void GappedSegmentIteratorIndelTest::createCallBack(AlignmentPtr alignment)
{
  addIdenticalParentChild(alignment, 1, 20, 5);
  Genome* parent = alignment->openGenome(alignment->getRootName());
  Genome* child = parent->getChild(0);
  TopSegmentIteratorPtr ti = child->getTopSegmentIterator();
  BottomSegmentIteratorPtr bi = parent->getBottomSegmentIterator();
//  int i = 0;
//  bool reversed = true;

  bi = parent->getBottomSegmentIterator(0);
  makeDelGap(bi);
  bi = parent->getBottomSegmentIterator(3);
  makeDelGap(bi);
/*
  ti = child->getTopSegmentIterator(1);
  makeInsGap(ti);
  ti = child->getTopSegmentIterator(21);
  makeInsGap(ti);
  ti = child->getTopSegmentIterator(28);
  makeInsGap(ti);
*/  
/*  for (size_t i = 0; i < 20; ++i)
  {
    cout << i << ": ";
    bi = parent->getBottomSegmentIterator(i);
    ti = child->getTopSegmentIterator(i);
    cout << "ci=" << bi->getBottomSegment()->getChildIndex(0) 
         << " pi=" << ti->getTopSegment()->getParentIndex() << endl;
         }*/
}
void GappedSegmentSimpleIteratorTest2::createCallBack(AlignmentPtr alignment)
{
  addIdenticalParentChild(alignment, 2, 100, 5);
  Genome* parent = alignment->openGenome(alignment->getRootName());
  Genome* child = parent->getChild(0);
  TopSegmentIteratorPtr ti = child->getTopSegmentIterator();
  BottomSegmentIteratorPtr bi = parent->getBottomSegmentIterator();
  hal_index_t i = 0;
  bool reversed = true;
  while (ti != child->getTopSegmentEndIterator())
  {
    if (i % 5 == 0)
    {
      reversed = !reversed;
      if (reversed && i < (hal_index_t)(parent->getNumBottomSegments() - 1))
      {
        makeInversion(ti, 5);
      }
    }

    ti->toRight();
    bi->toRight();
    ++i;
  }
}
示例#3
0
void TopSegmentSimpleIteratorTest::createCallBack(Alignment *alignment) {
    Genome *ancGenome = alignment->addRootGenome("Anc0", 0);
    size_t numChildren = 9;
    for (size_t i = 0; i < numChildren; ++i) {
        alignment->addLeafGenome("Leaf" + std::to_string(i), "Anc0", 0.1);
    }
    vector<Sequence::Info> seqVec(1);
    seqVec[0] = Sequence::Info("Sequence", 1000000, 5000, 10000);
    ancGenome->setDimensions(seqVec);

    CuAssertTrue(_testCase, ancGenome->getNumChildren() == numChildren);

    _topSegments.clear();
    for (size_t i = 0; i < ancGenome->getNumTopSegments(); ++i) {
        TopSegmentStruct topSeg;
        topSeg.setRandom();
        topSeg._length = ancGenome->getSequenceLength() / ancGenome->getNumTopSegments();
        topSeg._startPosition = i * topSeg._length;
        _topSegments.push_back(topSeg);
    }

    TopSegmentIteratorPtr tsIt = ancGenome->getTopSegmentIterator(0);
    for (size_t i = 0; not tsIt->atEnd(); tsIt->toRight(), ++i) {
        CuAssertTrue(_testCase, (size_t)tsIt->getTopSegment()->getArrayIndex() == i);
        _topSegments[i].applyTo(tsIt);
    }
}
示例#4
0
void TopSegmentSequenceTest::createCallBack(Alignment *alignment) {
    Genome *ancGenome = alignment->addRootGenome("Anc0", 0);
    vector<Sequence::Info> seqVec(1);
    seqVec[0] = Sequence::Info("Sequence", 1000000, 5000, 700000);
    ancGenome->setDimensions(seqVec);

