def __init__(self, seq, id, extension_handler, rc = None):
# type: (str, str, ExtensionHandler, Optional[NewLine]) -> None
self.extensionHandler = extension_handler
self.seq = seq
self.id = id # type: str
self.circular = False
self.name_printer = None
self.max_extension = False
if rc is None:
self.initial = AlignmentStorage()
self.correct_segments = SegmentStorage()
self.completely_resolved = SegmentStorage()
self.disjointig_alignments = AlignmentStorage()
self.read_alignments = ReadAlignmentStorage()
self.listeners = [self.initial, self.correct_segments, self.completely_resolved, self.disjointig_alignments, self.read_alignments, extension_handler] # type: List[LineListener]
rc = NewLine(basic.RC(seq), basic.Reverse(self.id), extension_handler.rc, self) #type: NewLine
self.rc = rc
self.addListener(ReadAlignmentListener(self))
# self.initial.add(AlignmentPiece.Identical(self.asSegment().asContig().asSegment(), self.asSegment()))
else:
self.initial = rc.initial.rc # type: AlignmentStorage
self.correct_segments = rc.correct_segments.rc # type: SegmentStorage
self.completely_resolved = rc.completely_resolved.rc # type: SegmentStorage
self.disjointig_alignments = rc.disjointig_alignments.rc # type: AlignmentStorage
self.read_alignments = rc.read_alignments.rc # type: ReadAlignmentStorage
self.listeners = [listener.rc for listener in rc.listeners] # type: List[LineListener]
Contig.__init__(self, seq, id, rc)
self.rc = rc #type: NewLine
self.knot = None # type: Knot
def splitBad(self, line, lines):
# type: (NewLine, NewLineStorage) -> None
s = AlignmentStorage()
s.addAll(al for al in line.read_alignments if not al.contradictingRTC())
segs = list(s.filterByCoverage(mi=params.reliable_coverage, k=params.k)) # type: List[Segment]
segs = filter(lambda seg: len(seg) >= params.k, segs)
if len(segs) == 0:
sys.stdout.warn("No part of a unique edge is covered by reads", line.id)
lines.removeLine(line)
return
if len(segs) == 1 and len(segs[0]) > len(line) - 10:
sys.stdout.info("Whole line", line.id, "is covered by reads")
return
sys.stdout.info( "Line", line.id, "has poorly covered regions. Splitting into", len(segs), "parts")
sys.stdout.trace( segs)
next_left = segs[-1].left
line.cutRight(segs[-1].right)
for seg in list(segs)[-2::-1]:
if next_left < seg.right:
line, new_line = lines.splitLine(line.segment(next_left, seg.right))
else:
line, new_line = lines.splitLine(line.segment(next_left, next_left))
line.cutRight(seg.right)
next_left = seg.left
line.rc.cutRight(len(segs[0]))
def splitBad(self, lines):
# type: (NewLineStorage) -> None
all_covs = []
for line in lines:
for rec in line.read_alignments.calculateCoverage(params.k):
all_covs.append(rec)
median = self.medianCoverage(all_covs)
sys.stdout.info("Median coverage determined as", median)
lids = [line.id for line in lines.unique()]
for line_id in lids:
line = lines[line_id]
s = AlignmentStorage()
s.addAll(al for al in line.read_alignments if not al.contradictingRTC())
segs = SegmentStorage().addAll(s.filterByCoverage(mi=params.reliable_coverage, ma=median * 7 /4, k=params.k))
segs.mergeSegments(max(params.k - params.bad_end_length * 2, params.k / 2))
if len(segs) == 0:
sys.stdout.warn("No part of a unique edge is covered by reads", line.id)
lines.removeLine(line)
continue
if len(segs) == 1 and len(segs[0]) > len(line) - 10:
sys.stdout.info("Whole line", line.id, "is covered by reads")
continue
sys.stdout.info( "Line", line.id, "has poorly covered regions. Splitting into", len(segs), "parts")
sys.stdout.trace(segs)
next_left = segs[-1].left
line.cutRight(segs[-1].right)
for seg in list(segs)[-2::-1]:
if next_left < seg.right:
line, new_line = lines.splitLine(line.segment(next_left, seg.right))
else:
line, new_line = lines.splitLine(line.segment(next_left, next_left))
line.cutRight(seg.right)
next_left = seg.left
line.rc.cutRight(len(segs[0]))
def FireMergedLines(self, al1, al2):
# type: (AlignmentPiece, AlignmentPiece) -> None
sys.stdout.trace("Fire merged lines", al1, al2)
new_line = al1.seg_to.contig
line1 = al1.seg_from.contig
line2 = al2.seg_from.contig
sys.stdout.trace(list(self.allInter(line1.asSegment())))
sys.stdout.trace(list(self.allInter(line2.asSegment())))
self.addLine(new_line)
self.auto_alignments[line1.id].setState(-1)
self.auto_alignments[line2.id].setState(-1)
auto1 = AutoAlignmentStorage(new_line).addAll([
