.append(eachsubitem)
repeatList.append([
abunHouseKeeper.getDistinct(leftList),
abunHouseKeeper.getDistinct(rightList)
])
with open(folderName + repeatFilename, 'w') as outfile:
json.dump(repeatList, outfile)
json_data = open(folderName + repeatFilename, 'r')
loadData = json.load(json_data)
assert (loadData == repeatList)
def defineRepeatAndFlanking(folderName, mummerLink,contigFilename,contigReadGraph,repeatFilename,repeatSpec ):
'''
Input :
V a) String graph : G
V b) Repeat Pairing : repeatList
Output :
V a) chain of repeat indices (e.g. [S= R1, R33, R45, R24= E])
V b) chain of flanking region indices for in1/2 out1/2 middle (e.g. [C1, R2, R4] )
V c) in1/2 out1/2 and middle reads per repeat (e.g. [R1, R33, R45, R24])
Algorithm :
V 1. Find repeat by graph operations
V 2. Find flanking region by graph operations
V 3. Find associated reads by graph operations
'''
print "defineRepeatAndFlanking: "
# 0. Load previous data
G = abunGraphLib.seqGraphWt(0)
G.loadFromFile(folderName, contigReadGraph)
Grev = abunGraphLib.formReverseGraph(G)
json_data = open(folderName + repeatFilename, 'r')
repeatList = json.load(json_data)
lenDicCC = IORobot.obtainLength(folderName, contigFilename+"_Double.fasta")
N1 = len(lenDicCC)
print "repeatList: ", repeatList
print "len(G.graphNodesList)", len(G.graphNodesList)
bigDumpList = []
print "len(repeatList)", len(repeatList) , repeatList
for r in repeatList:
rIn, rOut = [], []
for eachitem in r[0]:
rIn.append(eachitem / 2)
for eachitem in r[1]:
rOut.append(eachitem / 2)
if ( len(rIn) == 2 and len(rOut) == 2) or (len(rIn) == 1 and len(rOut) == 1):
print rIn, rOut
if (len(rIn) == 1 and len(rOut) == 1):
rIn = [rIn[0], rIn[0]]
rOut = [rOut[0], rOut[0]]
# 1. Records reachable indices
kkIn , kkOut = [], []
for eachkk in rIn:
kkIn.append(str(eachkk)+"_"+"in")
for eachkk in rOut:
kkOut.append(str(eachkk)+"_"+"out")
abunGraphLib.markReachableIndices(G, Grev, kkIn, kkOut, N1)
# 2. Marks inside nodes
singleMissList, allPassList = abunGraphLib.markInsideNodes(G, kkIn, kkOut)
for i in range(4):
print "len(singleMissList[i]), len(allPassList)", len(singleMissList[i]), len(allPassList)
# 3. Finds start/end of repeat
myStartIndex, myEndIndex = abunGraphLib.markStartEndNodes(G, rIn, rOut, singleMissList, allPassList)
print myStartIndex, myEndIndex
# 4. Find repeat interior by shortest path joining S/E
repeatPathway = abunGraphLib.markInterior(G , myStartIndex, myEndIndex, N1)
print "repeatPathway", repeatPathway
#checkPathLength(repeatPathway, G, N1, folderName)
# 5. Find flanking region by shortest path search again
flankingPathsList = abunGraphLib.markFlankingRegion(G, rIn, rOut, myStartIndex, myEndIndex, N1)
print flankingPathsList
# 6. Find associated reads by graph node query
flankingList, repeatList = abunGraphLib.markAssociatedReads(G, singleMissList, allPassList)
# ## Experimental
repeatList = allPassList
# ## End Experimental
for eachlist in flankingList:
print len(eachlist), len(repeatList)
bigDumpList.append([flankingList, repeatList, repeatPathway, flankingPathsList])
# 7. Format return and move on to the phasing
with open(folderName + repeatSpec, 'w') as outfile:
json.dump(bigDumpList, outfile)
def BResolution(Gnew, folderName, contigReadGraph, N1, myCountDic, lenDic,
mummerLink):
if abunHouseKeeper.abunGlobalSplitParameterRobot.runBResolve:
print "abunHouseKeeper.abunGlobalSplitParameterRobot.runBResolve", abunHouseKeeper.abunGlobalSplitParameterRobot.runBResolve
maxRThres = abunHouseKeeper.abunGlobalSplitParameterRobot.RThres
repeatFinder.adjListToRepeatList(Gnew.adj, folderName,
"phaseRepeatTR.txt")
json_data = open(folderName + "phaseRepeatTR.txt", 'r')
repeatPairs = json.load(json_data)
repeatPairs = obtainNonEmpty(repeatPairs)
biResolvedCombineList = []
G = abunGraphLib.seqGraphWt(0)
G.loadFromFile(folderName, contigReadGraph)
Grev = abunGraphLib.formReverseGraphFast(G)
abunAnalysisList = []
for eachitem in repeatPairs:
inList, outList = eachitem[0], eachitem[1]
if not abunHouseKeeper.abunGlobalRunEM:
resolvedList, brResolvedList = [], []
if abunHouseKeeper.abunGlobalSplitParameterRobot.toRunAbunB:
if abunHouseKeeper.abunGlobalSplitParameterRobot.AbunLowerB > 0:
abunHouseKeeper.abunGlobalSplitParameterRobot.AbunLower = abunHouseKeeper.abunGlobalSplitParameterRobot.AbunLowerB
if abunHouseKeeper.abunGlobalSplitParameterRobot.AbunUpperB > 0:
abunHouseKeeper.abunGlobalSplitParameterRobot.AbunUpper = abunHouseKeeper.abunGlobalSplitParameterRobot.AbunUpperB
if not abunHouseKeeper.abunGlobalSplitParameterRobot.toRunAggB:
resolvedList = determindMatch(inList, outList,
myCountDic, folderName,
contigReadGraph, N1)
else:
resolvedList = determindMatchAggregate(
inList, outList, myCountDic, folderName,
contigReadGraph, N1, Gnew, lenDic)
if abunHouseKeeper.abunGlobalSplitParameterRobot.toRunBRB:
if abunHouseKeeper.abunGlobalSplitParameterRobot.BRThresB > 0:
abunHouseKeeper.abunGlobalSplitParameterRobot.BRThres = abunHouseKeeper.abunGlobalSplitParameterRobot.BRThresB
brResolvedList = formBRReolve(folderName, inList, outList,
G, Grev, True, N1)
combinedList = abunHouseKeeper.getDistinct(resolvedList +
brResolvedList)
print "resolvedList, brResolvedList, inList, outList", resolvedList, brResolvedList, inList, outList
print "resolveConflict(combinedList)", resolveConflict(
combinedList)
abunAnalysisList.append([
inList, outList, resolvedList, brResolvedList,
resolveConflict(combinedList)
])
if len(inList) <= maxRThres and len(
outList) <= maxRThres and len(inList) > 0 and len(
outList) > 0:
resolvedCombine = resolveConflict(combinedList)
Gnew.bipartiteLocalResolve(resolvedCombine, inList,
outList, folderName)
else:
import emalgo
resolvedCombine = emalgo.BResolvePreparation(
folderName, inList, outList, G, Grev, N1, mummerLink)
Gnew.bipartiteLocalResolve(resolvedCombine, inList, outList,
folderName)
Gnew.condense()
with open(folderName + "biResolvedCombineList.json", 'w') as f:
json.dump(biResolvedCombineList, f)
with open(folderName + "abunAnalysisList.json", 'w') as f:
json.dump(abunAnalysisList, f)
#assert(1==2)
return Gnew
else:
return Gnew