int main(int argc, char *argv[])
{
if (argc == 1) usage(); // Print usage
for (int i=0; i<argc; i++)
{
cerr << argv[i] << ' ';
cerr << endl;
}
Initial_Time();
cerr<<"Start at: "<<Curr_Time()<<endl;
int noptions = mGetOptions(argc, argv);
// Mutltithreads part
#if defined (_OPENMP)
if (param.ncpu) omp_set_num_threads(param.ncpu);
#endif
fin_db.open(ref_file.c_str());
if (!fin_db)
{
cerr << "fatal error: failed to open ref file\n";
exit(1);
}
ref.Run_ConvertBinseq(fin_db);
cerr << "Load in " << ref.total_num << " reference seqs, total size " << ref.sum_length << " bp. " << Cal_AllTime() << " secs passed" << endl;
ref.CreateIndex(); // Qgram_Index();
cerr << "Create refseq k-mer index table. " << Cal_AllTime() << " secs passed" << endl;
RunProcess();
return 0;
}
// Can be either unique or multi-mapping reads
inline void PROBerReadModel_iCLIP::update(AlignmentGroup& ag) {
int size = ag.size();
BamAlignment *ba = NULL;
char dir;
if (size > 1) {
assert(model_type >= 2);
double frac = 1.0 / size;
for (int i = 0; i < size; ++i) {
ba = ag.getAlignment(i);
fld->update(ba->getInsertSize(), frac);
}
return;
}
assert(ag.getSEQ(seq));
if (model_type & 1) assert(ag.getQUAL(qual));
for (int i = 0; i < size; ++i) {
ba = ag.getAlignment(i);
dir = ba->getMateDir();
assert(ba->getCIGAR(cigar));
assert(ba->getMD(mdstr));
refseq.setUp(dir, cigar, mdstr, seq);
seqmodel->update(1.0, dir, 0, &refseq, &cigar, &seq, ((model_type & 1) ? &qual : NULL));
}
if (model_type >= 2) {
assert(ag.getSEQ(seq, 2));
if (model_type & 1) assert(ag.getQUAL(qual, 2));
for (int i = 0; i < size; ++i) {
ba = ag.getAlignment(i);
dir = ba->getMateDir(2);
assert(ba->getCIGAR(cigar, 2));
assert(ba->getMD(mdstr, 2));
refseq.setUp(dir, cigar, mdstr, seq);
seqmodel->update(1.0, dir, 0, &refseq, &cigar, &seq, ((model_type & 1) ? &qual : NULL));
}
}
}
inline void PROBerReadModel_iCLIP::calcProbs(AlignmentGroup& ag, double* conprbs) {
int size = ag.size();
BamAlignment *ba = NULL;
char dir;
assert(ag.getSEQ(seq));
if (model_type & 1) assert(ag.getQUAL(qual));
for (int i = 0; i < size; ++i) {
ba = ag.getAlignment(i);
dir = ba->getMateDir();
assert(ba->getCIGAR(cigar));
assert(ba->getMD(mdstr));
refseq.setUp(dir, cigar, mdstr, seq);
conprbs[i] = seqmodel->getProb(dir, 0, &refseq, &cigar, &seq, ((model_type & 1) ? &qual : NULL));
}
if (model_type >= 2) {
assert(ag.getSEQ(seq, 2));
if (model_type & 1) assert(ag.getQUAL(qual, 2));
for (int i = 0; i < size; ++i) {
ba = ag.getAlignment(i);
dir = ba->getMateDir(2);
assert(ba->getCIGAR(cigar, 2));
assert(ba->getMD(mdstr, 2));
refseq.setUp(dir, cigar, mdstr, seq);
conprbs[i] *= seqmodel->getProb(dir, 0, &refseq, &cigar, &seq, ((model_type & 1) ? &qual : NULL));
conprbs[i] *= fld->getProb(ba->getInsertSize()); // fragment length distribution
}
}
double sum = 0.0;
for (int i = 0; i < size; ++i) sum += conprbs[i];
//assert(sum > 0.0);
if (sum <= 0.0) sum = 1.0;
for (int i = 0; i < size; ++i) conprbs[i] /= sum;
}
void Do_Formatdb()
{
ref.InitialIndex();
pthread_t pids_ab, pids_ac, pids_ad;
//cal kmer freq
pthread_create(&pids_ab, NULL, t_SeedFreq_ab, NULL);
pthread_create(&pids_ac, NULL, t_SeedFreq_ac, NULL);
pthread_create(&pids_ad, NULL, t_SeedFreq_ad, NULL);
pthread_join(pids_ab, NULL);
pthread_join(pids_ac, NULL);
pthread_join(pids_ad, NULL);
ref.AllocIndex();
//record kmer locations
pthread_create(&pids_ab, NULL, t_Index_ab, NULL);
pthread_create(&pids_ac, NULL, t_Index_ac, NULL);
