int calcCoverage(char *fName, Slice *slice, htsFile *in, hts_idx_t *idx, int flags) {
int ref;
int begRange;
int endRange;
char region[1024];
char region_name[512];
if (Slice_getChrStart(slice) != 1) {
fprintf(stderr, "Currently only allow a slice start position of 1\n");
return 1;
}
if (flags & M_UCSC_NAMING) {
sprintf(region,"chr%s", Slice_getSeqRegionName(slice));
} else {
sprintf(region,"%s", Slice_getSeqRegionName(slice));
}
bam_hdr_t *header = bam_hdr_init();
header = bam_hdr_read(in->fp.bgzf);
ref = bam_name2id(header, region);
if (ref < 0) {
fprintf(stderr, "Invalid region %s\n", region);
exit(1);
}
sprintf(region,"%s:%ld-%ld", region_name,
Slice_getSeqRegionStart(slice),
Slice_getSeqRegionEnd(slice));
if (hts_parse_reg(region, &begRange, &endRange) == NULL) {
fprintf(stderr, "Could not parse %s\n", region);
exit(2);
}
bam_hdr_destroy(header);
hts_itr_t *iter = sam_itr_queryi(idx, ref, begRange, endRange);
bam1_t *b = bam_init1();
Coverage *coverage = calloc(Slice_getLength(slice),sizeof(Coverage));
long counter = 0;
long overlapping = 0;
long bad = 0;
int startIndex = 0;
while (bam_itr_next(in, iter, b) >= 0) {
if (b->core.flag & (BAM_FUNMAP | BAM_FSECONDARY | BAM_FQCFAIL | BAM_FDUP)) {
bad++;
continue;
}
int end;
//end = bam_calend(&b->core, bam1_cigar(b));
end = bam_endpos(b);
// There is a special case for reads which have zero length and start at begRange (so end at begRange ie. before the first base we're interested in).
// That is the reason for the || end == begRange test
if (end == begRange) {
continue;
}
counter++;
if (!(counter%1000000)) {
if (verbosity > 1) { printf("."); }
fflush(stdout);
}
// Remember: b->core.pos is zero based!
int cigInd;
int refPos;
int readPos;
uint32_t *cigar = bam_get_cigar(b);
for (cigInd = readPos = 0, refPos = b->core.pos; cigInd < b->core.n_cigar; ++cigInd) {
int k;
int lenCigBlock = cigar[cigInd]>>4;
int op = cigar[cigInd]&0xf;
if (op == BAM_CMATCH || op == BAM_CEQUAL || op == BAM_CDIFF) {
for (k = 0; k < lenCigBlock; ++k) {
//if (ref[refPos+k] == 0) break; // out of boundary
coverage[refPos+k].coverage++;
}
if (k < lenCigBlock) break;
refPos += lenCigBlock; readPos += lenCigBlock;
} else if (op == BAM_CDEL) {
for (k = 0; k < lenCigBlock; ++k) {
// if (ref[refPos+k] == 0) break;
coverage[refPos+k].coverage++;
}
if (k < lenCigBlock) break;
refPos += lenCigBlock;
} else if (op == BAM_CSOFT_CLIP) {
readPos += lenCigBlock;
} else if (op == BAM_CHARD_CLIP) {
} else if (op == BAM_CINS) {
readPos += lenCigBlock;
} else if (op == BAM_CREF_SKIP) {
refPos += lenCigBlock;
}
}
#ifdef DONE
int j;
int done = 0;
int hadOverlap = 0;
for (j=startIndex; j < Vector_getNumElement(genes) && !done; j++) {
Gene *gene = Vector_getElementAt(genes,j);
if (!gene) {
continue;
}
// Remember: b->core.pos is zero based!
