def trimMatchOverlapsInBoth(inpfile, outfile, trim_subtype):
gp = MyFile.myfile()
MatchRecord.sortInXorderAP(inpfile, gp)
# The following coalescing assumes perfect runs.
hp = MyFile.myfile()
coalesceMatches(gp, hp, ((trim_subtype == 'x') or (trim_subtype == 'u')))
gp = MyFile.myfile()
trimMatchOverlapsInX(hp, gp, trim_subtype)
hp = MyFile.myfile()
MatchRecord.sortInYorderAP(gp, hp)
trimMatchOverlapsInY(hp, outfile, trim_subtype)
return
def trimMatchOverlapsInBoth(inpfile,outfile,trim_subtype):
gp = MyFile.myfile()
MatchRecord.sortInXorderAP(inpfile,gp)
# The following coalescing assumes perfect runs.
hp = MyFile.myfile()
coalesceMatches( gp, hp, ((trim_subtype == 'x') or (trim_subtype == 'u')) )
gp = MyFile.myfile()
trimMatchOverlapsInX(hp,gp,trim_subtype)
hp = MyFile.myfile()
MatchRecord.sortInYorderAP(gp,hp)
trimMatchOverlapsInY(hp,outfile,trim_subtype)
return
def __init__( self, runName):
"You must supply a atac file called runName.atac."
self.runName = runName
self.comments = []
self.metacommands = []
self.globals = {}
self.tableformat = {}
self.tabledata = {}
self.matches = MyFile.myfile()
self.runs = MyFile.myfile()
fp = open(runName,"r")
for line in fp:
self.atac_file_parse_line(line)
def findCoverageIntervals( inpfile, outfile, processFirstAxis):
# The input file is an ATAC matches file.
# The output file is an ATAC coverage intervals file.
inpfile.seek(0)
outfile.seek(0)
t0 = time.time()
tmpfile3 = MyFile.myfile()
for line in inpfile:
if(line[0]=="M"):
fields = line.split()
if(fields[1]=="u" or fields[1]=="x"):
if(processFirstAxis):
axis = fields[4]
bgn = int(fields[5])
end = bgn+int(fields[6])
else:
axis = fields[8]
bgn = int(fields[9])
end = bgn+int(fields[10])
print >>tmpfile3, "E", axis,bgn,1
print >>tmpfile3, "E", axis,end,-1
tmpfile3.close()
tmpname = tempfile.mktemp()
cmd = "sort -T . -k 1,1 -k 2,2 -k 3n -k 4nr %s > %s" % (tmpfile3.name, tmpname)
print >>sys.stderr, cmd
iret = os.system(cmd); assert(iret==0)
print >>sys.stderr,"time elapsed is ", (time.time() - t0)
tmpfile4 = open(tmpname)
t0 = time.time()
findUniformCoverageIntervals( tmpfile4, outfile)
print >>sys.stderr,"time elapsed is ", (time.time() - t0)
tmpfile4.close()
os.system("rm -f " + tmpname)
outfile.seek(0)
def main( inpname, outname):
obj = AtacFile.AtacFile(inpname)
assemblyId1 = obj.globals['assemblyId1']
assemblyId2 = obj.globals['assemblyId2']
assemblyFilePrefix1 = obj.globals['assemblyFilePrefix1']
assemblyFilePrefix2 = obj.globals['assemblyFilePrefix2']
if(not obj.globals.has_key('fillIntraRunGapsErate')):
obj.globals['fillIntraRunGapsErate'] = 0.10
if(not obj.globals.has_key('fillIntraRunGapsMaxGap')):
obj.globals['fillIntraRunGapsMaxGap'] = 100000
fillIntraRunGapsErate = float(obj.globals['fillIntraRunGapsErate'])
fillIntraRunGapsMaxGap = int(obj.globals['fillIntraRunGapsMaxGap'])
# mismatches = checkExactMatches( x, y, inpfile)
# sys.stderr.write("mismatches = %d\n" % mismatches)
xIdx = IdxStore.IdxStore(assemblyFilePrefix1,assemblyId1)
yIdx = IdxStore.IdxStore(assemblyFilePrefix2,assemblyId2)
tempfile = MyFile.myfile()
mainLoop( obj.matches, tempfile, xIdx, yIdx,
fillIntraRunGapsMaxGap, fillIntraRunGapsErate)
obj.matches = tempfile
obj.checkpoint(outname)
def main(inpname, outname):
obj = AtacFile.AtacFile(inpname)
assemblyId1 = obj.globals['assemblyId1']
assemblyId2 = obj.globals['assemblyId2']
assemblyFilePrefix1 = obj.globals['assemblyFilePrefix1']
assemblyFilePrefix2 = obj.globals['assemblyFilePrefix2']
if (not obj.globals.has_key('fillIntraRunGapsErate')):
obj.globals['fillIntraRunGapsErate'] = 0.10
if (not obj.globals.has_key('fillIntraRunGapsMaxGap')):
obj.globals['fillIntraRunGapsMaxGap'] = 100000
fillIntraRunGapsErate = float(obj.globals['fillIntraRunGapsErate'])