al1.composeTargetDifference(al.compose(al1))
for al in self.auto_alignments[line1.id].content
])
auto2 = AutoAlignmentStorage(new_line).addAll([
al2.composeTargetDifference(al.compose(al2))
for al in self.auto_alignments[line2.id].content
])
auto3 = AutoAlignmentStorage(new_line).addAll([
al2.composeTargetDifference(al.compose(al1))
for al in self.getAlignmentsToFrom(line1, line2)
])
self.auto_alignments[new_line.id].addAll(
auto1.merge(auto3).merge(auto2).content)
rc1 = RCAlignmentStorage(new_line).addAll([
al1.rc.composeTargetDifference(al.compose(al1))
for al in self.rc_alignments[line1.id]
])
rc2 = RCAlignmentStorage(new_line).addAll([
al2.rc.composeTargetDifference(al.compose(al2))
for al in self.rc_alignments[line2.id]
])
rc3 = RCAlignmentStorage(new_line).addAll([
al2.rc.composeTargetDifference(al.compose(al1))
for al in self.getAlignmentsToFrom(line1, line2.rc)
])
self.rc_alignments[new_line.id].addAll(rc1.merge(rc3).merge(rc2))
common = set(self.alignmentsToFrom[line1.id].keys()).intersection(
set(self.alignmentsToFrom[line2.id].keys()))
for storage in self.alignmentsToFrom[line1.id].values():
if storage.line_from.id not in common and storage.line_from != line2 and storage.line_from != line2.rc:
self.addTwoLineStorage(new_line, storage.line_from).addAll(
[al.compose(al1) for al in storage])
for storage in self.alignmentsToFrom[line2.id].values():
if storage.line_from.id not in common and storage.line_from != line1 and storage.line_from != line1.rc:
self.addTwoLineStorage(new_line, storage.line_from).addAll(
[al.compose(al2) for al in storage])
for c in common:
storage1 = self.alignmentsToFrom[line1.id][c]
storage2 = self.alignmentsToFrom[line2.id][c]
als1 = AlignmentStorage().addAll(
[al.compose(al1) for al in storage1])
als2 = AlignmentStorage().addAll(
[al.compose(al2) for al in storage2])
self.addTwoLineStorage(new_line,
storage1.line_from).addAll(als1.merge(als2))
self.removeLine(al1.seg_from.contig)
self.removeLine(al2.seg_from.contig)
sys.stdout.trace(list(self.allInter(new_line.asSegment())))
def __init__(self, line, rc = None):
# type: (Contig, Optional[RCAlignmentStorage]) -> None
self.line = line
if rc is None:
self.content = AlignmentStorage()
rc = RCAlignmentStorage(line.rc, self)
rc.content = self.content.rc
LineListener.__init__(self, rc)
self.rc = rc # type: AutoAlignmentStorage
def __init__(self, line, rc = None):
# type: (Contig, Optional[AutoAlignmentStorage]) -> None
self.line = line
if rc is None:
self.content = AlignmentStorage()
rc = AutoAlignmentStorage(line.rc, self)
rc.content = self.content.rc
rc.state = -1
LineListener.__init__(self, rc)
self.rc = rc # type: AutoAlignmentStorage
self.state = 1 # from precedes to
def __init__(self, seq, id, rc=None):
# type: (str, str, Optional[Disjointig]) -> None
self.seq = seq
self.id = id
if rc is None:
self.read_alignments = AlignmentStorage() # type: AlignmentStorage
rc = Disjointig(basic.RC(seq), basic.Reverse(id),
self) # type: Disjointig
self.rc = rc
else:
self.rc = rc
self.read_alignments = self.rc.read_alignments.rc # type: AlignmentStorage
Contig.__init__(self, seq, id, rc)
self.rc = rc # type:Disjointig
def polyshSegments(self, line, to_polysh):
# type: (NewLine, Iterable[Segment]) -> List[Segment]
segs = SegmentStorage()
corrections = AlignmentStorage()
line.addListener(segs)
segs.addAll(to_polysh)
segs.mergeSegments()
segs.sort()
for seg in segs:
corrections.add(
self.polisher.polishSegment(
seg, list(line.read_alignments.allInter(seg))))
line.correctSequence(list(corrections))
line.removeListener(segs)
return list(segs)
class Disjointig(Contig):
def __init__(self, seq, id, rc=None):
# type: (str, str, Optional[Disjointig]) -> None
self.seq = seq
self.id = id
if rc is None:
self.read_alignments = AlignmentStorage() # type: AlignmentStorage
rc = Disjointig(basic.RC(seq), basic.Reverse(id),
self) # type: Disjointig
self.rc = rc
else:
self.rc = rc
self.read_alignments = self.rc.read_alignments.rc # type: AlignmentStorage
Contig.__init__(self, seq, id, rc)
self.rc = rc # type:Disjointig
def addAlignments(self, als):
# type: (Iterable[AlignmentPiece]) -> None
self.read_alignments.addAll(als)
def addAlignment(self, al):
# type: (AlignmentPiece) -> None
self.read_alignments.add(al)
def getAlignmentsTo(self, seg):
# type: (Segment) -> Generator[AlignmentPiece]
return self.read_alignments.getAlignmentsTo(seg)
