int cmpRedundant(GffObj& a, GffObj& b) {
if (a.exons.Count()==b.exons.Count()) {
if (a.covlen==b.covlen) {
return strcmp(a.getID(), b.getID());
}
else return (a.covlen>b.covlen)? 1 : -1;
}
else return (a.exons.Count()>b.exons.Count())? 1: -1;
}
bool unsplContained(GffObj& ti, GffObj& tj, bool fuzzSpan) {
//returns true only if ti (which MUST be single-exon) is "almost" contained in any of tj's exons
//but it does not cross any intron-exon boundary of tj
int imax=ti.exons.Count()-1;
int jmax=tj.exons.Count()-1;
if (imax>0) GError("Error: bad unsplContained() call, 1st param must be single-exon transcript!\n");
int minovl = (int)(0.8 * ti.len()); //minimum overlap for fuzzSpan
if (fuzzSpan) {
for (int j=0;j<=jmax;j++) {
//must NOT overlap the introns
if ((j>0 && ti.start<tj.exons[j]->start)
|| (j<jmax && ti.end>tj.exons[j]->end))
return false;
if (ti.exons[0]->overlapLen(tj.exons[j])>=minovl)
return true;
}
} else {
for (int j=0;j<=jmax;j++) {
//must NOT overlap the introns
if ((j>0 && ti.start<tj.exons[j]->start)
|| (j<jmax && ti.end>tj.exons[j]->end))
return false;
//strict containment
if (ti.end<=tj.exons[j]->end && ti.start>=tj.exons[j]->start)
return true;
}
}
return false;
}
int main(int argc, char* argv[]) {
if (argc == 1 or argc > 2) {
std::cerr << "Usage: TestGFFParse input\n";
std::exit(1);
}
GffReader reader(argv[1]);
reader.readAll(true);
std::cerr << "had count of " << reader.gflst.Count() << "\n";
size_t nfeat = reader.gflst.Count();
for (size_t i=0; i < nfeat; ++i) {
GffObj* f = reader.gflst[i];
if (f->isTranscript()) {
std::cout << f->getID() << '\t' << f->getGeneID() << '\t';
if (f->attrs) {
for (size_t j=0; j < f->attrs->Count(); ++j) {
std::cout << f->getAttrName(j) << "\t" << f->getAttrValue(j) << "\t";
}
}
std::cout << "\n";
}
}
std::exit(0);
}
bool exonOverlap2Gene(GffObj* t, GffObj& g) {
if (t->exons.Count()>0) {
return t->exonOverlap(g.start, g.end);
}
else return g.overlap(*t);
}
bool process_transcript(GFastaDb& gfasta, GffObj& gffrec) {
//returns true if the transcript passed the filter
char* gname=gffrec.getGeneName();
if (gname==NULL) gname=gffrec.getGeneID();
GStr defline(gffrec.getID());
if (f_out && !fmtGTF) {
const char* tname=NULL;
if ((tname=gffrec.getAttr("transcript_name"))!=NULL) {
gffrec.addAttr("Name", tname);
gffrec.removeAttr("transcript_name");
}
}
if (ensembl_convert && startsWith(gffrec.getID(), "ENS")) {
const char* biotype=gffrec.getAttr("gene_biotype");
if (biotype) {
gffrec.addAttr("type", biotype);
gffrec.removeAttr("gene_biotype");
}
else { //old Ensembl files lacking gene_biotype
gffrec.addAttr("type", gffrec.getTrackName());
}
//bool is_gene=false;
bool is_pseudo=false;
if (strcmp(biotype, "protein_coding")==0 || gffrec.hasCDS())
gffrec.setFeatureName("mRNA");
else {
if (strcmp(biotype, "processed_transcript")==0)
gffrec.setFeatureName("proc_RNA");
else {
//is_gene=endsWith(biotype, "gene");
is_pseudo=strifind(biotype, "pseudo");
if (is_pseudo) {
gffrec.setFeatureName("pseudo_RNA");
}
else if (endsWith(biotype, "RNA")) {
gffrec.setFeatureName(biotype);
} else gffrec.setFeatureName("misc_RNA");
}
}
}
if (gname && strcmp(gname, gffrec.getID())!=0) {
int* isonum=isoCounter.Find(gname);
if (isonum==NULL) {