fillIntraRunGapsMaxGap = int(obj.globals['fillIntraRunGapsMaxGap'])
# mismatches = checkExactMatches( x, y, inpfile)
# sys.stderr.write("mismatches = %d\n" % mismatches)
xIdx = IdxStore.IdxStore(assemblyFilePrefix1, assemblyId1)
yIdx = IdxStore.IdxStore(assemblyFilePrefix2, assemblyId2)
tempfile = MyFile.myfile()
mainLoop(obj.matches, tempfile, xIdx, yIdx, fillIntraRunGapsMaxGap,
fillIntraRunGapsErate)
obj.matches = tempfile
obj.checkpoint(outname)
def addMyFile(crtFile, nrFiles):
global nrNodes
nrFiles += 1
nrNodes += 1
Label[nrNodes] = (crtFile, 'File')
fileNodes.append(nrNodes)
fileDict[crtFile] = nrNodes
fileIssues[nrNodes] = {}
nrFileIssues[nrNodes] = 0
files.append(MyFile(nrFiles, crtFile, 0, 0, 0, 0))
return nrFiles
def boxRecovery( inpfile, rawfile, outname):
inpfile.seek(0)
rawfile.seek(0)
outfile = MyFile.myfile()
rawfileIter = iter(rawfile)
# This is a modified merge operation?
# The two input files must be sorted the same manner.
leftMatch = None
for line in inpfile:
if(line[0] == 'M'):
rightMatch = MatchRecord.MatchRecord(line)
if( leftMatch != None and leftMatch.inSameRunAs(rightMatch) ):
# print >>STDERR, "In same run leftMatch=", leftMatch, " rightMatch=", rightMatch
for rawline in rawfileIter:
if( rawline[0] == 'M'):
rawMatch = MatchRecord.MatchRecord(rawline)
if(rawMatch.sameAs(rightMatch)):
print >>outfile, rightMatch
break
else:
# print "Inside run rawMatch=", rawMatch
if(rawMatch.isInsideBox(leftMatch,rightMatch)):
print >>outfile, rawMatch
# end if
# end if
# end if
# end for
# We should die here if there is no rawMatch that matched the rightMatch ...
else:
# print >>STDERR, "Between runs leftMatch=", leftMatch, " rightMatch=", rightMatch
for rawline in rawfileIter:
if( rawline[0] == 'M'):
rawMatch = MatchRecord.MatchRecord(rawline)
if(rawMatch.sameAs(rightMatch)):
print >>outfile, rightMatch
break
else:
# print >>STDERR, "Discard rawMatch=", rawMatch
pass
# end if
# end if
# end for
# We should die here if there is no rawMatch that matched the rightMatch ...
# Discard raw Matches until it is ge to the right match.
# end if
leftMatch = rightMatch
# end if
# end for
return outfile
def formPerfectRuns ( inpfile, firstSort, secondSort, maxJump, runIdPrefix ):
inpfile.seek(0)
step = 0
print >>sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
tmpfile = MyFile.myfile()
firstSort( inpfile, tmpfile)
print >>sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
outfile = createSignedEnumeration(tmpfile)
print >>sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
tmpfile = MyFile.myfile()
secondSort( outfile, tmpfile)
print >>sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
outfile = findPerfectRuns( tmpfile, maxJump, runIdPrefix)
return outfile
def formPerfectRuns(inpfile, firstSort, secondSort, maxJump, runIdPrefix):
inpfile.seek(0)
step = 0
print >> sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
tmpfile = MyFile.myfile()
firstSort(inpfile, tmpfile)
print >> sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
outfile = createSignedEnumeration(tmpfile)
print >> sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
tmpfile = MyFile.myfile()
secondSort(outfile, tmpfile)
print >> sys.stderr, 'formPerfectRuns step=' + str(step)
step += 1
outfile = findPerfectRuns(tmpfile, maxJump, runIdPrefix)
return outfile
def onlyKeepLongRuns ( inpfile, outname, lengthThreshold ):
outfile = MyFile.myfile()
rejectsfile = MyFile.myfile()
FL = None
store = []
lenInMatches = 0
inpfile.seek(0)
for line in inpfile:
if(line[0] == 'M'):
FM = MatchRecord.MatchRecord(line)
SL = FM.x_length
if FL != None and FL.runid != FM.runid :
for x in store:
print >>rejectsfile, x
# end for
store = []
lenInMatches = SL
else:
lenInMatches += SL