if (b->core.pos < Gene_getEnd(gene) && end >= Gene_getStart(gene)) {
int k;
int doneGene = 0;
for (k=0; k<Gene_getTranscriptCount(gene) && !doneGene; k++) {
Transcript *trans = Gene_getTranscriptAt(gene,k);
if (b->core.pos < Transcript_getEnd(trans) && end >= Transcript_getStart(trans)) {
int m;
for (m=0; m<Transcript_getExonCount(trans) && !doneGene; m++) {
Exon *exon = Transcript_getExonAt(trans,m);
if (b->core.pos < Exon_getEnd(exon) && end >= Exon_getStart(exon)) {
// Only count as overlapping once (could be that a read overlaps more than one gene)
if (!hadOverlap) {
overlapping++;
hadOverlap = 1;
}
gs = IDHash_getValue(geneCountsHash, Gene_getDbID(gene));
gs->score++;
doneGene = 1;
}
}
}
}
} else if (Gene_getStart(gene) > end) {
done = 1;
} else if (Gene_getEnd(gene) < b->core.pos+1) {
gs = IDHash_getValue(geneCountsHash, Gene_getDbID(gene));
printf("Gene %s (%s) score %ld\n",Gene_getStableId(gene),
Gene_getDisplayXref(gene) ? DBEntry_getDisplayId(Gene_getDisplayXref(gene)) : "",
gs->score);
if (verbosity > 1) {
printf("Removing gene %s (index %d) with extent %d to %d\n",
Gene_getStableId(gene),
gs->index,
Gene_getStart(gene),
Gene_getEnd(gene));
}
Vector_setElementAt(genes,j,NULL);
// Magic (very important for speed) - move startIndex to first non null gene
int n;
startIndex = 0;
for (n=0;n<Vector_getNumElement(genes);n++) {
void *v = Vector_getElementAt(genes,n);
if (v != NULL) {
break;
}
startIndex++;
}
if (verbosity > 1) {
printf("startIndex now %d\n",startIndex);
}
}
}
#endif
}
if (verbosity > 1) { printf("\n"); }
#ifdef DONE
// Print out read counts for what ever's left in the genes array
int n;
for (n=0;n<Vector_getNumElement(genes);n++) {
Gene *gene = Vector_getElementAt(genes,n);
if (gene != NULL) {
gs = IDHash_getValue(geneCountsHash, Gene_getDbID(gene));
printf("Gene %s (%s) score %ld\n",Gene_getStableId(gene),
Gene_getDisplayXref(gene) ? DBEntry_getDisplayId(Gene_getDisplayXref(gene)) : "",
gs->score);
}
}
#endif
printf("Read %ld reads. Number of bad reads (unmapped, qc fail, secondary, dup) %ld\n", counter, bad);
long i;
for (i=0; i< Slice_getLength(slice); i++) {
printf("%ld %ld\n", i+1, coverage[i].coverage);
}
sam_itr_destroy(iter);
bam_destroy1(b);
return 1;
}
int dumpGenes(Vector *genes, int withSupport) {
FILE *fp = stderr;
int i;
int failed = 0;
for (i=0;i<Vector_getNumElement(genes) && !failed;i++) {
Gene *g = Vector_getElementAt(genes,i);
fprintf(fp,"Gene %s (%s) coords: %ld %ld %d\n",Gene_getStableId(g),(Gene_getDisplayXref(g) ? DBEntry_getDisplayId(Gene_getDisplayXref(g)) : ""),Gene_getStart(g),Gene_getEnd(g),Gene_getStrand(g));
int j;
for (j=0;j<Gene_getTranscriptCount(g);j++) {
Transcript *t = Gene_getTranscriptAt(g,j);
int k;
fprintf(fp," Trans %s coords: %ld %ld %d biotype: %s\n",Transcript_getStableId(t), Transcript_getStart(t),Transcript_getEnd(t),Transcript_getStrand(t),Transcript_getBiotype(t));
if (withSupport) {
Vector *support = Transcript_getAllSupportingFeatures(t);
for (k=0; k<Vector_getNumElement(support); k++) {
BaseAlignFeature *baf = Vector_getElementAt(support, k);
fprintf(fp," support %s coords: %ld %ld %d\n", BaseAlignFeature_getHitSeqName(baf), BaseAlignFeature_getStart(baf), BaseAlignFeature_getEnd(baf), BaseAlignFeature_getStrand(baf));
}
Vector *intronSupport = Transcript_getAllIntronSupportingEvidence(t);
for (k=0; k<Vector_getNumElement(intronSupport); k++) {
IntronSupportingEvidence *ise = Vector_getElementAt(intronSupport, k);
fprintf(fp," intron support %s coords: %ld %ld %d\n", IntronSupportingEvidence_getHitName(ise), IntronSupportingEvidence_getStart(ise), IntronSupportingEvidence_getEnd(ise), IntronSupportingEvidence_getStrand(ise));
}
}
for (k=0;k<Transcript_getExonCount(t);k++) {
Exon *e = Transcript_getExonAt(t,k);
fprintf(fp," exon %s (%p) coords: %ld %ld %d\n",Exon_getStableId(e), e, Exon_getStart(e), Exon_getEnd(e), Exon_getStrand(e));
if (withSupport) {
Vector *support = Exon_getAllSupportingFeatures(e);
int m;
for (m=0; m<Vector_getNumElement(support); m++) {
BaseAlignFeature *baf = Vector_getElementAt(support, m);
fprintf(fp," support %s coords: %ld %ld %d\n", BaseAlignFeature_getHitSeqName(baf), BaseAlignFeature_getStart(baf), BaseAlignFeature_getEnd(baf), BaseAlignFeature_getStrand(baf));
}
}
}
Translation *tln = Transcript_getTranslation(t);
if (tln) {
fprintf(fp," translation id: %s %s %d %s %d\n",Translation_getStableId(tln),
Exon_getStableId(Translation_getStartExon(tln)), Translation_getStart(tln),
Exon_getStableId(Translation_getEndExon(tln)), Translation_getEnd(tln));
char *tSeq = Transcript_translate(t);
fprintf(fp," translation: %s\n",tSeq);
free(tSeq);
Vector *tlnAttribs = Translation_getAllAttributes(tln, NULL);
if (Vector_getNumElement(tlnAttribs)) {
fprintf(fp, " translation attributes:\n");
int n;
for (n=0; n<Vector_getNumElement(tlnAttribs); n++) {
Attribute *attrib = Vector_getElementAt(tlnAttribs, n);
fprintf(fp, " code %s name %s desc %s value %s\n",
Attribute_getCode(attrib),
Attribute_getName(attrib),
Attribute_getDescription(attrib),
Attribute_getValue(attrib));
}
}
}
}
}
return failed;
}