isonum=new int(1);
isoCounter.Add(gname,isonum);
}
else (*isonum)++;
defline.appendfmt(" gene=%s", gname);
}
int seqlen=0;
const char* tlabel=tracklabel;
if (tlabel==NULL) tlabel=gffrec.getTrackName();
//defline.appendfmt(" track:%s",tlabel);
char* cdsnt = NULL;
char* cdsaa = NULL;
int aalen=0;
for (int i=1;i<gffrec.exons.Count();i++) {
int ilen=gffrec.exons[i]->start-gffrec.exons[i-1]->end-1;
if (ilen>4000000)
GMessage("Warning: very large intron (%d) for transcript %s\n",
ilen, gffrec.getID());
if (ilen>maxintron) {
return false;
}
}
GList<GSeg> seglst(false,true);
GFaSeqGet* faseq=fastaSeqGet(gfasta, gffrec);
if (spliceCheck && gffrec.exons.Count()>1) {
//check introns for splice site consensi ( GT-AG, GC-AG or AT-AC )
if (faseq==NULL) GError("Error: no genomic sequence available!\n");
int glen=gffrec.end-gffrec.start+1;
const char* gseq=faseq->subseq(gffrec.start, glen);
bool revcompl=(gffrec.strand=='-');
bool ssValid=true;
for (int e=1;e<gffrec.exons.Count();e++) {
const char* intron=gseq+gffrec.exons[e-1]->end+1-gffrec.start;
int intronlen=gffrec.exons[e]->start-gffrec.exons[e-1]->end-1;
GSpliceSite acceptorSite(intron,intronlen,true, revcompl);
GSpliceSite donorSite(intron,intronlen, false, revcompl);
//GMessage("%c intron %d-%d : %s .. %s\n",
// gffrec.strand, istart, iend, donorSite.nt, acceptorSite.nt);
if (acceptorSite=="AG") { // GT-AG or GC-AG
if (!donorSite.canonicalDonor()) {
ssValid=false;break;
}
}
else if (acceptorSite=="AC") { //
if (donorSite!="AT") { ssValid=false; break; }
}
else { ssValid=false; break; }
}
//GFREE(gseq);
if (!ssValid) {
if (verbose)
GMessage("Invalid splice sites found for '%s'\n",gffrec.getID());
return false; //don't print this one!
}
}
bool trprint=true;
int stopCodonAdjust=0;
int mCDphase=0;
bool hasStop=false;
if (gffrec.CDphase=='1' || gffrec.CDphase=='2')
mCDphase = gffrec.CDphase-'0';
if (f_y!=NULL || f_x!=NULL || validCDSonly) {
if (faseq==NULL) GError("Error: no genomic sequence provided!\n");
//if (protmap && fullCDSonly) {
//if (protmap && (fullCDSonly || (gffrec.qlen>0 && gffrec.qend==gffrec.qlen))) {
if (validCDSonly) { //make sure the stop codon is always included
//adjust_stopcodon(gffrec,3);
stopCodonAdjust=adjust_stopcodon(gffrec,3);
}
int strandNum=0;
int phaseNum=0;
CDS_CHECK:
cdsnt=gffrec.getSpliced(faseq, true, &seqlen, NULL, NULL, &seglst);
if (cdsnt==NULL) trprint=false;
else { //has CDS
if (validCDSonly) {
cdsaa=translateDNA(cdsnt, aalen, seqlen);
char* p=strchr(cdsaa,'.');
hasStop=false;
if (p!=NULL) {
if (p-cdsaa>=aalen-2) { //stop found as the last codon
*p='0';//remove it
hasStop=true;
if (aalen-2==p-cdsaa) {
//previous to last codon is the stop codon
//so correct the CDS stop accordingly
adjust_stopcodon(gffrec,-3, &seglst);
stopCodonAdjust=0; //clear artificial stop adjustment
seqlen-=3;
cdsnt[seqlen]=0;
}
aalen=p-cdsaa;
}
else {//stop found before the last codon
trprint=false;
}
}//stop codon found
if (trprint==false) { //failed CDS validity check
//in-frame stop codon found
if (altPhases && phaseNum<3) {
phaseNum++;
gffrec.CDphase = '0'+((mCDphase+phaseNum)%3);
GFREE(cdsaa);
goto CDS_CHECK;
}
if (gffrec.exons.Count()==1 && bothStrands) {
strandNum++;
phaseNum=0;
if (strandNum<2) {
GFREE(cdsaa);