pthread_create(&pids_ad, NULL, t_Index_ad, NULL);
pthread_join(pids_ab, NULL);
pthread_join(pids_ac, NULL);
pthread_join(pids_ad, NULL);
//
ref._blocks.clear();
cout<<"Create seed table. "<<Cal_AllTime()<<" secs passed"<<endl;
};
void SingleAlign::SnpAlign(RefSeq &ref, bit32_t mode) {
bit32_t i,j,m, modeindex, cmodeindex, mc, h, read_chain_index_mask;
Hit prefetch_hit, _hHit;
//cout<<_pread->seq<<endl;
if(param.RRBS_flag){ //RRBS mode
for(read_chain_index=0; read_chain_index<2;read_chain_index++){
if(!xflag_chain[read_chain_index]) continue;
cmodeindex=modeindex=xseedindex[read_chain_index][mode].second;
if(read_chain_index) cmodeindex=map_readlen/param.seed_size-1-modeindex;
_seed=xseeds[read_chain_index][modeindex][0];
//cout<<"mode:"<<mode<<" m="<<ref.index[_seed].n1<<" seed:"<<_seed;
//cout<<" "; disp_bfa(param.map3to4(_seed), param.seed_size); cout<<endl;
if((m=ref.index[_seed].n1)==0) continue;
_refloc=ref.index[_seed].loc1; //list of seeded reference locations
h=param.profile[modeindex][0]+cseed_offset*read_chain_index;
read_chain_index_mask=read_chain_index<<24;
for(j=0; j!=m; j++) {
_hHit=_refloc[j];
if(((_hHit.chr^read_chain_index_mask)>>16)!=cmodeindex) continue; // mode or strand not match
_hit.chr=_hHit.chr&0xffff;
if(_hHit.loc<h) continue; //underflow the start of refseq
_hit.loc=_hHit.loc-h;
//cout<<" j="<<j<<" chr"<<(int)_hit.chr<<":"<<_hit.loc<<endl;
CountMismatch(xseq[read_chain_index], (_hit.loc%SEGLEN)<<1, ref.bfa[_hit.chr].s+_hit.loc/SEGLEN);
//cout<<" mis:"<<tmp_snp<<endl;
if(tmp_snp<=snp_thres) {
_ghit=int2hit(ref,_hit,0,0);
if(!param.pairend){
seg_info=ref.CCGG_seglen(_ghit.chr, _ghit.loc, map_readlen); //get fragment information
//cout<<"seg1:"<<seg_info.first<<" seg2:"<<seg_info.second<<endl;
//if(seg_info.second>param.max_insert) continue; // fragment too large
}
if(AddHit(ref, tmp_snp, mode)) return;
}
if(param.gap>0) {
if(GapAlign(ref, mode, h)) return;
}
}
}
}
else{ //WGBS mode
for(read_chain_index=0; read_chain_index<2;read_chain_index++){
int main(int argc, char *argv[])
{
//print usage
if (argc == 1)
{
usage();
}
Initial_Time();
cout<<"Start at: "<<Curr_Time()<<endl;
int noptions;
noptions=mGetOptions(argc, argv);
fin_db.open(ref_file.c_str());
if(!fin_db) {
cerr<<"fatal error: failed to open ref file\n";
exit(1);
}
ref.Run_ConvertBinseq(fin_db);
cout<<"Load in "<<ref.total_num<<" db seqs, total size "<<ref.sum_length<<" bp. "<<Cal_AllTime()<<" secs passed"<<endl;
//single command:
if(noptions==argc) {
Do_Formatdb();
RunProcess();
}
else {
int old_seed_size=0;
char * margv[1000];
for(int i=0; i<1000; i++) {
margv[i] = new char[1000];
}
char ch[10000];
ifstream fin_batch(argv[noptions]);
while(!fin_batch.eof()) {
fin_batch.getline(ch, 10000);
if(fin_batch.eof())
break;
cout<<"Line of options: "<<ch<<endl;
bool is_word=0;
int margc=0;
char *q=margv[margc];
for(int i=0; ch[i]!='\0'; i++) {
if((ch[i]>=33)&&(ch[i]<=126)) {
if(!is_word) {
*q='\0';
margc++;
q=margv[margc];
}
is_word=1;
*q++=ch[i];
}
else {
is_word=0;
}
}
*q='\0';
margc++;
mGetOptions(margc, margv);
if(param.seed_size!=old_seed_size) {
if(ref.total_kmers>0)
ref.ReleaseIndex();
Do_Formatdb();
old_seed_size=param.seed_size;
}
RunProcess();
}
fin_batch.close();
}
return 0;
}
void Do_Formatdb()
{
ref.CreateIndex();
cout<<"Create seed table. "<<Cal_AllTime()<<" secs passed"<<endl;
};
void *t_Index_ad(void *)
{
ref.t_CreateIndex_ad();
};
void *t_SeedFreq_ad(void *)
{
ref.t_CalKmerFreq_ad();
};