def allInter(self, seg, min_inter=1):
# type: (Segment, int) -> Generator[AlignmentPiece]
return self.read_alignments.allInter(seg, min_inter)
def save(self, handler):
# type: (TokenWriter) -> None
handler.writeTokenLine(self.id)
handler.writeTokenLine(self.seq)
self.read_alignments.save(handler)
def loadDisjointig(self, handler, reads):
# type: (TokenReader, ReadCollection) -> None
self.id = handler.readToken()
self.rc.id = basic.Reverse(self.id)
seq = handler.readToken()
self.read_alignments.load(handler, reads, self)
def __init__(self, line_from, line_to, rc = None, reverse = None):
# type: (Contig, Contig, Optional[TwoLineAlignmentStorage], Optional[TwoLineAlignmentStorage]) -> None
assert line_from.id != line_to.id and line_from.rc.id != line_to.id
self.line_from = line_from
self.line_to = line_to
self.reverse = reverse
if rc is None:
self.content = AlignmentStorage()
self.rc = TwoLineAlignmentStorage(line_from.rc, line_to.rc, self, None)
else:
self.rc = rc
self.content = rc.content.rc # type: AlignmentStorage
LineListener.__init__(self, self.rc)
self.rc = self.rc # type: TwoLineAlignmentStorage
if reverse is None and rc is None:
reverse = TwoLineAlignmentStorage(line_to, line_from, None, self)
self.reverse = reverse
self.reverse.reverse = self
self.rc.reverse = self.reverse.rc
self.rc.reverse.reverse = self.rc
def fireAfterExtendRight(self, line, seq, relevant_als = None):
# type: (NewLine, str, Optional[List[AlignmentPiece]]) -> None
line = line # type: NewLine
if relevant_als is not None:
tmp = line.read_alignments.merge(AlignmentStorage().addAll(relevant_als).targetAsSegment(line.asSegment()))
line.read_alignments.clean()
line.read_alignments.addAll(tmp)
new_seg = line.asSegment().suffix(length = min(len(line), len(seq) + params.k + 100))
for al in self.aligner.dotplotAlign([new_seg.asContig()], self.disjointigs):
if len(al) < params.k:
continue
al = al.reverse().targetAsSegment(new_seg)
line.disjointig_alignments.addAndMergeRight(al)
class TwoLineAlignmentStorage(LineListener):
def __init__(self, line_from, line_to, rc = None, reverse = None):
# type: (Contig, Contig, Optional[TwoLineAlignmentStorage], Optional[TwoLineAlignmentStorage]) -> None
assert line_from.id != line_to.id and line_from.rc.id != line_to.id
self.line_from = line_from
self.line_to = line_to
self.reverse = reverse
if rc is None:
self.content = AlignmentStorage()
self.rc = TwoLineAlignmentStorage(line_from.rc, line_to.rc, self, None)
else:
self.rc = rc
self.content = rc.content.rc # type: AlignmentStorage
LineListener.__init__(self, self.rc)
self.rc = self.rc # type: TwoLineAlignmentStorage
if reverse is None and rc is None:
reverse = TwoLineAlignmentStorage(line_to, line_from, None, self)
self.reverse = reverse
self.reverse.reverse = self
self.rc.reverse = self.reverse.rc
self.rc.reverse.reverse = self.rc
def add(self, al):
# type: (AlignmentPiece) -> None
assert al.seg_from.contig == self.line_from
assert al.seg_to.contig == self.line_to
self.content.add(al)
reverse = al.reverse()
self.reverse.content.add(reverse)
def addAll(self, als):
for al in als:
self.add(al)
return self
def __iter__(self):
# type: () -> Generator[AlignmentPiece]
return self.content.__iter__()
def getAlignmentsTo(self, seg):
# type: (Segment) -> Generator[AlignmentPiece]
return self.content.getAlignmentsTo(seg)
def allInter(self, seg):
return self.content.allInter(seg)
def normalizeReverse(self):
self.reverse.content = self.content.reverse()
self.reverse.rc.content = self.reverse.content.rc
def fireBeforeExtendRight(self, line, new_seq, seq):
# type: (Any, Contig, str) -> None
self.content.fireBeforeExtendRight(line, new_seq, seq)
self.normalizeReverse()
def fireBeforeCutRight(self, line, new_seq, pos):
# type: (Any, Contig, int) -> None
self.content.fireBeforeCutRight(line, new_seq, pos)
self.normalizeReverse()
# alignments from new sequence to new sequence
def fireBeforeCorrect(self, alignments):
# type: (Correction) -> None
self.content.fireBeforeCorrect(alignments)
self.normalizeReverse()
def fireAfterExtendRight(self, line, seq, relevant_als = None):
# type: (Any, str, Optional[List[AlignmentPiece]]) -> None
self.content.fireAfterExtendRight(line, seq)
self.normalizeReverse()
def fireAfterCutRight(self, line, pos):
# type: (Any, int) -> None
self.content.fireAfterCutRight(line, pos)
