// ----------------------------------------------------------------------------
bool CGtfReader::x_FeatureSetDataCDS(
const CGff2Record& record,
CRef< CSeq_feat > pFeature )
// ----------------------------------------------------------------------------
{
if ( ! CGff2Reader::x_FeatureSetDataCDS( record, pFeature ) ) {
return false;
}
CCdregion& cdr = pFeature->SetData().SetCdregion();
string strValue;
if ( record.GetAttribute( "protein_id", strValue ) ) {
CRef<CSeq_id> pId = CReadUtil::AsSeqId(strValue,m_iFlags);
if (pId->IsGenbank()) {
pFeature->SetProduct().SetWhole(*pId);
}
}
if ( record.GetAttribute( "ribosomal_slippage", strValue ) ) {
pFeature->SetExcept( true );
pFeature->SetExcept_text( "ribosomal slippage" );
}
if ( record.GetAttribute( "transl_table", strValue ) ) {
CRef< CGenetic_code::C_E > pGc( new CGenetic_code::C_E );
pGc->SetId( NStr::StringToUInt( strValue ) );
cdr.SetCode().Set().push_back( pGc );
}
return true;
}
// ----------------------------------------------------------------------------
bool CGtfReader::xFeatureSetQualifiersGene(
const CGff2Record& record,
CRef< CSeq_feat > pFeature )
// ----------------------------------------------------------------------------
{
//
// Create GB qualifiers for the record attributes:
//
CRef< CGb_qual > pQual(0);
const CGff2Record::TAttributes& attrs = record.Attributes();
CGff2Record::TAttrCit it = attrs.begin();
for (/*NOOP*/; it != attrs.end(); ++it) {
// gtf genes don't get transcript_id
if (it->first == "transcript_id") {
continue;
}
// special case some well-known attributes
if (x_ProcessQualifierSpecialCase(it, pFeature)) {
continue;
}
// turn everything else into a qualifier
pQual.Reset(new CGb_qual);
pQual->SetQual(it->first);
pQual->SetVal(it->second);
pFeature->SetQual().push_back(pQual);
}
return true;
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_MergeFeatureLocationSingleInterval(
const CGff2Record& record,
CRef< CSeq_feat > pFeature )
// ----------------------------------------------------------------------------
{
const CSeq_interval& gene_int = pFeature->GetLocation().GetInt();
if ( gene_int.GetFrom() > record.SeqStart() -1 ) {
pFeature->SetLocation().SetInt().SetFrom( record.SeqStart() );
}
if ( gene_int.GetTo() < record.SeqStop() - 1 ) {
pFeature->SetLocation().SetInt().SetTo( record.SeqStop() );
}
if (record.Type() == "CDS" && pFeature->GetData().IsCdregion()) {
return x_FeatureTrimQualifiers(record, pFeature);
}
return true;
}
// ----------------------------------------------------------------------------
string s_FeatureKey(
const CGff2Record& gff )
// ----------------------------------------------------------------------------
{
string strGeneId = s_GeneKey( gff );
if ( gff.Type() == "gene" ) {
return strGeneId;
}
string strTranscriptId;
if ( ! gff.GetAttribute( "transcript_id", strTranscriptId ) ) {
cerr << "Unexpected: GTF feature without a transcript_id." << endl;
strTranscriptId = "transcript_id";
}
return strGeneId + "|" + strTranscriptId;
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_CreateFeatureLocation(
const CGff2Record& record,
CRef< CSeq_feat > pFeature )
// ----------------------------------------------------------------------------
{
CRef<CSeq_id> pId = CReadUtil::AsSeqId(
record.Id(), m_iFlags & fAllIdsAsLocal);
CSeq_interval& location = pFeature->SetLocation().SetInt();
location.SetId( *pId );
location.SetFrom( record.SeqStart() );
if (record.Type() != "mRNA") {
location.SetTo(record.SeqStop());
}
else {
// place holder
// actual location will be computed from the exons and CDSs living on
// this feature.
location.SetTo(record.SeqStart());
}
if ( record.IsSetStrand() ) {
location.SetStrand( record.Strand() );
}
return true;
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_MergeFeatureLocationMultiInterval(
const CGff2Record& record,
CRef< CSeq_feat > pFeature )
// ----------------------------------------------------------------------------
{
CRef<CSeq_id> pId = CReadUtil::AsSeqId(
record.Id(), m_iFlags & fAllIdsAsLocal);
CRef< CSeq_loc > pLocation( new CSeq_loc );
pLocation->SetInt().SetId( *pId );
pLocation->SetInt().SetFrom( record.SeqStart() );
pLocation->SetInt().SetTo( record.SeqStop() );
if ( record.IsSetStrand() ) {
pLocation->SetInt().SetStrand( record.Strand() );
}
pLocation = pLocation->Add(
pFeature->SetLocation(), CSeq_loc::fSortAndMerge_All, 0 );
pFeature->SetLocation( *pLocation );
return true;
}
// ----------------------------------------------------------------------------
string s_GeneKey(
const CGff2Record& gff )
// ----------------------------------------------------------------------------
{
string strGeneId;
if ( ! gff.GetAttribute( "gene_id", strGeneId ) ) {
cerr << "Unexpected: GTF feature without a gene_id." << endl;
return "gene_id";
}
return strGeneId;
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_CdsIsPartial(
const CGff2Record& record )
// ----------------------------------------------------------------------------
{
string strPartial;
// if ( record.Type() != "CDS" ) {
// return false;
// }
if ( record.GetAttribute( "partial", strPartial ) ) {
return true;
}
CRef< CSeq_feat > mRna;
if ( ! x_FindParentMrna( record, mRna ) ) {
return false;
}
return ( mRna->IsSetPartial() && mRna->GetPartial() );
}
// -----------------------------------------------------------------------------
bool CGtfReader::x_CreateParentCds(
const CGff2Record& gff,
CRef< CSeq_annot > pAnnot )
// -----------------------------------------------------------------------------
{
//
// Create a single cds feature.
