void CGeneFinder::CGeneSearchPlugin::setUpFeatureIterator (
CBioseq_Handle &ignored_bioseq_handle,
auto_ptr<CFeat_CI> &feat_ci,
TSeqPos circular_length,
CRange<TSeqPos> &range,
const CSeq_loc& loc,
SAnnotSelector &sel,
CScope &scope,
ENa_strand &strand )
{
if ( m_BioseqHandle ) {
// if we're circular, we may need to split our range into two pieces
if( ( circular_length != kInvalidSeqPos ) &&
( range.GetFrom() > range.GetTo() ))
{
// For circular locations, the "from" is greater than the "to", which
// would not work properly if given to CFeat_CI.
// So, as a work around, we transform the range
// into a mix location of the form "join(0..to, from..MAXINT)"
CRef<CSeq_loc> new_loc( new CSeq_loc );
new_loc->SetInt().SetFrom( 0 );
new_loc->SetInt().SetTo( range.GetTo() );
CRef<CSeq_loc> otherHalfOfRange( new CSeq_loc );
otherHalfOfRange->SetInt().SetFrom( range.GetFrom() );
otherHalfOfRange->SetInt().SetTo( kMax_Int );
new_loc->Add( *otherHalfOfRange );
new_loc->SetStrand( loc.GetStrand() );
new_loc->SetId( *loc.GetId() );
feat_ci.reset( new CFeat_CI(scope, *new_loc, sel) );
} else {
// remove far parts, if necessary
bool loc_change_needed = false;
ITERATE( CSeq_loc, loc_iter, loc ) {
if( ! m_BioseqHandle.IsSynonym( loc_iter.GetSeq_id() ) ) {
loc_change_needed = true;
break;
}
}
if( loc_change_needed ) {
CRef<CSeq_loc> new_loc( new CSeq_loc );
ITERATE( CSeq_loc, loc_iter, loc ) {
if( m_BioseqHandle.IsSynonym( loc_iter.GetSeq_id() ) ) {
new_loc->Add( *loc_iter.GetRangeAsSeq_loc() );
}
}
feat_ci.reset( new CFeat_CI(scope, *new_loc, sel) );
} else {
feat_ci.reset( new CFeat_CI(scope, loc, sel) );
}
}
} else {
pair<double, bool> CScoreUniqSeqCoverage::MakeScore(CBioseq_Handle const& query_handle, vector<CSeq_align const*>::const_iterator begin, vector<CSeq_align const*>::const_iterator end)
{
CConstRef<CBioseq> bioseq = query_handle.GetCompleteBioseq();
unsigned int qlen = 0;
if ( !bioseq.Empty() && bioseq->IsSetLength()) {
qlen = bioseq->GetLength();
}
if ( !qlen ) {
return make_pair(0, false);
}
bool isDenDiag = ( (*begin)->GetSegs().Which() == CSeq_align::C_Segs::e_Dendiag) ?