# end if
if lenInMatches < lengthThreshold:
store.append(FM)
else:
for x in store:
print >>outfile, x
# end for
store = []
print >>outfile, FM
# end if
FL = FM
# end if
# end for
rejectsfile.close()
return outfile
def createSignedEnumeration(inpfile):
outfile = MyFile.myfile()
p = 1
inpfile.seek(0)
for line in inpfile:
if (line[0] == 'M'):
FM = MatchRecord.MatchRecord(line)
forwardX = FM.x_orientation
forwardY = FM.y_orientation
srank = cvm(forwardX == forwardY, p, -p)
p += 1
FM.extend['srank'] = srank
print >> outfile, FM
# end if
# end while
return outfile
def createSignedEnumeration(inpfile):
outfile = MyFile.myfile()
p = 1
inpfile.seek(0)
for line in inpfile:
if(line[0] == 'M'):
FM = MatchRecord.MatchRecord(line)
forwardX = FM.x_orientation
forwardY = FM.y_orientation
srank = cvm(forwardX == forwardY, p, -p)
p += 1
FM.extend['srank'] = srank
print >>outfile, FM
# end if
# end while
return outfile
def newmain():
inpname = sys.argv[1]
outname = sys.argv[2]
obj = AtacFile.AtacFile(inpname)
xname = obj.globals["assemblyFilePrefix1"]
yname = obj.globals["assemblyFilePrefix1"]
assemblyId1 = obj.globals["assemblyId1"]
assemblyId2 = obj.globals["assemblyId2"]
xIdx = IdxStore.IdxStore(xname, assemblyId1)
yIdx = IdxStore.IdxStore(yname, assemblyId2)
inpfile = obj.matches
outfile = MyFile.myfile()
main(inpfile, outfile, xIdx, yIdx)
obj.matches = outfile
obj.checkpoint(outname)
outfile.close()
def newmain():
inpname = sys.argv[1]
outname = sys.argv[2]
obj = AtacFile.AtacFile(inpname)
xname = obj.globals["assemblyFilePrefix1"]
yname = obj.globals["assemblyFilePrefix1"]
assemblyId1 = obj.globals["assemblyId1"]
assemblyId2 = obj.globals["assemblyId2"]
xIdx = IdxStore.IdxStore(xname,assemblyId1)
yIdx = IdxStore.IdxStore(yname,assemblyId2)
inpfile = obj.matches
outfile = MyFile.myfile()
main( inpfile, outfile, xIdx, yIdx)
obj.matches = outfile
obj.checkpoint(outname)
outfile.close()
def findCoverageIntervals(inpfile, outfile, processFirstAxis):
# The input file is an ATAC matches file.
# The output file is an ATAC coverage intervals file.
inpfile.seek(0)
outfile.seek(0)
t0 = time.time()
tmpfile3 = MyFile.myfile()
for line in inpfile:
if (line[0] == "M"):
fields = line.split()
if (fields[1] == "u" or fields[1] == "x"):
if (processFirstAxis):
axis = fields[4]
bgn = int(fields[5])
end = bgn + int(fields[6])
else:
axis = fields[8]
bgn = int(fields[9])
end = bgn + int(fields[10])
print >> tmpfile3, "E", axis, bgn, 1
print >> tmpfile3, "E", axis, end, -1
tmpfile3.close()
tmpname = tempfile.mktemp()
cmd = "sort -T . -k 1,1 -k 2,2 -k 3n -k 4nr %s > %s" % (tmpfile3.name,
tmpname)
print >> sys.stderr, cmd
iret = os.system(cmd)
assert (iret == 0)
print >> sys.stderr, "time elapsed is ", (time.time() - t0)
tmpfile4 = open(tmpname)
t0 = time.time()
findUniformCoverageIntervals(tmpfile4, outfile)
print >> sys.stderr, "time elapsed is ", (time.time() - t0)
tmpfile4.close()
os.system("rm -f " + tmpname)
outfile.seek(0)
def applyBothKeepMasks( inpfile, outfile ):
# Maybe we can think of a masking implementation where each ATAC match
# is treated atomicly. Assume that the keep mask intervals are sorted
# by start postition. Assume that the ATAC matches are sorted by start
# postion. Assert that all keep mask intervals are non-overlapping and
# were cut from only one ATAC match. Thus the mapping from keep mask
# intervals is a function. Note that this requires that we do not
# coalesce abutting keep mask intervals that originate from multiple
# matches. Note this still allows an ATAC match to overlap more than
# one keep mask interval. Ignore all keep mask intervals with zero
# length their creation has tie breaking problems. See notes on 2003
# Jul 29.
debug = 0
debugnum = 0
inpfile.seek(0)
outfile.seek(0)