self.normalizeReverse()
def fireAfterCorrect(self, line, alignments):
# type: (Any, Correction) -> None
self.content.fireAfterCorrect(line, alignments)
self.normalizeReverse()
def addAndMergeRight(self, al):
self.content.addAndMergeRight(al)
self.normalizeReverse()
def merge(self, other):
# type: (TwoLineAlignmentStorage) -> TwoLineAlignmentStorage
res = TwoLineAlignmentStorage(self.line_from, self.line_to)
res.content.addAll(self.content.merge(other.content))
res.normalizeReverse()
def save(self, handler):
# type: (TokenWriter) -> None
self.content.save(handler)
def load(self, handler, lines):
# type: (TokenReader, Any) -> None
self.content.load(handler, lines, lines)
self.normalizeReverse()
class NewLine(Contig):
def __init__(self, seq, id, extension_handler, rc = None):
# type: (str, str, ExtensionHandler, Optional[NewLine]) -> None
self.extensionHandler = extension_handler
self.seq = seq
self.id = id # type: str
self.circular = False
self.name_printer = None
self.max_extension = False
if rc is None:
self.initial = AlignmentStorage()
self.correct_segments = SegmentStorage()
self.completely_resolved = SegmentStorage()
self.disjointig_alignments = AlignmentStorage()
self.read_alignments = ReadAlignmentStorage()
self.listeners = [self.initial, self.correct_segments, self.completely_resolved, self.disjointig_alignments, self.read_alignments, extension_handler] # type: List[LineListener]
rc = NewLine(basic.RC(seq), basic.Reverse(self.id), extension_handler.rc, self) #type: NewLine
self.rc = rc
self.addListener(ReadAlignmentListener(self))
# self.initial.add(AlignmentPiece.Identical(self.asSegment().asContig().asSegment(), self.asSegment()))
else:
self.initial = rc.initial.rc # type: AlignmentStorage
self.correct_segments = rc.correct_segments.rc # type: SegmentStorage
self.completely_resolved = rc.completely_resolved.rc # type: SegmentStorage
self.disjointig_alignments = rc.disjointig_alignments.rc # type: AlignmentStorage
self.read_alignments = rc.read_alignments.rc # type: ReadAlignmentStorage
self.listeners = [listener.rc for listener in rc.listeners] # type: List[LineListener]
Contig.__init__(self, seq, id, rc)
self.rc = rc #type: NewLine
self.knot = None # type: Knot
def updateCorrectSegments(self, seg, threshold = params.reliable_coverage):
# type: (Segment, int) -> None
segs = AlignmentStorage().addAll(self.read_alignments.allInter(seg)).filterByCoverage(mi=threshold)
self.correct_segments.addAll(segs)
self.correct_segments.mergeSegments()
def addReads(self, alignments):
# type: (Iterable[AlignmentPiece]) -> None
self.read_alignments.addAll(alignments)
self.max_extension = False
def getReadAlignmentsTo(self, seg):
# type: (Segment) -> Iterable[AlignmentPiece]
return self.read_alignments.getAlignmentsTo(seg)
def getPotentialAlignmentsTo(self, seg):
# type: (Segment) -> Generator[AlignmentPiece]
result = []
for alDL in self.disjointig_alignments.getAlignmentsTo(seg):
reduced = alDL.reduce(target=seg)
dt = alDL.seg_from.contig # type: Disjointig
for alRD in dt.getAlignmentsTo(reduced.seg_from):
result.append(alRD.compose(alDL))
result = sorted(result, key = lambda al: (al.seg_from.contig.id, -len(al.seg_from)))
for read, iter in itertools.groupby(result, key = lambda al: al.seg_from.contig):
readRes = []
for al in iter:
found = False
for al1 in readRes:
inter = al.matchingSequence(True).inter(al1.matchingSequence(True))
if len(inter.matches) != 0:
found = True
if not found:
yield al
readRes.append(al)
def getRelevantAlignmentsFor(self, seg):
# type: (Segment) -> Generator[AlignmentPiece]
sys.stdout.trace("Requesting read alignments for", seg)
result = []
if params.debug:
print self.disjointig_alignments
print list(self.disjointig_alignments.allInter(seg))
for alDL in self.disjointig_alignments.allInter(seg):
if len(alDL.seg_to) < params.k:
continue
reduced = alDL.reduce(target=seg)
dt = alDL.seg_from.contig # type: Disjointig
cnt = 0
als = filter(lambda al: al.seg_to.interSize(alDL.seg_from) > 8 * params.k / 10, dt.allInter(reduced.seg_from))
compositions = alDL.massComposeBack(als)
for al in compositions:
if len(al.seg_to) >= params.k:
result.append(al)
cnt += 1
sys.stdout.trace("Request for read alignments for", seg, " collecting finished. Started filtering")
result = sorted(result, key = lambda al: (al.seg_from.contig.id, -len(al.seg_from)))
