void Clustal::writeAlignment(std::ostream& output, const SiteContainer& sc) const throw (Exception)
{
output << "CLUSTAL W (1.81) multiple sequence alignment" << endl;
output << endl;
if (sc.getNumberOfSequences() == 0)
return;
vector<string> text;
size_t length = 0;
for (size_t i = 0; i < sc.getNumberOfSequences(); ++i ) {
const Sequence& seq = sc.getSequence(i);
if (seq.getName().size() > length)
length = seq.getName().size();
text.push_back(sc.getSequence(i).toString());
}
length += nbSpacesBeforeSeq_;
for (unsigned int j = 0; j < text[0].size(); j += charsByLine_) {
for (unsigned int i = 0; i < sc.getNumberOfSequences(); ++i ) {
output << TextTools::resizeRight(sc.getSequence(i).getName(), length);
output << text[i].substr(j, charsByLine_) << endl;
}
output << endl;
}
}
void SiteContainerTools::getSequencePositions(const SiteContainer& sites, Matrix<size_t>& positions)
{
positions.resize(sites.getNumberOfSequences(), sites.getNumberOfSites());
int gap = sites.getAlphabet()->getGapCharacterCode();
for (size_t i = 0; i < sites.getNumberOfSequences(); ++i) {
const Sequence& seq = sites.getSequence(i);
unsigned int pos = 0;
for (size_t j = 0; j < sites.getNumberOfSites(); ++j) {
if (seq[j] != gap) {
++pos;
positions(i, j) = pos;
} else {
positions(i, j) = 0;
}
}
}
}
std::map<size_t, size_t> SiteContainerTools::translateSequence(const SiteContainer& sequences, size_t i1, size_t i2)
{
const Sequence* seq1 = &sequences.getSequence(i1);
const Sequence* seq2 = &sequences.getSequence(i2);
map<size_t, size_t> tln;
size_t count1 = 0; // Sequence 1 counter
size_t count2 = 0; // Sequence 2 counter
int state1;
int state2;
for (size_t i = 0; i < sequences.getNumberOfSites(); i++)
{
state1 = (*seq1)[i];
if (state1 != -1)
count1++;
state2 = (*seq2)[i];
if (state2 != -1)
count2++;
if (state1 != -1)
{
tln[count1] = (state2 == -1 ? 0 : count2);
}
}
return tln;
}
void RecursiveLikelihoodTree::initLikelihoodsWithoutPatterns_(const Node* node, const SiteContainer& sequences, const SubstitutionProcess& process) throw (Exception)
{
int nId = node->getId();
// Initialize likelihood vector:
if (!node->hasFather())
{
resetAboveLikelihoods(nId, nbDistinctSites_, nbStates_);
resetLikelihoods(nId, nbDistinctSites_, nbStates_, ComputingNode::D0);
resetLikelihoods(nId, nbDistinctSites_, nbStates_, ComputingNode::D1);
resetLikelihoods(nId, nbDistinctSites_, nbStates_, ComputingNode::D2);
}
resetBelowLikelihoods(nId, nbDistinctSites_, nbStates_, ComputingNode::D0);
resetBelowLikelihoods(nId, nbDistinctSites_, nbStates_, ComputingNode::D1);
resetBelowLikelihoods(nId, nbDistinctSites_, nbStates_, ComputingNode::D2);
// Now initialize likelihood values and pointers:
if (node->hasNoSon())
{
const Sequence* seq;
try
{
seq = &sequences.getSequence(node->getName());
}
catch (SequenceNotFoundException snfe)
{
throw SequenceNotFoundException("RecursiveLikelihoodTree::initTreelikelihoods. Leaf name in tree not found in site conainer: ", (node->getName()));
}
for (size_t c = 0; c < nbClasses_; c++)
{
RecursiveLikelihoodNode& lNode = *dynamic_cast<RecursiveLikelihoodNode*>(vTree_[c]->getNode(nId));
VVdouble& array = lNode.getBelowLikelihoodArray_(ComputingNode::D0);
for (size_t i = 0; i < nbDistinctSites_; i++)
{
Vdouble* array_i = &array[i];
int state = seq->getValue(i);
double test = 0.;
for (size_t s = 0; s < nbStates_; s++)
{
double x = process.getInitValue(s, state);
if (lNode.usesLog())
{
if (x <= 0)
(*array_i)[s] = -10000;
else
(*array_i)[s] = log(x);
}
else
(*array_i)[s] = x;
test += x;
}
if (test < 0.000001)
std::cerr << "WARNING!!! Likelihood will be 0 for this site " << TextTools::toString(i) << std::endl;
}
lNode.updateBelow_(true, ComputingNode::D0);
}
}
else
{
// 'node' is an internal node.
std::map<int, std::vector<size_t> >* patternLinks_node = &patternLinks_[nId];
int nbSonNodes = static_cast<int>(node->getNumberOfSons());
for (int l = 0; l < nbSonNodes; ++l)
{
// For each son node,
const Node* son = (*node)[l];
initLikelihoodsWithoutPatterns_(son, sequences, process);
std::vector<size_t>* patternLinks_node_son = &(*patternLinks_node)[son->getId()];
// Init map:
patternLinks_node_son->resize(nbDistinctSites_);
for (size_t i = 0; i < nbDistinctSites_; i++)
{
(*patternLinks_node_son)[i] = i;
}
}
}
if (!node->hasFather())
setAboveLikelihoods(nId, process.getRootFrequencies());
}