# Apply the keepMask for the first axis.
# Make the sorted the keep mask intervals for the first axis.
processFirstAxis = 1
keepMaskFile = MyFile.myfile()
tmpfile2 = inpfile
tmpfile3 = MyFile.myfile()
tmpfile4 = MyFile.myfile()
findCoverageIntervals( inpfile, keepMaskFile, processFirstAxis)
if debug:
debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
for line in keepMaskFile: print >>debugfile, line,
MatchRecord.sortInXorderAP(tmpfile2,tmpfile3)
if debug:
#tmpfile2.seek(0)
#debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
#for line in tmpfile2: print >>debugfile, line,
tmpfile3.seek(0)
debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile3: print >>debugfile, line,
applyOneKeepMask( tmpfile3, tmpfile4, keepMaskFile, processFirstAxis)
if debug:
tmpfile4.seek(0)
debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile4: print >>debugfile, line,
# Apply the keepMask for the second axis.
# Make the sorted the keep mask intervals for the second axis.
processFirstAxis = 0
keepMaskFile = MyFile.myfile()
tmpfile2 = tmpfile4
tmpfile3 = MyFile.myfile()
tmpfile4 = outfile
findCoverageIntervals( inpfile, keepMaskFile, processFirstAxis)
if debug:
debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
for line in keepMaskFile: print >>debugfile, line,
MatchRecord.sortInYorderAP(tmpfile2,tmpfile3)
if debug:
#tmpfile2.seek(0)
#debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
#for line in tmpfile2: print >>debugfile, line,
tmpfile3.seek(0)
debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile3: print >>debugfile, line,
applyOneKeepMask( tmpfile3, tmpfile4, keepMaskFile, processFirstAxis)
if debug:
tmpfile4.seek(0)
debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile4: print >>debugfile, line,
def runOld(self):
self.globals['atacAlgorithmVersion'] = str(17)
print >>STDERR, "runName = %s\n" % self.runName
# The ATAC globals used by this script:
opt_t = int(self.globals['globalMatchMinSize'])
opt_l = int(self.globals['globalPerfectRunMinLen'])
maxdiff = int(self.globals['globalPerfectRunMaxGapLen'])
assemblyId1 = self.globals['assemblyId1']
assemblyId2 = self.globals['assemblyId2']
assemblyFile1 = self.globals['assemblyFile1']
assemblyFile2 = self.globals['assemblyFile2']
boxRecoveryOn = 0 # Deprecated for same species comparisons 2003/09/09.
if(self.globals.has_key("boxRecoveryOn")):
boxRecoveryOn = int(self.globals['boxRecoveryOn'])
t0 = time.time()
assemblyIdx1 = IdxStore.IdxStore(assemblyFile1,assemblyId1)
assemblyIdx2 = IdxStore.IdxStore(assemblyFile2,assemblyId2)
rawfile = None
###################################################################
# Setup for checkpointing scheme.
redo = 0
keep = 0
step = 0
if(self.globals.has_key("ckpKeep")):
keep = int(self.globals['ckpKeep'])
ckpName = "AllDone"
###################################################################
print >>STDERR, 'Keep step=' + str(keep)
print >>STDERR, 'At step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
outprefix = self.runName
step += 1
print >>STDERR, 'At uniqueFilter, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
if(not(self.globals.has_key('uniqueFilterOn') and self.globals['uniqueFilterOn']=="0")):
print >>STDERR, 'Running UniqueFilter'
outfile = MyFile.myfile()
UniqueFilter.main( self.matches, outfile)
self.matches = outfile
outprefix += '.uniq'
self.checkpoint(outprefix)
step += 1
print >>STDERR, 'At filterByMatchLength, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, 'Running filterByMatchLength'
outfile = MyFile.myfile()
filterByMatchLength( self.matches, outfile, opt_t)
self.matches = outfile
outprefix += '.t' + str(opt_t)
self.checkpoint(outprefix)
step += 1
print >>STDERR, 'At trimMatchOverlaps, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, "Start trimming for bp one-to-one-ness"
tempdata = MyFile.myfile()
TrimMatchOverlaps.trimMatchOverlapsInBoth(self.matches,tempdata,'u')
self.matches = tempdata
print >>STDERR, "Finished trimming for bp one-to-one-ness"
outprefix += '.trim'
self.checkpoint(outprefix)
if( boxRecoveryOn == 1 ):