for read, iter in itertools.groupby(result, key = lambda al: al.seg_from.contig): # type: AlignedRead, Generator[AlignmentPiece]
readRes = []
for al in iter:
found = False
for al1 in readRes:
inter = al.matchingSequence(True).inter(al1.matchingSequence(True))
if len(inter.matches) != 0:
found = True
break
if not found:
if params.debug:
print al
yield al
readRes.append(al)
sys.stdout.trace("Request for read alignments for", seg, "finished")
def position(self, pos):
# type: (int) -> LinePosition
return LinePosition(self, pos)
def extendRight(self, seq, relevant_als = None):
# type: (str, List[AlignmentPiece]) -> None
sys.stdout.trace("Line operation Extend:", self, len(seq), relevant_als)
assert self.knot is None
if relevant_als is None:
relevant_als = []
new_seq = Contig(self.seq + seq, "TMP2_" + self.id)
self.notifyBeforeExtendRight(new_seq, seq)
self.seq = self.seq + seq
self.rc.seq = basic.RC(seq) + self.rc.seq
self.notifyAfterExtendRight(seq, relevant_als)
self.updateCorrectSegments(self.asSegment())
self.max_extension = True
def notifyBeforeExtendRight(self, new_seq, seq):
# type: (Contig, str) -> None
for listener in self.listeners:
listener.fireBeforeExtendRight(self, new_seq, seq)
def notifyAfterExtendRight(self, seq, relevant_als):
# type: (str, Optional[List[AlignmentPiece]]) -> None
for listener in self.listeners:
listener.fireAfterExtendRight(self, seq, relevant_als)
def cutRight(self, pos):
sys.stdout.trace("Line operation Cut:", self, pos)
assert pos > 0 and pos <= len(self)
cut_length = len(self) - pos
if cut_length == 0:
return
new_seq = Contig(self.seq[:pos], "TMP3_" + self.id)
self.notifyBeforeCutRight(new_seq, pos)
self.seq = self.seq[:-cut_length]
self.rc.seq = self.rc.seq[cut_length:]
self.notifyAfterCutRight(pos)
def notifyBeforeCutRight(self, new_seq, pos):
# type: (Contig, int) -> None
for listener in self.listeners:
listener.fireBeforeCutRight(self, new_seq, pos)
def notifyAfterCutRight(self, pos):
# type: (int) -> None
for listener in self.listeners:
listener.fireAfterCutRight(self, pos)
def correctSequence(self, alignments):
# type: (Iterable[AlignmentPiece]) -> None
sys.stdout.trace("Line operation Correct:", alignments)
alignments = [al.cutIdenticalEnds() for al in alignments if al.seg_from.Seq() != al.seg_to.Seq()]
if len(alignments) == 0:
sys.stdout.trace("Skipping trivial correction operation")
return
assert len(alignments) > 0
correction = Correction.constructCorrection(alignments)
self.notifyBeforeCorrect(correction)
old = Contig(self.seq, "old")
self.seq = correction.seq_from.seq
self.rc.seq = basic.RC(self.seq)
correction.changeQT(self, old)
self.notifyAfterCorrect(correction)
def notifyBeforeCorrect(self, alignments):
# type: (Correction) -> None
for listener in self.listeners:
listener.fireBeforeCorrect(alignments)
def notifyAfterCorrect(self, alignments):
# type: (Correction) -> None
for listener in self.listeners:
listener.fireAfterCorrect(self, alignments)
def addReadAlignment(self, al):
# type: (AlignmentPiece) -> AlignmentPiece
self.read_alignments.add(al)
self.max_extension = False
return al
def addListener(self, listener):
self.listeners.append(listener)
self.rc.listeners.append(listener.rc)
def removeListener(self, listener):
self.listeners.remove(listener)
self.rc.listeners.remove(listener.rc)
def save(self, handler):
# type: (TokenWriter) -> None
handler.writeTokenLine(self.id)
handler.writeTokenLine(self.seq)
self.initial.save(handler)
self.correct_segments.save(handler)
self.completely_resolved.save(handler)
self.disjointig_alignments.save(handler)
self.read_alignments.save(handler)
def loadLine(self, handler, disjointigs, reads, contigs):
# type: (TokenReader, DisjointigCollection, ReadCollection, ContigCollection) -> None
self.id = handler.readToken()
self.seq = handler.readToken()
self.rc.id = basic.Reverse(self.id)
n = handler.readInt()
for i in range(n):
handler.readToken()
handler.readToken()
handler.readToken()
seg = Segment.load(handler, self)
handler.readToken()
self.initial.add(AlignmentPiece.Identical(seg.asContig().asSegment(), seg))
# self.add(AlignmentPiece.load(handler, collection_from, collection_to))
self.correct_segments.load(handler, self)
self.completely_resolved.load(handler, self)
self.disjointig_alignments.load(handler, disjointigs, self)
