double FONSEModel::calculateLogLikelihoodRatioPerAA(Gene& gene, std::string grouping, double *mutation, double *selection, double phiValue)
{
int numCodons = SequenceSummary::GetNumCodonsForAA(grouping);
double logLikelihood = 0.0;
std::vector <unsigned> *positions;
double codonProb[6];
unsigned maxIndexVal = 0u;
for (int i = 1; i < (numCodons - 1); i++)
{
if (selection[maxIndexVal] < selection[i])
{
maxIndexVal = i;
}
}
unsigned aaStart, aaEnd;
SequenceSummary::AAToCodonRange(grouping, aaStart, aaEnd, false);
for (unsigned i = aaStart, k = 0; i < aaEnd; i++, k++) {
positions = gene.geneData.getCodonPositions(i);
for (unsigned j = 0; j < positions->size(); j++) {
calculateCodonProbabilityVector(numCodons, positions->at(j), maxIndexVal, mutation, selection, phiValue, codonProb);
if (codonProb[k] == 0) continue;
logLikelihood += std::log(codonProb[k]);
}
//positions->clear();
}
return logLikelihood;
}
void FONSEModel::simulateGenome(Genome & genome)
{
unsigned codonIndex;
std::string curAA;
std::string tmpDesc = "Simulated Gene";
for (unsigned geneIndex = 0; geneIndex < genome.getGenomeSize(); geneIndex++) //loop over all genes in the genome
{
if (geneIndex % 100 == 0) my_print("Simulating Gene %\n", geneIndex);
Gene gene = genome.getGene(geneIndex);
SequenceSummary sequenceSummary = gene.geneData;
std::string tmpSeq = "ATG"; //Always will have the start amino acid
unsigned mixtureElement = getMixtureAssignment(geneIndex);
unsigned mutationCategory = getMutationCategory(mixtureElement);
unsigned selectionCategory = getSelectionCategory(mixtureElement);
unsigned synthesisRateCategory = getSynthesisRateCategory(mixtureElement);
double phi = getSynthesisRate(geneIndex, synthesisRateCategory, false);
std::string geneSeq = gene.getSequence();
for (unsigned position = 1; position < (geneSeq.size() / 3); position++)
{
std::string codon = geneSeq.substr((position * 3), 3);
curAA = SequenceSummary::codonToAA(codon);
//TODO: Throw an error here instead
if (curAA == "X") {
if (position < (geneSeq.size() / 3) - 1) my_print("Warning: Internal stop codon found in gene % at position %. Ignoring and moving on.\n", gene.getId(), position);
continue;
}
unsigned numCodons = SequenceSummary::GetNumCodonsForAA(curAA);
double* codonProb = new double[numCodons](); //size the arrays to the proper size based on # of codons.
double* mutation = new double[numCodons - 1]();
double* selection = new double[numCodons - 1]();
if (curAA == "M" || curAA == "W")
{
codonProb[0] = 1;
}
else
{
getParameterForCategory(mutationCategory, FONSEParameter::dM, curAA, false, mutation);
getParameterForCategory(selectionCategory, FONSEParameter::dOmega, curAA, false, selection);
calculateCodonProbabilityVector(numCodons, position, mutation, selection, phi, codonProb);
}
codonIndex = Parameter::randMultinom(codonProb, numCodons);
unsigned aaStart, aaEnd;
SequenceSummary::AAToCodonRange(curAA, aaStart, aaEnd, false); //need the first spot in the array where the codons for curAA are
codon = sequenceSummary.indexToCodon(aaStart + codonIndex);//get the correct codon based off codonIndex
tmpSeq += codon;
}
std::string codon = sequenceSummary.indexToCodon((unsigned)Parameter::randUnif(61.0, 64.0)); //randomly choose a stop codon, from range 61-63
tmpSeq += codon;
Gene simulatedGene(tmpSeq, tmpDesc, gene.getId());
genome.addGene(simulatedGene, true);
}
}