void processLast(FILE *of, Config *config, struct lastCall *last, bam_hdr_t *hdr, int32_t tid, int32_t pos, int width, uint32_t nmethyl, uint32_t nunmethyl) {
if(last->tid == tid && last->pos == pos) {
nmethyl += last->nmethyl;
nunmethyl += last->nunmethyl;
writeCall(of, config, hdr->target_name[tid], pos, width, nmethyl, nunmethyl);
last->tid = -1;
} else {
if(last->tid != -1) {
writeCall(of, config, hdr->target_name[tid], pos, width, last->nmethyl+nmethyl, last->nunmethyl+nunmethyl);
}
last->tid = tid;
last->pos = pos;
last->nmethyl = nmethyl;
last->nunmethyl = nunmethyl;
}
}
void CRecordingEngExtInterceptor::PopFunc(CRecorderFuncCall* func)
{
if (func != m_LastFunc)
rehlds_syserror("%s: stack corrupted", __FUNCTION__);
writeCall(!func->m_StartWritten, true, func->m_FuncCall);
if (m_LastFunc->m_Prev == NULL) {
m_LastFunc = m_RootFunc = NULL;
}
else
{
m_LastFunc = m_LastFunc->m_Prev;
m_LastFunc->m_Next = NULL;
}
}
void CRecordingEngExtInterceptor::PushFunc(CRecorderFuncCall* func)
{
CRecorderFuncCall* cc = m_LastFunc;
while (cc != NULL) {
if (!cc->m_StartWritten) {
writeCall(true, false, cc->m_FuncCall);
cc->m_StartWritten = true;
}
cc = cc->m_Prev;
}
if (m_LastFunc != NULL) {
m_LastFunc->m_Next = func;
func->m_Next = NULL;
func->m_Prev = m_LastFunc;
m_LastFunc = func;
}
else
{
m_LastFunc = m_RootFunc = func;
func->m_Next = func->m_Prev = NULL;
}
}
void extractCalls(Config *config) {
bam_hdr_t *hdr = sam_hdr_read(config->fp);
bam_mplp_t iter;
int ret, tid, pos, i, seqlen, type, rv, o = 0;
int beg0 = 0, end0 = 1u<<29;
int n_plp; //This will need to be modified for multiple input files
int ctid = -1; //The tid of the contig whose sequence is stored in "seq"
int idxBED = 0, strand;
uint32_t nmethyl = 0, nunmethyl = 0;
const bam_pileup1_t **plp = NULL;
char *seq = NULL, base;
mplp_data *data = NULL;
struct lastCall *lastCpG = NULL;
struct lastCall *lastCHG = NULL;
if(config->merge) {
if(config->keepCpG) {
lastCpG = calloc(1, sizeof(struct lastCall));
assert(lastCpG);
lastCpG->tid = -1;
}
if(config->keepCHG) {
lastCHG = calloc(1, sizeof(struct lastCall));
assert(lastCHG);
lastCHG->tid = -1;
}
}
data = calloc(1,sizeof(mplp_data));
if(data == NULL) {
fprintf(stderr, "Couldn't allocate space for the data structure in extractCalls()!\n");
return;
}
data->config = config;
data->hdr = hdr;
if (config->reg) {
if((data->iter = sam_itr_querys(config->bai, hdr, config->reg)) == 0) {
fprintf(stderr, "failed to parse regions %s", config->reg);
return;
}
}
if(config->bedName) {
config->bed = parseBED(config->bedName, hdr);
if(config->bed == NULL) return;
}
plp = calloc(1, sizeof(bam_pileup1_t *)); //This will have to be modified for multiple input files
if(plp == NULL) {
fprintf(stderr, "Couldn't allocate space for the plp structure in extractCalls()!\n");
return;
}
//Start the pileup
iter = bam_mplp_init(1, filter_func, (void **) &data);
bam_mplp_init_overlaps(iter);
bam_mplp_set_maxcnt(iter, config->maxDepth);
while((ret = cust_mplp_auto(iter, &tid, &pos, &n_plp, plp)) > 0) {
//Do we need to process this position?
if (config->reg) {
beg0 = data->iter->beg, end0 = data->iter->end;
if ((pos < beg0 || pos >= end0)) continue; // out of the region requested
}
if(tid != ctid) {
if(seq != NULL) free(seq);
seq = faidx_fetch_seq(config->fai, hdr->target_name[tid], 0, faidx_seq_len(config->fai, hdr->target_name[tid]), &seqlen);
if(seqlen < 0) {
fprintf(stderr, "faidx_fetch_seq returned %i while trying to fetch the sequence for tid %i (%s)!\n",\
seqlen, tid, hdr->target_name[tid]);
fprintf(stderr, "Note that the output will be truncated!\n");
return;
}
ctid = tid;
}
if(config->bed) { //Handle -l
while((o = posOverlapsBED(tid, pos, config->bed, idxBED)) == -1) idxBED++;
if(o == 0) continue; //Wrong strand
}
if(isCpG(seq, pos, seqlen)) {
if(!config->keepCpG) continue;
type = 0;
} else if(isCHG(seq, pos, seqlen)) {
if(!config->keepCHG) continue;
type = 1;
} else if(isCHH(seq, pos, seqlen)) {
if(!config->keepCHH) continue;
type = 2;
} else {
continue;
}
nmethyl = nunmethyl = 0;
base = *(seq+pos);
for(i=0; i<n_plp; i++) {
if(plp[0][i].is_del) continue;
if(plp[0][i].is_refskip) continue;
if(config->bed) if(!readStrandOverlapsBED(plp[0][i].b, config->bed->region[idxBED])) continue;
strand = getStrand((plp[0]+i)->b);
if(strand & 1) {
if(base != 'C' && base != 'c') continue;
} else {
if(base != 'G' && base != 'g') continue;
}
rv = updateMetrics(config, plp[0]+i);
if(rv > 0) nmethyl++;
else if(rv<0) nunmethyl++;
}
if(nmethyl+nunmethyl==0) continue;
if(!config->merge || type==2) {
writeCall(config->output_fp[type], config, hdr->target_name[tid], pos, 1, nmethyl, nunmethyl);
} else {
//Merge into per-CpG/CHG metrics
if(type==0) {
if(base=='G' || base=='g') pos--;
processLast(config->output_fp[0], config, lastCpG, hdr, tid, pos, 2, nmethyl, nunmethyl);
} else {
if(base=='G' || base=='g') pos-=2;
processLast(config->output_fp[1], config, lastCHG, hdr, tid, pos, 3, nmethyl, nunmethyl);
}
}
}
//Don't forget the last CpG/CHG
if(config->merge) {
if(config->keepCpG && lastCpG->tid != -1) {
processLast(config->output_fp[0], config, lastCpG, hdr, tid, pos, 2, nmethyl, nunmethyl);
}
if(config->keepCHG && lastCHG->tid != -1) {
processLast(config->output_fp[1], config, lastCHG, hdr, tid, pos, 3, nmethyl, nunmethyl);
}
}
bam_hdr_destroy(hdr);
if(data->iter) hts_itr_destroy(data->iter);
bam_mplp_destroy(iter);
free(data);
free(plp);
if(seq != NULL) free(seq);
}