self.read_alignments.load(handler, reads, self)
for al in self.read_alignments:
read = al.seg_from.contig #type: AlignedRead
read.addAlignment(al)
self.max_extension = False
def __str__(self):
if self.name_printer is not None:
return self.name_printer(self)
points = [self.left()]
if len(self.initial) == 0:
points.append("NA")
else:
points.append(self.initial[0].seg_to.left)
points.append(self.initial[-1].seg_to.right)
points.append(self.right())
points = map(str, points)
return "Line:" + str(self.id) + ":" + "[" + ":".join(points) +"]"
def __repr__(self):
points = [self.left()]
points.extend(self.initial)
points.append(self.right())
points = map(str, points)
return "Line:" + str(self.id) + ":" + "[" + ":".join(points) +"]"
def setCircular(self):
self.circular = True
self.rc.circular = True
def cleanReadAlignments(self):
for read in self.read_alignments:
read.seg_from.contig.removeContig(self)
self.read_alignments.clean()
self.max_extension = False
def tie(self, other, gap, gap_seq):
self.knot = Knot(self, other, gap, gap_seq)
other.rc.knot = self.knot.rc
if self == other:
self.setCircular()
def unTie(self):
if self.knot is not None:
self.knot.line_right.rc.knot = None
if self.knot is not None:
self.knot = None
class RCAlignmentStorage(LineListener):
def __init__(self, line, rc = None):
# type: (Contig, Optional[RCAlignmentStorage]) -> None
self.line = line
if rc is None:
self.content = AlignmentStorage()
rc = RCAlignmentStorage(line.rc, self)
rc.content = self.content.rc
LineListener.__init__(self, rc)
self.rc = rc # type: AutoAlignmentStorage
def __iter__(self):
# type: () -> Generator[AlignmentPiece]
return self.content.__iter__()
def getAlignmentsTo(self, seg):
# type: (Segment) -> Generator[AlignmentPiece]
return self.content.getAlignmentsTo(seg)
def allInter(self, seg):
return self.content.allInter(seg)
def add(self, alignment):
self.content.add(alignment)
self.content.add(alignment.reverse().rc)
def addAndMergeRight(self, al):
# type: (AlignmentPiece) -> None
self.content.addAndMergeRight(al)
self.content.addAndMergeLeft(al.reverse().rc)
def addAll(self, als):
for al in als:
self.add(al)
return self
def fireBeforeExtendRight(self, line, new_seq, seq):
# type: (Any, Contig, str) -> None
self.content.fireBeforeExtendRight(line, new_seq, seq)
self.reverse()
self.content.fireBeforeExtendRight(line, new_seq, seq)
def fireBeforeCutRight(self, line, new_seq, pos):
# type: (Any, Contig, int) -> None
self.content.fireBeforeCutRight(line, new_seq, pos)
self.reverse()
self.content.fireBeforeCutRight(line, new_seq, pos)
# alignments from new sequence to new sequence
def fireBeforeCorrect(self, alignments):
# type: (Correction) -> None
self.content.fireBeforeCorrect(alignments)
self.reverse()
self.content.fireBeforeCorrect(alignments)
def fireAfterExtendRight(self, line, seq, relevant_als = None):
# type: (Any, str, Optional[List[AlignmentPiece]]) -> None
self.content.fireAfterExtendRight(line, seq)
self.reverse()
self.content.fireAfterExtendRight(line, seq)
def fireAfterCutRight(self, line, pos):
# type: (Any, int) -> None
self.content.fireAfterCutRight(line, pos)
self.reverse()
self.content.fireAfterCutRight(line, pos)
def fireAfterCorrect(self, line, alignments):
# type: (Any, Correction) -> None
self.content.fireAfterCorrect(line, alignments)
self.reverse()
self.content.fireAfterCorrect(line, alignments)
# This is CRAAAZY!!! But correct.
def reverse(self):
self.rc.content = self.content.reverse()
self.content = self.rc.content.rc
def merge(self, other):
# type: (RCAlignmentStorage) -> RCAlignmentStorage
res = RCAlignmentStorage(self.line)
res.content.addAll(self.content.merge(other.content))
return res
def save(self, handler):
# type: (TokenWriter) -> None
self.content.save(handler)
def load(self, handler):
# type: (TokenReader) -> None
self.content.load(handler, self.line.rc, self.line)
def updateCorrectSegments(self, seg, threshold = params.reliable_coverage):
# type: (Segment, int) -> None
segs = AlignmentStorage().addAll(self.read_alignments.allInter(seg)).filterByCoverage(mi=threshold)
self.correct_segments.addAll(segs)
self.correct_segments.mergeSegments()
class AutoAlignmentStorage(LineListener):
def __init__(self, line, rc = None):
# type: (Contig, Optional[AutoAlignmentStorage]) -> None
self.line = line
if rc is None:
self.content = AlignmentStorage()