    ancGenome->setSubString("CACACATTC", 500, 9);
    TopSegmentIteratorPtr tsIt = ancGenome->getTopSegmentIterator(100);
    tsIt->getTopSegment()->setCoordinates(500, 9);
}
示例#5
0
void LodExtract::writeHomologies(const Genome* inParent,
                                 const vector<const Genome*>& inChildren)
{
  vector<const Genome*> inGenomes = inChildren;
  inGenomes.push_back(inParent);
  Genome* outParent = _outAlignment->openGenome(inParent->getName());
  assert(outParent != NULL && outParent->getNumBottomSegments() > 0);
  assert(inChildren.size() > 0);
  Genome* outChild = _outAlignment->openGenome(inChildren[0]->getName());
  BottomSegmentIteratorPtr bottom = outParent->getBottomSegmentIterator();
  TopSegmentIteratorPtr top = outChild->getTopSegmentIterator();

  // FOR EVERY BLOCK
  for (hal_size_t blockIdx = 0; blockIdx < _graph.getNumBlocks(); ++blockIdx)
  {
    SegmentMap segMap;
    const LodBlock* block = _graph.getBlock(blockIdx);

    for (hal_size_t segIdx = 0; segIdx < block->getNumSegments(); ++segIdx)
    {
      const LodSegment* segment = block->getSegment(segIdx);
      const Genome* genome = segment->getSequence()->getGenome();

      // ADD TO MAP
      pair<SegmentMap::iterator, bool> res = segMap.insert(
        pair<const Genome*, SegmentSet*>(genome, NULL));
      if (res.second == true)
      {
        assert(res.first->second == NULL);
        res.first->second = new SegmentSet();
      }
      res.first->second->insert(segment);    
    }      
    updateBlockEdges(inParent, segMap, block, bottom, top);
    
    // free the temporary sets! 
    for (SegmentMap::iterator mapIt = segMap.begin(); mapIt != segMap.end();
         ++mapIt)
    {
      delete mapIt->second;
    }
  }
}
void GappedSegmentSimpleIteratorTest::createCallBack(AlignmentPtr alignment)
{
  addIdenticalParentChild(alignment, 2, 100, 5);
  Genome* parent = alignment->openGenome(alignment->getRootName());
  Genome* child = parent->getChild(0);
  TopSegmentIteratorPtr ti = child->getTopSegmentIterator();
  BottomSegmentIteratorPtr bi = parent->getBottomSegmentIterator();
  int i = 0;
  while (ti != child->getTopSegmentEndIterator())
  {
    if (i++ % 2)
    {
      ti->getTopSegment()->setParentReversed(true);
      bi->getBottomSegment()->setChildReversed(0, true);
    }
    ti->toRight();
    bi->toRight();
  }
}
示例#7
0
void GenomeCopyTest::createCallBack(Alignment *alignment) {
    hal_size_t alignmentSize = alignment->getNumGenomes();
    CuAssertTrue(_testCase, alignmentSize == 0);

    // Hacky: Need a different alignment to test copying the bottom
    // segments correctly.  (the names of a node's children are used
    // when copying bottom segments, and two genomes can't have the same
    // name in the same alignment)
    _path = getTempFile();
    _secondAlignment =
        AlignmentPtr(getTestAlignmentInstances(alignment->getStorageFormat(), _path, WRITE_ACCESS | CREATE_ACCESS));

    Genome *ancGenome = alignment->addRootGenome("AncGenome", 0);
    Genome *leafGenome = alignment->addLeafGenome("LeafGenome1", "AncGenome", 0);
    // This genome will test copyDimensions, copyTopSegments,
    // copyBottomSegments, copySequence, copyMetadata
    Genome *copyRootGenome = _secondAlignment->addRootGenome("copyRootGenome", 0);
    Genome *copyLeafGenome = _secondAlignment->addLeafGenome("LeafGenome1", "copyRootGenome", 0);