gffrec.strand = (gffrec.strand=='-') ? '+':'-';
goto CDS_CHECK; //repeat the CDS check for a different frame
}
}
if (verbose) GMessage("In-frame STOP found for '%s'\n",gffrec.getID());
} //has in-frame STOP
if (fullCDSonly) {
if (!hasStop || cdsaa[0]!='M') trprint=false;
}
} // CDS check requested
} //has CDS
} //translation or codon check/output was requested
if (!trprint) {
GFREE(cdsnt);
GFREE(cdsaa);
return false;
}
if (stopCodonAdjust>0 && !hasStop) {
//restore stop codon location
adjust_stopcodon(gffrec, -stopCodonAdjust, &seglst);
if (cdsnt!=NULL && seqlen>0) {
seqlen-=stopCodonAdjust;
cdsnt[seqlen]=0;
}
if (cdsaa!=NULL) aalen--;
}
if (f_y!=NULL) { //CDS translation fasta output requested
//char*
if (cdsaa==NULL) { //translate now if not done before
cdsaa=translateDNA(cdsnt, aalen, seqlen);
}
if (fullattr && gffrec.attrs!=NULL) {
//append all attributes found for each transcripts
for (int i=0;i<gffrec.attrs->Count();i++) {
defline.append(" ");
defline.append(gffrec.getAttrName(i));
defline.append("=");
defline.append(gffrec.getAttrValue(i));
}
}
printFasta(f_y, defline, cdsaa, aalen);
}
if (f_x!=NULL) { //CDS only
if (writeExonSegs) {
defline.append(" loc:");
defline.append(gffrec.getGSeqName());
defline.appendfmt("(%c)",gffrec.strand);
//warning: not CDS coordinates are written here, but the exon ones
defline+=(int)gffrec.start;
defline+=(char)'-';
defline+=(int)gffrec.end;
// -- here these are CDS substring coordinates on the spliced sequence:
defline.append(" segs:");
for (int i=0;i<seglst.Count();i++) {
if (i>0) defline.append(",");
defline+=(int)seglst[i]->start;
defline.append("-");
defline+=(int)seglst[i]->end;
}
}
if (fullattr && gffrec.attrs!=NULL) {
//append all attributes found for each transcript
for (int i=0;i<gffrec.attrs->Count();i++) {
defline.append(" ");
defline.append(gffrec.getAttrName(i));
defline.append("=");
defline.append(gffrec.getAttrValue(i));
}
}
printFasta(f_x, defline, cdsnt, seqlen);
}
GFREE(cdsnt);
GFREE(cdsaa);
if (f_w!=NULL) { //write spliced exons
uint cds_start=0;
uint cds_end=0;
seglst.Clear();
char* exont=gffrec.getSpliced(faseq, false, &seqlen, &cds_start, &cds_end, &seglst);
if (exont!=NULL) {
if (gffrec.CDstart>0) {
defline.appendfmt(" CDS=%d-%d", cds_start, cds_end);
}
if (writeExonSegs) {
defline.append(" loc:");
defline.append(gffrec.getGSeqName());
defline+=(char)'|';
defline+=(int)gffrec.start;
defline+=(char)'-';
defline+=(int)gffrec.end;
defline+=(char)'|';
defline+=(char)gffrec.strand;
defline.append(" exons:");
for (int i=0;i<gffrec.exons.Count();i++) {
if (i>0) defline.append(",");
defline+=(int)gffrec.exons[i]->start;
defline.append("-");
defline+=(int)gffrec.exons[i]->end;
}
defline.append(" segs:");
for (int i=0;i<seglst.Count();i++) {
if (i>0) defline.append(",");
defline+=(int)seglst[i]->start;
defline.append("-");
defline+=(int)seglst[i]->end;
}
}
if (fullattr && gffrec.attrs!=NULL) {
//append all attributes found for each transcripts
for (int i=0;i<gffrec.attrs->Count();i++) {
defline.append(" ");
defline.append(gffrec.getAttrName(i));
defline.append("=");
defline.append(gffrec.getAttrValue(i));
}
}
printFasta(f_w, defline, exont, seqlen);
GFREE(exont);
}
} //writing f_w (spliced exons)
return true;
}