rc = AutoAlignmentStorage(line.rc, self)
rc.content = self.content.rc
rc.state = -1
LineListener.__init__(self, rc)
self.rc = rc # type: AutoAlignmentStorage
self.state = 1 # from precedes to
def makeCanonical(self, al):
if (self.state == 1) == (al.seg_from.left < al.seg_to.left):
return al
else:
return al.reverse()
def isCanonical(self, al):
return (self.state == 1) == (al.seg_from.left < al.seg_to.left)
def add(self, al):
# type: (AlignmentPiece) -> None
if al.isIdentical():
return
self.content.add(self.makeCanonical(al))
def addAll(self, als):
for al in als:
self.add(al)
return self
def addAndMergeRight(self, al):
if al.isIdentical():
return
if self.isCanonical(al):
self.content.addAndMergeRight(al)
else:
self.content.addAndMergeRight(al.reverse())
def __iter__(self):
# type: () -> Generator[AlignmentPiece]
for al in self.content:
yield al
for al in self.content:
yield al.reverse()
yield AlignmentPiece.Identical(self.line.asSegment())
def getAlignmentsTo(self, seg):
# type: (Segment) -> Generator[AlignmentPiece]
for al in self:
if al.seg_to.contains(seg):
yield al
def allInter(self, seg):
for al in self:
if al.seg_to.inter(seg):
yield al
def fireBeforeExtendRight(self, line, new_seq, seq):
# type: (Any, Contig, str) -> None
self.content.fireBeforeExtendRight(line, new_seq, seq)
self.reverse()
self.content.fireBeforeExtendRight(line, new_seq, seq)
# alignments from new sequence to new sequence
def fireBeforeCutRight(self, line, new_seq, pos):
# type: (Any, Contig, int) -> None
self.content.fireBeforeCutRight(line, new_seq, pos)
self.reverse()
self.content.fireBeforeCutRight(line, new_seq, pos)
def fireBeforeCorrect(self, alignments):
# type: (Correction) -> None
self.content.fireBeforeCorrect(alignments)
self.reverse()
self.content.fireBeforeCorrect(alignments)
def fireAfterExtendRight(self, line, seq, relevant_als = None):
# type: (Any, str, Optional[List[AlignmentPiece]]) -> None
self.content.fireAfterExtendRight(line, seq)
self.reverse()
self.content.fireAfterExtendRight(line, seq)
def fireAfterCutRight(self, line, pos):
# type: (Any, int) -> None
self.content.fireAfterCutRight(line, pos)
self.reverse()
self.content.fireAfterCutRight(line, pos)
def fireAfterCorrect(self, line, alignments):
# type: (Any, Correction) -> None
self.content.fireAfterCorrect(line, alignments)
self.reverse()
self.content.fireAfterCorrect(line, alignments)
def reverse(self):
self.state = -self.state
self.rc.state = -self.rc.state
self.content = self.content.reverse()
self.rc.content = self.content.rc
def merge(self, other):
# type: (AutoAlignmentStorage) -> AutoAlignmentStorage
if self.state != other.state:
self.reverse()
res = AutoAlignmentStorage(self.line)
res.state = self.state
res.content.addAll(self.content.merge(other.content))
return res
def save(self, handler):
# type: (TokenWriter) -> None
self.content.save(handler)
def load(self, handler):
# type: (TokenReader) -> None
self.content.load(handler, self.line, self.line)
def setState(self, state):
assert state in [-1, 1]
if self.state != state:
self.reverse()
def mergeLines(self, alignment, k):
# type: (AlignmentPiece, int) -> NewLine
sys.stdout.trace("Line operation Merge", alignment.seg_from.contig,
alignment.seg_to.contig, alignment)
line1 = alignment.seg_from.contig #type: NewLine
line2 = alignment.seg_to.contig #type: NewLine
assert line1 != line2
if len(alignment) < k + 100:
sys.stdout.trace(
"Prolonging line to ensure alignment of at least k")
seg = line2.segment(
alignment.seg_to.right,
alignment.seg_to.right + k + 100 - len(alignment))
line1.extendRight(seg.Seq())
alignment = alignment.mergeDistant(
AlignmentPiece.Identical(
line1.asSegment().suffix(length=len(seg)), seg))
# Cutting hanging tips of both lines
al_storage = AlignmentStorage()
al_storage.add(alignment)
storage = TwoLineAlignmentStorage(line1, line2)
line2.addListener(storage)
line1.addListener(storage.reverse)
storage.add(alignment)
if alignment.seg_from.right < len(line1):
line1.cutRight(alignment.seg_from.right)
sys.stdout.trace("Cut right")
sys.stdout.trace(list(storage.content)[0])
sys.stdout.trace("\n".join(
list(storage.content)[0].asMatchingStrings()))
sys.stdout.trace(list(storage.content)[0].cigar)
if alignment.seg_to.left > 0:
line2.rc.cutRight(len(line2) - alignment.seg_to.left)
sys.stdout.trace("Cut left")
sys.stdout.trace(list(storage.content)[0])