# For box recovery later ... but what if we start from a checkpoint?
rawfile = self.matches
step += 1
print >>STDERR, 'At formPerfectRuns, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, 'from ' + outprefix + ' making ' + outprefix + '.p6'
tempdata = PerfectRuns.formPerfectRuns(self.matches,
MatchRecord.sortInXorderAP,
MatchRecord.sortInYorderAP,
maxdiff,
'r')
self.matches = tempdata
outprefix += ".p6"
# end if
step += 1
print >>STDERR, 'At onlyKeepLongRuns, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, 'from ' + outprefix + ' making ' + outprefix + '.l' + str(opt_l)
tempdata = onlyKeepLongRuns( self.matches, outprefix, opt_l)
self.matches = tempdata
outprefix += '.l' + str(opt_l)
self.checkpoint(outprefix)
step += 1
print >>STDERR, 'At formPerfectRuns, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, 'Heal the perfect runs'
tempdata = PerfectRuns.formPerfectRuns(self.matches,
MatchRecord.sortInYorderAP,
MatchRecord.sortInXorderAP, maxdiff, 'r')
self.matches = tempdata
outprefix += '.pr'
self.checkpoint(outprefix)
if(boxRecoveryOn == 1):
# This is a box recovery step.
step += 1
print >>STDERR, 'At boxRecovery, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, 'from ' + outprefix + ' making ' + outprefix + '.br'
print >>STDERR, "Make sorted raw matches"
outfile = MyFile.myfile()
MatchRecord.sortInXorderAP( rawfile, outfile)
rawfile = outfile
print >>STDERR, "perform box recovery"
tempdata = boxRecovery( self.matches, rawfile, outprefix)
self.matches = tempdata
outprefix += '.br'
self.checkpoint(outprefix)
# end if
step += 1
print >>STDERR, 'At formPerfectRuns, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ( (keep < step) and not self.globals.has_key(ckpName))):
print >>STDERR, "form perfect runs"
redo = 1
print >>STDERR, 'from ' + outprefix + ' to ' + outprefix + '.p6'
tempdata = PerfectRuns.formPerfectRuns(self.matches,
MatchRecord.sortInXorderAP,
MatchRecord.sortInYorderAP, maxdiff, 'r')
self.matches = tempdata
outprefix += '.pr'
self.checkpoint(outprefix)
step += 1
print >>STDERR, 'At squeezeIntraRunGaps, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, 'from ' + outprefix + ' to ' + outprefix + '.sq'
tempdata = MyFile.myfile()
squeezeIntraRunGaps.mainLoop(
self.matches,
tempdata,
assemblyIdx1, assemblyIdx2)
tempy = MyFile.myfile()
# Beware the current match subtypes are 'x', 'L', and 'R'!
coalesceMatches( tempdata, tempy, 1)
self.matches = tempy
outprefix += '.sq'
self.checkpoint(outprefix)
step += 1
print >>STDERR, 'At TrimMatchOverlaps, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, "Start trimming for bp one-to-one-ness"
tempdata = MyFile.myfile()
TrimMatchOverlaps.trimMatchOverlapsInBoth(self.matches,tempdata,'u')
self.matches = tempdata
outprefix += '.trim'
print >>STDERR, "Finished trimming for bp one-to-one-ness"
step += 1
print >>STDERR, 'At RunsAsMatches, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
self.runs = PerfectRuns.runsAsMatches( self.matches)
outprefix += '.runs'
self.checkpoint(outprefix)
# end if
if(self.globals.has_key('fillIntraRunGapsOn') and self.globals['fillIntraRunGapsOn']=="1" ):
# Next comes the DNA sequence dependent stuff.
step += 1
print >>STDERR, 'At fillIntraRunGaps, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, "fill the intrarun gaps"
if(not self.globals.has_key('fillIntraRunGapsErate')):
self.globals['fillIntraRunGapsErate'] = 0.10
if(not self.globals.has_key('fillIntraRunGapsMaxGap')):
self.globals['fillIntraRunGapsMaxGap'] = 100000
fillIntraRunGapsErate = float(self.globals['fillIntraRunGapsErate'])
fillIntraRunGapsMaxGap = int(self.globals['fillIntraRunGapsMaxGap'])
tempdata = MyFile.myfile()
fillIntraRunGaps.mainLoop(self.matches, tempdata,
assemblyIdx1, assemblyIdx2,
fillIntraRunGapsMaxGap, fillIntraRunGapsErate)
self.matches = tempdata
outprefix += '.fill'