true : false;
CRangeCollection<TSeqPos> subj_rng_coll((*begin)->GetSeqRange(1));
CRange<TSeqPos> q_rng((*begin)->GetSeqRange(0));
CRangeCollection<TSeqPos> query_rng_coll(s_FixMinusStrandRange(q_rng));
for( ++begin; begin != end; ++begin ) {
const CRange<TSeqPos> align_subj_rng((*begin)->GetSeqRange(1));
// subject range should always be on the positive strand
assert(align_subj_rng.GetTo() > align_subj_rng.GetFrom());
CRangeCollection<TSeqPos> coll(align_subj_rng);
coll.Subtract(subj_rng_coll);
if ( coll.empty() ) {
continue;
}
if(coll[0] == align_subj_rng) {
CRange<TSeqPos> query_rng ((*begin)->GetSeqRange(0));
query_rng_coll += s_FixMinusStrandRange(query_rng);
subj_rng_coll += align_subj_rng;
}
else {
ITERATE (CRangeCollection<TSeqPos>, uItr, coll) {
CRange<TSeqPos> query_rng;
const CRange<TSeqPos> & subj_rng = (*uItr);
CRef<CSeq_align> densegAln;
if ( isDenDiag) {
densegAln = CreateDensegFromDendiag(**begin);
}
CAlnMap map( (isDenDiag) ? densegAln->GetSegs().GetDenseg() : (*begin)->GetSegs().GetDenseg());
TSignedSeqPos subj_aln_start = map.GetAlnPosFromSeqPos(1,subj_rng.GetFrom());
TSignedSeqPos subj_aln_end = map.GetAlnPosFromSeqPos(1,subj_rng.GetTo());
query_rng.SetFrom(map.GetSeqPosFromAlnPos(0,subj_aln_start));
query_rng.SetTo(map.GetSeqPosFromAlnPos(0,subj_aln_end));
query_rng_coll += s_FixMinusStrandRange(query_rng);
subj_rng_coll += subj_rng;
}
}
}
CRef<CSeq_loc> CFeatTrim::Apply(const CSeq_loc& loc,
const CRange<TSeqPos>& range)
{
const bool set_partial = true;
const TSeqPos from = range.GetFrom();
const TSeqPos to = range.GetTo();
CRef<CSeq_loc> trimmed_loc(new CSeq_loc());
trimmed_loc->Assign(loc);
x_TrimLocation(from, to, set_partial, trimmed_loc);
return trimmed_loc;
}
CRef<CSeq_feat> CFeatTrim::Apply(const CSeq_feat& feat,
const CRange<TSeqPos>& range)
{
CRef<CSeq_loc> loc = Ref(new CSeq_loc());
loc->Assign(feat.GetLocation());
const TSeqPos from = range.GetFrom();
const TSeqPos to = range.GetTo();
const bool set_partial = true;
x_TrimLocation(from, to, set_partial, loc);
if (loc->IsNull()) {
return Ref(new CSeq_feat());
}
// Create a new seq-feat with the trimmed location
CRef<CSeq_feat> new_sf(new CSeq_feat());
new_sf->Assign(feat);
new_sf->SetLocation(*loc);
if (!loc->IsNull() &&
(loc->IsPartialStart(eExtreme_Biological) ||
loc->IsPartialStop(eExtreme_Biological))) {
new_sf->SetPartial(true);
}
// If Cdregion need to consider changes in frameshift
if (new_sf->GetData().IsCdregion()) {
const TSeqPos offset = x_GetStartOffset(feat, from, to);
x_UpdateFrame(offset, new_sf->SetData().SetCdregion());
if (new_sf->SetData().SetCdregion().IsSetCode_break()) {
// iterate over code breaks and remove if they fall outside the range
list<CRef<CCode_break>>& code_breaks = new_sf->SetData().SetCdregion().SetCode_break();
code_breaks.remove_if(SOutsideRange(from,to));
if (code_breaks.empty()) {
new_sf->SetData().SetCdregion().ResetCode_break();
}
else {
const auto strand = loc->GetStrand();
// Trim the 3' end
if (strand != eNa_strand_minus) {
for (auto code_break : code_breaks) {
const TSeqPos cb_to = code_break->GetLoc().GetTotalRange().GetTo();
if (cb_to > to) {
x_TrimCodeBreak(from, to, *code_break);
}
}
}
else { // strand == eNa_strand_minus
for (auto code_break : code_breaks) {
const TSeqPos cb_from = code_break->GetLoc().GetTotalRange().GetFrom();
if (cb_from < from) {
x_TrimCodeBreak(from, to, *code_break);
}
}
}
}
}
}
else
if (new_sf->GetData().GetSubtype() == CSeqFeatData::eSubtype_tRNA) {
auto& rna = new_sf->SetData().SetRna();
if (rna.IsSetExt() && rna.GetExt().IsTRNA()) {
x_TrimTrnaExt(from, to, rna.SetExt().SetTRNA());
}
}
return new_sf;
}