int main(int argc, const char** argv)
{
usage(argc, argv);
string read = argv[1];
string germline = argv[2];
Fasta Vgenes(germline+"V.fa", 2, "|");
Fasta Jgenes(germline+"J.fa", 2, "|");
Fasta interestingV = extractInterestingGenes(Vgenes, argv[3]);
Fasta interestingJ = extractInterestingGenes(Jgenes, argv[4]);
if (interestingV.size() == 0) {
cerr << "No interesting V found" << endl;
exit(2);
}
if (interestingJ.size() == 0) {
cerr << "No interesting J found" << endl;
exit(2);
}
AlignBox box_V("5", V_COLOR);
AlignBox box_J("3", J_COLOR);
if (read == "-") {
// Read on stdin
read = read_sequence(cin);
}
align_against_collection(read, interestingV, -1, false, false, false, &box_V, VDJ);
align_against_collection(read, interestingJ, -1, false, true, false, &box_J, VDJ);
// This should be handled directly into align_against_collection
box_J.start = box_J.end ;
box_J.del_left = box_J.del_right;
box_J.end = read.size() - 1;
int align_V_length = min(GENE_ALIGN, box_V.end - box_V.start + 1);
int align_J_length = min(GENE_ALIGN, (int)read.size() - box_J.start + 1);
int start_V = box_V.end - align_V_length + 1;
int end_J = box_J.start + align_J_length - 1;
cout << "read \t" << start_V << "\t" ;
cout << V_COLOR << read.substr(start_V, align_V_length)
<< NO_COLOR
<< read.substr(box_V.end+1, (box_J.start - 1) - (box_V.end + 1) +1)
<< J_COLOR
<< read.substr(box_J.start, align_J_length)
<< NO_COLOR
<< "\t" << end_J << endl ;
cout << box_V.refToString(start_V, end_J) << "\t" << box_V << endl ;
cout << box_J.refToString(start_V, end_J) << "\t" << box_J << endl ;
exit (0);
}
Fasta extractInterestingGenes(Fasta &repertoire, string name) {
Fasta interesting;
int size = repertoire.size();
for (int i = 0; i < size; i++) {
if (repertoire.label(i).find(name) != string::npos) {
interesting.add(repertoire.read(i));
}
}
return interesting;
}
void align_against_collection(string &read, Fasta &rep, int forbidden_rep_id,
bool reverse_ref, bool reverse_both, bool local,
AlignBox *box, Cost segment_cost)
{
int best_score = MINUS_INF ;
box->ref_nb = MINUS_INF ;
int best_best_i = (int) string::npos ;
int best_best_j = (int) string::npos ;
int best_first_i = (int) string::npos ;
int best_first_j = (int) string::npos ;
vector<pair<int, int> > score_r;
DynProg::DynProgMode dpMode = DynProg::LocalEndWithSomeDeletions;
if (local==true) dpMode = DynProg::Local;
// With reverse_ref, the read is reversed to prevent calling revcomp on each reference sequence
string sequence_or_rc = revcomp(read, reverse_ref);
for (int r = 0 ; r < rep.size() ; r++)
{
if (r == forbidden_rep_id)
continue;
DynProg dp = DynProg(sequence_or_rc, rep.sequence(r),
dpMode, // DynProg::SemiGlobalTrans,
segment_cost, // DNA
reverse_both, reverse_both,
rep.read(r).marked_pos);
bool onlyBottomTriangle = !local ;
int score = dp.compute(onlyBottomTriangle, BOTTOM_TRIANGLE_SHIFT);
if (local==true){
dp.backtrack();
}
if (score > best_score)
{
best_score = score ;
best_best_i = dp.best_i ;
best_best_j = dp.best_j ;
best_first_i = dp.first_i ;
best_first_j = dp.first_j ;
box->ref_nb = r ;
box->ref_label = rep.label(r) ;
if (!local)
dp.backtrack();
box->marked_pos = dp.marked_pos_i ;
}
score_r.push_back(make_pair(score, r));
// #define DEBUG_SEGMENT
#ifdef DEBUG_SEGMENT
cout << rep.label(r) << " " << score << " " << dp.best_i << endl ;
#endif
}
sort(score_r.begin(),score_r.end(),comp_pair);
box->ref = rep.sequence(box->ref_nb);
box->del_right = reverse_both ? best_best_j : box->ref.size() - best_best_j - 1;
box->del_left = best_first_j;
box->start = best_first_i;
box->score = score_r;
#ifdef DEBUG_SEGMENT
cout << "best: " << box->ref_label << " " << best_score ;
cout << "del/del2/begin:" << (box->del_right) << "/" << (box->del_left) << "/" << (box->start) << endl;
cout << endl;
#endif
if (reverse_ref)
// Why -1 here and +1 in dynprog.cpp /// best_i = m - best_i + 1 ;
best_best_i = read.length() - best_best_i - 1 ;
box->end = best_best_i ;
}