self.checkpoint(outprefix)
step += 1
print >>STDERR, 'At TrimMatchOverlaps, step=' + str(step)
print >>STDERR, 'Time elapsed=' + str(time.time()-t0)
if (redo or ((keep < step) and not self.globals.has_key(ckpName))):
redo = 1
print >>STDERR, "trim the overlaps"
tempdata = MyFile.myfile()
TrimMatchOverlaps.trimMatchOverlapsInBoth(self.matches,tempdata,'u')
self.matches = tempdata
outprefix += '.trim'
self.checkpoint(outprefix)
def runsAsMatches(inpfile):
outfile = MyFile.myfile()
lastF = None
firstF = None
runFill = 0
inpfile.seek(0)
for line in inpfile:
if (line[0] == 'M'):
curF = MatchRecord.MatchRecord(line)
if ((lastF == None) or (curF.runid != lastF.runid)):
if ((lastF != None)
and (firstF.x_scaf_uid != lastF.x_scaf_uid)):
print >> sys.stderr, firstF
print >> sys.stderr, lastF
# end if
assert ((lastF == None)
or (firstF.x_scaf_uid == lastF.x_scaf_uid))
assert ((lastF == None)
or (firstF.y_scaf_uid == lastF.y_scaf_uid))
if (None != lastF):
x1 = firstF.x_start
x2 = lastF.x_start
startX = cvm(x1 < x2, x1, x2)
x1 += firstF.x_length
x2 += lastF.x_length
endX = cvm(x1 > x2, x1, x2)
y1 = firstF.y_start
y2 = lastF.y_start
startY = cvm(y1 < y2, y1, y2)
y1 += firstF.y_length
y2 += lastF.y_length
endY = cvm(y1 > y2, y1, y2)
lastF.subtype = 'r'
lastF.matchid = lastF.runid
lastF.runid = "." # the agreed NULL value
lastF.x_start = startX
lastF.y_start = startY
lastF.x_length = endX - startX
lastF.y_length = endY - startY
lastF.runFill = runFill
print >> outfile, lastF
# end if
firstF = curF
runFill = 0
# end if
runFill += curF.x_length
lastF = curF
# end if
# end for
if (None != lastF):
x1 = firstF.x_start
x2 = lastF.x_start
startX = cvm(x1 < x2, x1, x2)
x1 += firstF.x_length
x2 += lastF.x_length
endX = cvm(x1 > x2, x1, x2)
y1 = firstF.y_start
y2 = lastF.y_start
startY = cvm(y1 < y2, y1, y2)
y1 += firstF.y_length
y2 += lastF.y_length
endY = cvm(y1 > y2, y1, y2)
lastF.subtype = 'r'
lastF.matchid = lastF.runid
lastF.runid = "." # the agreed NULL value
lastF.x_start = startX
lastF.y_start = startY
lastF.x_length = endX - startX
lastF.y_length = endY - startY
lastF.runFill = runFill
print >> outfile, lastF
# end if
return outfile
def applyBothKeepMasks(inpfile, outfile):
# Maybe we can think of a masking implementation where each ATAC match
# is treated atomicly. Assume that the keep mask intervals are sorted
# by start postition. Assume that the ATAC matches are sorted by start
# postion. Assert that all keep mask intervals are non-overlapping and
# were cut from only one ATAC match. Thus the mapping from keep mask
# intervals is a function. Note that this requires that we do not
# coalesce abutting keep mask intervals that originate from multiple
# matches. Note this still allows an ATAC match to overlap more than
# one keep mask interval. Ignore all keep mask intervals with zero
# length their creation has tie breaking problems. See notes on 2003
# Jul 29.
debug = 0
debugnum = 0
inpfile.seek(0)
outfile.seek(0)
# Apply the keepMask for the first axis.
# Make the sorted the keep mask intervals for the first axis.
processFirstAxis = 1
keepMaskFile = MyFile.myfile()
tmpfile2 = inpfile
tmpfile3 = MyFile.myfile()
tmpfile4 = MyFile.myfile()
findCoverageIntervals(inpfile, keepMaskFile, processFirstAxis)
if debug:
debugnum += 1
debugfile = open("debugfile.%d" % debugnum, "w")
for line in keepMaskFile:
print >> debugfile, line,
MatchRecord.sortInXorderAP(tmpfile2, tmpfile3)
if debug:
#tmpfile2.seek(0)
#debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
#for line in tmpfile2: print >>debugfile, line,
tmpfile3.seek(0)
debugnum += 1
debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile3:
print >> debugfile, line,
applyOneKeepMask(tmpfile3, tmpfile4, keepMaskFile, processFirstAxis)
if debug:
tmpfile4.seek(0)
debugnum += 1
debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile4:
print >> debugfile, line,
# Apply the keepMask for the second axis.
# Make the sorted the keep mask intervals for the second axis.
processFirstAxis = 0
keepMaskFile = MyFile.myfile()
tmpfile2 = tmpfile4
tmpfile3 = MyFile.myfile()
tmpfile4 = outfile
findCoverageIntervals(inpfile, keepMaskFile, processFirstAxis)
if debug:
debugnum += 1
debugfile = open("debugfile.%d" % debugnum, "w")
for line in keepMaskFile:
print >> debugfile, line,
MatchRecord.sortInYorderAP(tmpfile2, tmpfile3)
if debug:
#tmpfile2.seek(0)
#debugnum += 1; debugfile = open("debugfile.%d" % debugnum, "w")
#for line in tmpfile2: print >>debugfile, line,
tmpfile3.seek(0)
debugnum += 1
debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile3:
print >> debugfile, line,
applyOneKeepMask(tmpfile3, tmpfile4, keepMaskFile, processFirstAxis)
if debug:
tmpfile4.seek(0)
debugnum += 1
debugfile = open("debugfile.%d" % debugnum, "w")
for line in tmpfile4:
print >> debugfile, line,
def findPerfectRuns(inpfile, maxJump, runIdPrefix):
outfile = MyFile.myfile()
left = None
runid = 1
inpfile.seek(0)
for line in inpfile:
if (line[0] == 'M'):
right = MatchRecord.MatchRecord(line)
pr = int(right.extend['srank'])
del (right.extend['srank'])
if (left != None):
maxGapInXandY = 0
if (left.x_scaf_uid == right.x_scaf_uid
and left.y_scaf_uid == right.y_scaf_uid):
# Find the maximum of the gap in x and y axis.
x_rs = right.x_start
x_re = x_rs + right.x_length
x_ls = left.x_start
x_le = x_ls + left.x_length
assert (x_rs < x_re)
assert (x_ls < x_le)
# All matches are positive length.
x_gapLeftBeforeRight = x_rs - x_le
x_gapRightBeforeLeft = x_ls - x_re
assert (not (x_gapLeftBeforeRight > 0
and x_gapRightBeforeLeft > 0))
x_gap = max(x_gapLeftBeforeRight, x_gapRightBeforeLeft)
# x_gap == 0 is abutting
# x_gap < 0 is overlapping
y_rs = right.y_start
y_re = y_rs + right.y_length
y_ls = left.y_start
y_le = y_ls + left.y_length
assert (y_rs < y_re)
assert (y_ls < y_le)
y_gapLeftBeforeRight = y_rs - y_le
y_gapRightBeforeLeft = y_ls - y_re
assert (not (y_gapLeftBeforeRight > 0
and y_gapRightBeforeLeft > 0))
y_gap = max(y_gapLeftBeforeRight, y_gapRightBeforeLeft)
# y_gap == 0 is abutting
# y_gap < 0 is overlapping
maxGapInXandY = max(x_gap, y_gap)
if 1:
# Check the sorting of the matches.
sorted_by_x = (x_ls <= x_rs)
sorted_by_y = (y_ls <= y_rs)
if (not (sorted_by_x or sorted_by_y)):
print >> sys.stderr, "bad sorting in findPerfectRuns"
print >> sys.stderr, left
print >> sys.stderr, right
assert (sorted_by_x or sorted_by_y)
dovetail_in_x = (x_ls <= x_rs) and (x_le <= x_re)
dovetail_in_y = (y_ls <= y_rs) and (y_ls <= y_re)
if (sorted_by_x and not (dovetail_in_x)):
print >> sys.stderr, "contained in x in findPerfectRuns"
print >> sys.stderr, left
print >> sys.stderr, right
if (sorted_by_y and not (dovetail_in_y)):
print >> sys.stderr, "contained in y in findPerfectRuns"
print >> sys.stderr, left
print >> sys.stderr, right
# endif
if ((left.x_scaf_uid != right.x_scaf_uid)
or # check first axis id
(left.y_scaf_uid != right.y_scaf_uid)
or # check second axis id
(maxGapInXandY > maxJump)
or (pr != lastpr + 1
) # Using the signed rank NOT the run id !!!!
):
runid += 1
# end if
# end if
lastpr = pr
right.runid = "%s%d" % (
runIdPrefix,
runid,
) # Assign the run id in the same slot as the signed rank.