// This creation may either be triggered by an actual CDS feature found in the
// gtf, or by a feature that would imply a CDS feature (such as a start codon
// or a stop codon). The latter is necessary because nothing the the gtf
// standard stipulates that gtf features have to be arranged in any particular
// order.
//
CRef< CSeq_feat > pFeature( new CSeq_feat );
string strType = gff.Type();
if ( strType != "CDS" && strType != "start_codon" && strType != "stop_codon" ) {
return false;
}
if ( ! x_FeatureSetDataCDS( gff, pFeature ) ) {
return false;
}
if ( ! x_CreateFeatureLocation( gff, pFeature ) ) {
return false;
}
if ( ! x_CreateFeatureId( gff, "cds", pFeature ) ) {
return false;
}
if ( ! x_CreateGeneXrefs( gff, pFeature ) ) {
return false;
}
if ( ! x_CreateMrnaXrefs( gff, pFeature ) ) {
return false;
}
if ( ! x_FeatureSetQualifiers( gff, pFeature ) ) {
return false;
}
m_CdsMap[ s_FeatureKey( gff ) ] = pFeature;
return xAddFeatureToAnnot( pFeature, pAnnot );
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_FeatureSetDataMRNA(
const CGff2Record& record,
CRef<CSeq_feat> pFeature)
// ----------------------------------------------------------------------------
{
if ( ! CGff2Reader::x_FeatureSetDataRna(
record, pFeature, CSeqFeatData::eSubtype_mRNA)) {
return false;
}
CRNA_ref& rna = pFeature->SetData().SetRna();
string strValue;
if (record.GetAttribute("product", strValue)) {
rna.SetExt().SetName(strValue);
}
return true;
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_FeatureSetDataGene(
const CGff2Record& record,
CRef< CSeq_feat > pFeature )
// ----------------------------------------------------------------------------
{
if ( ! CGff2Reader::x_FeatureSetDataGene( record, pFeature ) ) {
return false;
}
CGene_ref& gene = pFeature->SetData().SetGene();
string strValue;
if ( record.GetAttribute( "gene_synonym", strValue ) ) {
gene.SetSyn().push_back( strValue );
}
// mss-399: do -not- use gene_id for /gene_syn or /gene:
//if ( record.GetAttribute( "gene_id", strValue ) ) {
// gene.SetSyn().push_front( strValue );
//}
return true;
}
// ----------------------------------------------------------------------------
bool CGff3Reader::x_UpdateAnnot(
const CGff2Record& record,
CRef< CSeq_annot > pAnnot )
// ----------------------------------------------------------------------------
{
string gbkey;
record.GetAttribute("gbkey", gbkey);
CRef< CSeq_feat > pFeature(new CSeq_feat);
// Round trip info:
CRef< CGb_qual > pQual( new CGb_qual );
pQual->SetQual( "gff_source" );
pQual->SetVal( record.Source() );
pFeature->SetQual().push_back( pQual );
pQual.Reset( new CGb_qual );
pQual->SetQual( "gff_type" );
pQual->SetVal( record.Type() );
pFeature->SetQual().push_back( pQual );
if ( record.IsSetScore() ) {
pQual.Reset( new CGb_qual );
pQual->SetQual( "gff_score" );
pQual->SetVal( NStr::DoubleToString( record.Score() ) );
pFeature->SetQual().push_back( pQual );
}
// Special case: exon feature belonging to an RNA we have already seen
if (record.Type() == "exon") {
string parent;
if (record.GetAttribute("Parent", parent)) {
IdToFeatureMap::iterator it = m_MapIdToFeature.find(parent);
if (it != m_MapIdToFeature.end()) {
return record.UpdateFeature(m_iFlags, it->second);
}
}
}
// Special case: Piece of another feature we have already seen
string id;
if (record.GetAttribute("ID", id)) {
IdToFeatureMap::iterator it = m_MapIdToFeature.find(id);
if (it != m_MapIdToFeature.end()) {
return record.UpdateFeature(m_iFlags, it->second);
}
}
// General case: brand new regular feature
if (!record.InitializeFeature(m_iFlags, pFeature)) {
return false;
}
string strId;
if ( record.GetAttribute( "ID", strId ) ) {
m_MapIdToFeature[ strId ] = pFeature;
}
return x_AddFeatureToAnnot( pFeature, pAnnot );
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_UpdateAnnotCds(
const CGff2Record& gff,
CRef< CSeq_annot > pAnnot )
// ----------------------------------------------------------------------------
{
//
// If there is no gene feature to go with this CDS then make one. Otherwise,
// make sure the existing gene feature includes the location of the CDS.