sys.stdout.trace("\n".join(
list(storage.content)[0].asMatchingStrings()))
sys.stdout.trace(list(storage.content)[0].cigar)
alignment = list(storage.content)[0] # type: AlignmentPiece
line2.removeListener(storage)
line1.removeListener(storage.reverse)
# Making sure line sequences match on the overlap
if alignment.seg_from.left > 0:
new_seq = Contig(
line1.asSegment().prefix(pos=alignment.seg_from.left).Seq() +
line2.seq, "new_seq")
else:
new_seq = Contig(line2.seq, "new_seq")
al2 = AlignmentPiece.Identical(
line2.asSegment(),
new_seq.asSegment().suffix(length=len(line2)))
sys.stdout.trace("Al2:", al2)
alignment = alignment.compose(al2).reverse()
sys.stdout.trace("Composed alignment", alignment)
sys.stdout.trace("\n".join(alignment.asMatchingStrings()))
sys.stdout.trace(alignment.cigar)
assert alignment.seg_to.right == len(line1)
assert alignment.seg_from.left == al2.seg_to.left
line1.correctSequence([alignment])
# Now lines have exact match
name = "(" + ",".join(
basic.parseLineName(line1.id) +
basic.parseLineName(line2.id)) + ")"
line = self.addNew(new_seq.seq, name)
assert line.seq.startswith(line1.seq)
assert line.seq.endswith(line2.seq)
al1 = AlignmentPiece.Identical(
line1.asSegment(),
line.asSegment().prefix(length=len(line1)))
al2 = AlignmentPiece.Identical(
line2.asSegment(),
line.asSegment().suffix(length=len(line2)))
line.initial.addAll(
line1.initial.targetAsSegment(al1.seg_to).merge(
line2.initial.targetAsSegment(al2.seg_to)))
line.correct_segments.addAll(
line1.correct_segments.contigAsSegment(al1.seg_to).merge(
line2.correct_segments.contigAsSegment(al2.seg_to)))
line.completely_resolved.addAll(
line1.completely_resolved.contigAsSegment(al1.seg_to).merge(
line2.completely_resolved.contigAsSegment(al2.seg_to), k))
line.disjointig_alignments.addAll(
line1.disjointig_alignments.targetAsSegment(al1.seg_to).merge(
line2.disjointig_alignments.targetAsSegment(al2.seg_to)))
for al in line1.read_alignments.targetAsSegment(al1.seg_to).merge(
line2.read_alignments.targetAsSegment(al2.seg_to)):
line.addReadAlignment(al)
line1.cleanReadAlignments()
line2.cleanReadAlignments()
self.notifyMergedLines(al1, al2)
knot_right = line2.knot
knot_left = line1.rc.knot
self.remove(line1)
self.remove(line2)
if knot_right is not None:
if knot_right.line_right == line1:
line.tie(line, knot_right.gap, knot_right.gap_seq)
else:
line.tie(knot_right.line_right, knot_right.gap,
knot_right.gap_seq)
if knot_left is not None and knot_left.line_right != line2.rc:
line.rc.tie(knot_left.line_right, knot_left.gap, knot_left.gap_seq)
return line
def testManual(self):
contig1 = Contig("ACGTACGTACGT", "from")
contig2 = Contig("ACGTACGTACGT", "to")
al1 = AlignmentPiece.Identical(contig1.segment(0, 4),
contig2.segment(0, 4))
al2 = AlignmentPiece.Identical(contig1.segment(0, 4),
contig2.segment(4, 8))
al3 = AlignmentPiece.Identical(contig1.segment(4, 8),
contig2.segment(8, 12))
storage = AlignmentStorage()
storage.addAll([al1, al2, al3])
assert str(
list(storage)
) == "[(from[0:4]->to[0:4]:1.000), (from[0:4]->to[4:12-4]:1.000), (from[4:12-4]->to[8:12-0]:1.000)]"
assert str(
list(storage.rc)
) == "[(-from[4:12-4]->-to[0:4]:1.000), (-from[8:12-0]->-to[4:12-4]:1.000), (-from[8:12-0]->-to[8:12-0]:1.000)]"
assert str(list(storage.calculateCoverage())) == "[(to[0:12-0], 1)]"
assert str(list(storage.filterByCoverage(0, 1))) == "[]"
assert str(list(storage.filterByCoverage(1, 2))) == "[to[0:12-0]]"
assert str(list(storage.filterByCoverage(2))) == "[]"
storage.addAndMergeRight(al3)
assert str(
list(storage)
) == "[(from[0:4]->to[0:4]:1.000), (from[0:4]->to[4:12-4]:1.000), (from[4:12-4]->to[8:12-0]:1.000)]"
al4 = AlignmentPiece.Identical(contig1.segment(2, 8),
contig2.segment(2, 8))
al5 = AlignmentPiece.Identical(contig1.segment(4, 10),
contig2.segment(4, 10))
storage.addAll([al4, al5])
assert str(
list(storage.calculateCoverage())
) == "[(to[0:2], 1), (to[2:4], 2), (to[4:12-4], 3), (to[8:12-2], 2), (to[10:12-0], 1)]"
assert str(list(storage.filterByCoverage(
2, 3))) == "[to[2:4], to[8:12-2]]"
assert str(list(storage.filterByCoverage(2))) == "[to[2:12-2]]"
assert str(
list(storage.getAlignmentsTo(contig2.segment(2, 3)))
) == "[(from[0:4]->to[0:4]:1.000), (from[2:12-4]->to[2:12-4]:1.000)]"
assert str(list(storage.getAlignmentsTo(contig2.segment(
2, 6)))) == "[(from[2:12-4]->to[2:12-4]:1.000)]"