print >> outfile, right
left = right
# end if
# end for
return outfile
def runsAsMatches(inpfile):
outfile = MyFile.myfile()
lastF = None
firstF = None
runFill = 0
inpfile.seek(0)
for line in inpfile:
if(line[0] == 'M'):
curF = MatchRecord.MatchRecord(line)
if ((lastF == None) or (curF.runid != lastF.runid)):
if ((lastF != None) and (firstF.x_scaf_uid != lastF.x_scaf_uid)):
print >>sys.stderr, firstF
print >>sys.stderr, lastF
# end if
assert((lastF==None) or (firstF.x_scaf_uid == lastF.x_scaf_uid))
assert((lastF==None) or (firstF.y_scaf_uid == lastF.y_scaf_uid))
if (None != lastF):
x1 = firstF.x_start
x2 = lastF.x_start
startX = cvm(x1 < x2, x1, x2)
x1 += firstF.x_length
x2 += lastF.x_length
endX = cvm(x1 > x2, x1, x2)
y1 = firstF.y_start
y2 = lastF.y_start
startY = cvm( y1 < y2, y1, y2)
y1 += firstF.y_length
y2 += lastF.y_length
endY = cvm(y1 > y2, y1, y2)
lastF.subtype = 'r'
lastF.matchid = lastF.runid
lastF.runid = "." # the agreed NULL value
lastF.x_start = startX
lastF.y_start = startY
lastF.x_length = endX - startX
lastF.y_length = endY - startY
lastF.runFill = runFill
print >>outfile, lastF
# end if
firstF = curF
runFill = 0
# end if
runFill += curF.x_length
lastF = curF
# end if
# end for
if (None != lastF):
x1 = firstF.x_start
x2 = lastF.x_start
startX = cvm( x1 < x2, x1, x2)
x1 += firstF.x_length
x2 += lastF.x_length
endX = cvm( x1 > x2, x1, x2)
y1 = firstF.y_start
y2 = lastF.y_start
startY = cvm( y1 < y2, y1, y2)
y1 += firstF.y_length
y2 += lastF.y_length
endY = cvm( y1 > y2, y1, y2)
lastF.subtype = 'r'
lastF.matchid = lastF.runid
lastF.runid = "." # the agreed NULL value
lastF.x_start = startX
lastF.y_start = startY
lastF.x_length = endX - startX
lastF.y_length = endY - startY
lastF.runFill = runFill
print >>outfile, lastF
# end if
return outfile
def findPerfectRuns ( inpfile, maxJump, runIdPrefix ):
outfile = MyFile.myfile()
left = None
runid = 1
inpfile.seek(0)
for line in inpfile:
if(line[0] == 'M'):
right = MatchRecord.MatchRecord(line)
pr = int(right.extend['srank'])
del(right.extend['srank'])
if(left != None):
maxGapInXandY = 0
if(left.x_scaf_uid == right.x_scaf_uid
and
left.y_scaf_uid == right.y_scaf_uid ):
# Find the maximum of the gap in x and y axis.
x_rs = right.x_start
x_re = x_rs + right.x_length
x_ls = left.x_start
x_le = x_ls + left.x_length
assert(x_rs < x_re)
assert(x_ls < x_le)
# All matches are positive length.
x_gapLeftBeforeRight = x_rs - x_le
x_gapRightBeforeLeft = x_ls - x_re
assert(not(x_gapLeftBeforeRight>0 and x_gapRightBeforeLeft>0))
x_gap = max(x_gapLeftBeforeRight,x_gapRightBeforeLeft)
# x_gap == 0 is abutting
# x_gap < 0 is overlapping
y_rs = right.y_start
y_re = y_rs + right.y_length
y_ls = left.y_start
y_le = y_ls + left.y_length
assert(y_rs < y_re)
assert(y_ls < y_le)
y_gapLeftBeforeRight = y_rs - y_le
y_gapRightBeforeLeft = y_ls - y_re
assert(not(y_gapLeftBeforeRight>0 and y_gapRightBeforeLeft>0))
y_gap = max(y_gapLeftBeforeRight,y_gapRightBeforeLeft)
# y_gap == 0 is abutting
# y_gap < 0 is overlapping
maxGapInXandY = max(x_gap,y_gap)
if 1:
# Check the sorting of the matches.
sorted_by_x = (x_ls <= x_rs)
sorted_by_y = (y_ls <= y_rs)
if(not(sorted_by_x or sorted_by_y)):
print >>sys.stderr, "bad sorting in findPerfectRuns"
print >>sys.stderr, left
print >>sys.stderr, right
assert(sorted_by_x or sorted_by_y)
dovetail_in_x = (x_ls <= x_rs) and (x_le <= x_re)
dovetail_in_y = (y_ls <= y_rs) and (y_ls <= y_re)
if(sorted_by_x and not(dovetail_in_x)):
print >>sys.stderr, "contained in x in findPerfectRuns"
print >>sys.stderr, left
print >>sys.stderr, right
if(sorted_by_y and not(dovetail_in_y)):
print >>sys.stderr, "contained in y in findPerfectRuns"
print >>sys.stderr, left
print >>sys.stderr, right
# endif
if(
(left.x_scaf_uid != right.x_scaf_uid) or # check first axis id
(left.y_scaf_uid != right.y_scaf_uid) or # check second axis id
(maxGapInXandY > maxJump) or
(pr != lastpr + 1) # Using the signed rank NOT the run id !!!!
):
runid += 1
# end if
# end if
lastpr = pr
right.runid = "%s%d" % (runIdPrefix,runid,) # Assign the run id in the same slot as the signed rank.
print >>outfile, right
left = right
# end if
# end for
return outfile