    MetaData *ancMeta = ancGenome->getMetaData();
    ancMeta->set("Young", "Jeezy");

    vector<Sequence::Info> seqVec(1);
    seqVec[0] = Sequence::Info("Sequence", 1000000, 0, 700000);
    ancGenome->setDimensions(seqVec);
    seqVec[0] = Sequence::Info("Sequence", 1000000, 5000, 0);
    leafGenome->setDimensions(seqVec);
    string ancSeq = "CAT";
    hal_index_t n = ancGenome->getSequenceLength();
    DnaIteratorPtr dnaIt = ancGenome->getDnaIterator();
    for (; dnaIt->getArrayIndex() < n; dnaIt->toRight()) {
        size_t i = dnaIt->getArrayIndex() % ancSeq.size();
        dnaIt->setBase(ancSeq[i]);
    }
    dnaIt->flush();

    n = leafGenome->getSequenceLength();
    dnaIt = leafGenome->getDnaIterator();
    for (; dnaIt->getArrayIndex() < n; dnaIt->toRight()) {
        size_t i = dnaIt->getArrayIndex() % ancSeq.size();
        dnaIt->setBase(ancSeq[i]);
    }
    dnaIt->flush();

    TopSegmentIteratorPtr topIt = leafGenome->getTopSegmentIterator();
    n = leafGenome->getNumTopSegments();
    for (; topIt->getArrayIndex() < n; topIt->toRight()) {
        topIt->setCoordinates(topIt->getArrayIndex(), 1);
        topIt->tseg()->setParentIndex(3);
        topIt->tseg()->setParentReversed(true);
        topIt->tseg()->setBottomParseIndex(5);
        if (topIt->getArrayIndex() != 6) {
            topIt->tseg()->setNextParalogyIndex(6);
        } else {
            topIt->tseg()->setNextParalogyIndex(7);
        }
    }
    BottomSegmentIteratorPtr botIt = ancGenome->getBottomSegmentIterator();
    n = ancGenome->getNumBottomSegments();
    for (; botIt->getArrayIndex() < n; botIt->toRight()) {
        botIt->setCoordinates(botIt->getArrayIndex(), 1);
        botIt->bseg()->setChildIndex(0, 3);
        botIt->bseg()->setChildReversed(0, true);
        botIt->bseg()->setTopParseIndex(5);
    }

    seqVec[0] = Sequence::Info("Sequence", 3300, 0, 1100);
    copyRootGenome->setDimensions(seqVec);
    seqVec[0] = Sequence::Info("Sequence", 3300, 2200, 0);
    copyLeafGenome->setDimensions(seqVec);
    string copySeq = "TAG";
    dnaIt = copyRootGenome->getDnaIterator();
    n = copyRootGenome->getSequenceLength();
    for (; dnaIt->getArrayIndex() < n; dnaIt->toRight()) {
        size_t i = dnaIt->getArrayIndex() % copySeq.size();
        dnaIt->setBase(copySeq[i]);
    }
    dnaIt->flush();

    dnaIt = copyLeafGenome->getDnaIterator();
    n = copyLeafGenome->getSequenceLength();
    for (; dnaIt->getArrayIndex() < n; dnaIt->toRight()) {
        size_t i = dnaIt->getArrayIndex() % copySeq.size();
        dnaIt->setBase(copySeq[i]);
    }
    dnaIt->flush();

