void AssemblyGraph::pointEachNodeToItsReverseComplement()
{
QMapIterator<QString, DeBruijnNode*> i(m_deBruijnGraphNodes);
while (i.hasNext())
{
i.next();
DeBruijnNode * positiveNode = i.value();
if (positiveNode->isPositiveNode())
{
DeBruijnNode * negativeNode = m_deBruijnGraphNodes[getOppositeNodeName(positiveNode->getName())];
if (negativeNode != 0)
{
positiveNode->setReverseComplement(negativeNode);
negativeNode->setReverseComplement(positiveNode);
}
}
}
}
void AssemblyGraph::buildDeBruijnGraphFromGfa(QString fullFileName)
{
m_graphFileType = GFA;
QFile inputFile(fullFileName);
if (inputFile.open(QIODevice::ReadOnly))
{
std::vector<QString> edgeStartingNodeNames;
std::vector<QString> edgeEndingNodeNames;
std::vector<int> edgeOverlaps;
QTextStream in(&inputFile);
while (!in.atEnd())
{
QApplication::processEvents();
QString line = in.readLine();
QStringList lineParts = line.split(QRegExp("\t"));
if (lineParts.size() < 1)
continue;
//Lines beginning with "S" are sequence (node) lines
if (lineParts.at(0) == "S")
{
if (lineParts.size() < 3)
throw "load error";
QString nodeName = lineParts.at(1);
QString posNodeName = nodeName + "+";
QString negNodeName = nodeName + "-";
QByteArray sequence = lineParts.at(2).toLocal8Bit();
QByteArray revCompSequence = getReverseComplement(sequence);
//If there is an attribute holding the read depth, we'll use that.
//If there isn't, then we'll use zero.
double nodeReadDepth = 0.0;
for (int i = 3; i < lineParts.size(); ++i)
{
QString part = lineParts.at(i);
if (part.size() < 6)
continue;
else if (part.left(5) == "KC:f:")
nodeReadDepth = part.right(part.length() - 5).toDouble();
}
DeBruijnNode * node = new DeBruijnNode(posNodeName, nodeReadDepth, sequence);
DeBruijnNode * reverseComplementNode = new DeBruijnNode(negNodeName, nodeReadDepth, revCompSequence);
node->setReverseComplement(reverseComplementNode);
reverseComplementNode->setReverseComplement(node);
m_deBruijnGraphNodes.insert(posNodeName, node);
m_deBruijnGraphNodes.insert(negNodeName, reverseComplementNode);
}
//Lines beginning with "L" are link (edge) lines
else if (lineParts.at(0) == "L")
{
//Edges aren't made now, in case their sequence hasn't yet been specified.
//Instead, we save the starting and ending nodes and make the edges after
//we're done looking at the file.
if (lineParts.size() < 6)
throw "load error";
//Parts 1 and 3 hold the node names and parts 2 and 4 hold the corresponding +/-.
QString startingNode = lineParts.at(1) + lineParts.at(2);
QString endingNode = lineParts.at(3) + lineParts.at(4);
edgeStartingNodeNames.push_back(startingNode);
edgeEndingNodeNames.push_back(endingNode);
//Part 5 holds the node overlap cigar string
QString cigar = lineParts.at(5);
edgeOverlaps.push_back(getLengthFromCigar(cigar));
}
}
//Create all of the edges
for (size_t i = 0; i < edgeStartingNodeNames.size(); ++i)
{
QString node1Name = edgeStartingNodeNames[i];
QString node2Name = edgeEndingNodeNames[i];
int overlap = edgeOverlaps[i];
createDeBruijnEdge(node1Name, node2Name, overlap);
}
}
if (m_deBruijnGraphNodes.size() == 0)
throw "load error";
}
void AssemblyGraph::buildDeBruijnGraphFromLastGraph(QString fullFileName)
{
m_graphFileType = LAST_GRAPH;
bool firstLine = true;
QFile inputFile(fullFileName);
if (inputFile.open(QIODevice::ReadOnly))
{
QTextStream in(&inputFile);
while (!in.atEnd())
{
QApplication::processEvents();
QString line = in.readLine();
if (firstLine)
{
QStringList firstLineParts = line.split(QRegExp("\\s+"));
if (firstLineParts.size() > 2)
m_kmer = firstLineParts[2].toInt();
firstLine = false;
}
if (line.startsWith("NODE"))
{
QStringList nodeDetails = line.split(QRegExp("\\s+"));
if (nodeDetails.size() < 4)
throw "load error";
QString nodeName = nodeDetails.at(1);
QString posNodeName = nodeName + "+";
QString negNodeName = nodeName + "-";
int nodeLength = nodeDetails.at(2).toInt();
double nodeReadDepth;
if (nodeLength > 0)
nodeReadDepth = double(nodeDetails.at(3).toInt()) / nodeLength; //IS THIS COLUMN ($COV_SHORT1) THE BEST ONE TO USE?
else
nodeReadDepth = double(nodeDetails.at(3).toInt());
QByteArray sequence = in.readLine().toLocal8Bit();
QByteArray revCompSequence = in.readLine().toLocal8Bit();
DeBruijnNode * node = new DeBruijnNode(posNodeName, nodeReadDepth, sequence);
DeBruijnNode * reverseComplementNode = new DeBruijnNode(negNodeName, nodeReadDepth, revCompSequence);
node->setReverseComplement(reverseComplementNode);
reverseComplementNode->setReverseComplement(node);
m_deBruijnGraphNodes.insert(posNodeName, node);
m_deBruijnGraphNodes.insert(negNodeName, reverseComplementNode);
}
else if (line.startsWith("ARC"))
{
QStringList arcDetails = line.split(QRegExp("\\s+"));
if (arcDetails.size() < 3)
throw "load error";
QString node1Name = convertNormalNumberStringToBandageNodeName(arcDetails.at(1));
QString node2Name = convertNormalNumberStringToBandageNodeName(arcDetails.at(2));
createDeBruijnEdge(node1Name, node2Name);
}
}
inputFile.close();
setAllEdgesExactOverlap(0);
}
if (m_deBruijnGraphNodes.size() == 0)
throw "load error";
}