//
CRef< CSeq_feat > pGene;
if ( ! x_FindParentGene( gff, pGene ) ) {
if ( ! x_CreateParentGene( gff, pAnnot ) ) {
return false;
}
}
else {
if ( ! x_MergeParentGene( gff, pGene ) ) {
return false;
}
}
//
// If there is no CDS feature with this gene_id|transcript_id then make one.
// Otherwise, fix up the location of the existing one.
//
CRef< CSeq_feat > pCds;
if ( ! x_FindParentCds( gff, pCds ) ) {
//
// Create a brand new CDS feature:
//
if ( ! x_CreateParentCds( gff, pAnnot ) ) {
return false;
}
x_FindParentCds( gff, pCds );
}
else {
//
// Update an already existing CDS features:
//
if ( ! x_MergeFeatureLocationMultiInterval( gff, pCds ) ) {
return false;
}
if (!x_FeatureTrimQualifiers(gff, pCds)) {
return false;
}
}
if ( x_CdsIsPartial( gff ) ) {
CRef<CSeq_feat> pParent;
if ( x_FindParentMrna( gff, pParent ) ) {
CSeq_loc& loc = pCds->SetLocation();
size_t uCdsStart = gff.SeqStart();
size_t uMrnaStart = pParent->GetLocation().GetStart( eExtreme_Positional );
if ( uCdsStart == uMrnaStart ) {
loc.SetPartialStart( true, eExtreme_Positional );
// cerr << "fuzzed down: " << gff.SeqStart() << " " << gff.SeqStop() << " vs. " << uMrnaStart << endl;
}
size_t uCdsStop = gff.SeqStop();
size_t uMrnaStop = pParent->GetLocation().GetStop( eExtreme_Positional );
if ( uCdsStop == uMrnaStop && gff.Type() != "stop_codon" ) {
loc.SetPartialStop( true, eExtreme_Positional );
// cerr << "fuzzed up : " << gff.SeqStart() << " " << gff.SeqStop() << " vs. " << uMrnaStop << endl;
}
}
}
return true;
}
// ----------------------------------------------------------------------------
bool CGtfReader::x_UpdateAnnotFeature(
const CGff2Record& gff,
CRef< CSeq_annot > pAnnot,
ILineErrorListener* pEC)
// ----------------------------------------------------------------------------
{
CRef< CSeq_feat > pFeature( new CSeq_feat );
//
// Handle officially recognized GTF types:
//
string strType = gff.Type();
if ( strType == "CDS" ) {
//
// Observations:
// Location does not include the stop codon hence must be fixed up once
// the stop codon is seen.
//
return x_UpdateAnnotCds( gff, pAnnot );
}
if ( strType == "start_codon" ) {
//
// Observation:
// Comes in up to three pieces (depending on splicing).
// Location _is_ included in CDS.
//
return x_UpdateAnnotStartCodon( gff, pAnnot );
}
if ( strType == "stop_codon" ) {
//
// Observation:
// Comes in up to three pieces (depending on splicing).
// Location not included in CDS hence must be used to fix up location of
// the coding region.
//
return x_UpdateAnnotStopCodon( gff, pAnnot );
}
if ( strType == "5UTR" ) {
return x_UpdateAnnot5utr( gff, pAnnot );
}
if ( strType == "3UTR" ) {
return x_UpdateAnnot3utr( gff, pAnnot );
}
if ( strType == "inter" ) {
return x_UpdateAnnotInter( gff, pAnnot );
}
if ( strType == "inter_CNS" ) {
return x_UpdateAnnotInterCns( gff, pAnnot );
}
if ( strType == "intron_CNS" ) {
return x_UpdateAnnotIntronCns( gff, pAnnot );
}
if ( strType == "exon" ||
strType == "initial" ||
strType == "internal" ||
strType == "terminal" ||
strType == "single") {
return x_UpdateAnnotExon( gff, pAnnot );
}
//
// Every other type is not officially sanctioned GTF, and per spec we are
// supposed to ignore it. In the spirit of being lenient on input we may
// try to salvage some of it anyway.
//
if ( strType == "gene" ) {
//
// Not an official GTF feature type but seen frequently. Hence we give
// it some recognition.
//
if ( ! x_CreateParentGene( gff, pAnnot ) ) {
return false;
}
}
if (strType == "mRNA") {
if ( ! x_CreateParentMrna(gff, pAnnot) ) {
return false;
}
}
return x_UpdateAnnotMiscFeature( gff, pAnnot );
}