    topIt = copyLeafGenome->getTopSegmentIterator();
    n = copyLeafGenome->getNumTopSegments();
    for (; topIt->getArrayIndex() < n; topIt->toRight()) {
        topIt->setCoordinates(7, 8);
        topIt->tseg()->setParentIndex(9);
        topIt->tseg()->setParentReversed(false);
        topIt->tseg()->setBottomParseIndex(11);
        if (topIt->getArrayIndex() != 12) {
            topIt->tseg()->setNextParalogyIndex(12);
        } else {
            topIt->tseg()->setNextParalogyIndex(7);
        }
    }
    botIt = copyRootGenome->getBottomSegmentIterator();
    n = copyRootGenome->getNumBottomSegments();
    for (; botIt->getArrayIndex() < n; botIt->toRight()) {
        botIt->setCoordinates(6, 7);
        botIt->bseg()->setChildIndex(0, 8);
        botIt->bseg()->setChildReversed(0, false);
        botIt->bseg()->setTopParseIndex(10);
    }

    ancGenome->copy(copyRootGenome);
    leafGenome->copy(copyLeafGenome);
    _secondAlignment->close();
}
示例#8
0
void MappedSegmentMapExtraParalogsTest::createCallBack(AlignmentPtr alignment)
{
  vector<Sequence::Info> seqVec(1);

  BottomSegmentIteratorPtr bi;
  BottomSegmentStruct bs;
  TopSegmentIteratorPtr ti;
  TopSegmentStruct ts;

  // Set up a case where all the segments of grandChild1 coalesce with
  // the first segment of grandChild2, but only if using the root as
  // the coalescence limit. Otherwise only the first segments map to
  // each other.
  Genome* root = alignment->addRootGenome("root");
  Genome* parent = alignment->addLeafGenome("parent", "root", 1);
  Genome* grandChild1 = alignment->addLeafGenome("grandChild1", "parent", 1);
  Genome* grandChild2 = alignment->addLeafGenome("grandChild2", "parent", 1);
  seqVec[0] = Sequence::Info("Sequence", 3, 0, 1);
  root->setDimensions(seqVec);
  seqVec[0] = Sequence::Info("Sequence", 9, 3, 3);
  parent->setDimensions(seqVec);
  seqVec[0] = Sequence::Info("Sequence", 9, 3, 0);
  grandChild1->setDimensions(seqVec);
  seqVec[0] = Sequence::Info("Sequence", 9, 3, 0);
  grandChild2->setDimensions(seqVec);

  root->setString("CCC");
  parent->setString("CCCTACGTG");
  grandChild1->setString("CCCTACGTG");
  grandChild2->setString("CCCTACGTG");

  bi = root->getBottomSegmentIterator();
  bs.set(0, 3);
  bs._children.push_back(pair<hal_size_t, bool>(0, false));
  bs.applyTo(bi);

  ti = parent->getTopSegmentIterator();
  ts.set(0, 3, 0, false, NULL_INDEX, 1);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(3, 3, 0, false, NULL_INDEX, 2);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(6, 3, 0, false, NULL_INDEX, 0);
  ts.applyTo(ti);

  bi = parent->getBottomSegmentIterator();
  bs.set(0, 3);
  bs._children.clear();
  bs._children.push_back(pair<hal_size_t, bool>(0, true));
  bs._children.push_back(pair<hal_size_t, bool>(0, false));
  bs.applyTo(bi);
  bi->toRight();
  bs.set(3, 3);
  bs._children.clear();
  bs._children.push_back(pair<hal_size_t, bool>(1, true));
  bs._children.push_back(pair<hal_size_t, bool>(NULL_INDEX, true));
  bs.applyTo(bi);
  bi->toRight();
  bs.set(6, 3);
  bs._children.clear();
  bs._children.push_back(pair<hal_size_t, bool>(2, true));
  bs._children.push_back(pair<hal_size_t, bool>(NULL_INDEX, false));
  bs.applyTo(bi);

  ti = grandChild1->getTopSegmentIterator();
  ts.set(0, 3, 0, true);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(3, 3, 1, true);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(6, 3, 2, true);
  ts.applyTo(ti);

  ti = grandChild2->getTopSegmentIterator();
  ts.set(0, 3, 0, false);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(3, 3, NULL_INDEX, true);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(6, 3, NULL_INDEX, false);
  ts.applyTo(ti);

  parent->fixParseInfo();
}