#define MPLP_FMT_SP 0x200
#define MPLP_NO_INDEL 0x400
#define MPLP_EXT_BAQ 0x800
#define MPLP_FMT_SP 0x200
#define MPLP_NO_INDEL 0x400
#define MPLP_EXT_BAQ 0x800
+#define MPLP_ILLUMINA13 0x1000
+
+void *bed_read(const char *fn);
+void bed_destroy(void *_h);
+int bed_overlap(const void *_h, const char *chr, int beg, int end);
- int max_mq, min_mq, flag, min_baseQ, capQ_thres, max_depth;
+ int max_mq, min_mq, flag, min_baseQ, capQ_thres, max_depth, max_indel_depth;
int openQ, extQ, tandemQ, min_support; // for indels
double min_frac; // for indels
int openQ, extQ, tandemQ, min_support; // for indels
double min_frac; // for indels
int has_ref;
ret = ma->iter? bam_iter_read(ma->fp, ma->iter, b) : bam_read1(ma->fp, b);
if (ret < 0) break;
int has_ref;
ret = ma->iter? bam_iter_read(ma->fp, ma->iter, b) : bam_read1(ma->fp, b);
if (ret < 0) break;
+ if (b->core.tid < 0 || (b->core.flag&BAM_FUNMAP)) { // exclude unmapped reads
+ skip = 1;
+ continue;
+ }
+ if (ma->conf->bed) { // test overlap
+ skip = !bed_overlap(ma->conf->bed, ma->h->target_name[b->core.tid], b->core.pos, bam_calend(&b->core, bam1_cigar(b)));
+ if (skip) continue;
+ }
if (ma->conf->rghash) { // exclude read groups
uint8_t *rg = bam_aux_get(b, "RG");
skip = (rg && bcf_str2id(ma->conf->rghash, (const char*)(rg+1)) >= 0);
if (skip) continue;
}
if (ma->conf->rghash) { // exclude read groups
uint8_t *rg = bam_aux_get(b, "RG");
skip = (rg && bcf_str2id(ma->conf->rghash, (const char*)(rg+1)) >= 0);
if (skip) continue;
}
has_ref = (ma->ref && ma->ref_id == b->core.tid)? 1 : 0;
skip = 0;
if (has_ref && (ma->conf->flag&MPLP_REALN)) bam_prob_realn_core(b, ma->ref, (ma->conf->flag & MPLP_EXT_BAQ)? 3 : 1);
has_ref = (ma->ref && ma->ref_id == b->core.tid)? 1 : 0;
skip = 0;
if (has_ref && (ma->conf->flag&MPLP_REALN)) bam_prob_realn_core(b, ma->ref, (ma->conf->flag & MPLP_EXT_BAQ)? 3 : 1);
int q = bam_cap_mapQ(b, ma->ref, ma->conf->capQ_thres);
if (q < 0) skip = 1;
else if (b->core.qual > q) b->core.qual = q;
int q = bam_cap_mapQ(b, ma->ref, ma->conf->capQ_thres);
if (q < 0) skip = 1;
else if (b->core.qual > q) b->core.qual = q;
else if (b->core.qual < ma->conf->min_mq) skip = 1;
else if ((ma->conf->flag&MPLP_NO_ORPHAN) && (b->core.flag&1) && !(b->core.flag&2)) skip = 1;
} while (skip);
else if (b->core.qual < ma->conf->min_mq) skip = 1;
else if ((ma->conf->flag&MPLP_NO_ORPHAN) && (b->core.flag&1) && !(b->core.flag&2)) skip = 1;
} while (skip);
q = bam_aux_get(p->b, "RG");
if (q) id = bam_smpl_rg2smid(sm, fn[i], (char*)q+1, buf);
if (id < 0) id = bam_smpl_rg2smid(sm, fn[i], 0, buf);
q = bam_aux_get(p->b, "RG");
if (q) id = bam_smpl_rg2smid(sm, fn[i], (char*)q+1, buf);
if (id < 0) id = bam_smpl_rg2smid(sm, fn[i], 0, buf);
- assert(id >= 0 && id < m->n);
+ if (id < 0 || id >= m->n) {
+ assert(q); // otherwise a bug
+ fprintf(stderr, "[%s] Read group %s used in file %s but not defined in the header.\n", __func__, (char*)q+1, fn[i]);
+ exit(1);
+ }
if (m->n_plp[id] == m->m_plp[id]) {
m->m_plp[id] = m->m_plp[id]? m->m_plp[id]<<1 : 8;
m->plp[id] = realloc(m->plp[id], sizeof(bam_pileup1_t) * m->m_plp[id]);
if (m->n_plp[id] == m->m_plp[id]) {
m->m_plp[id] = m->m_plp[id]? m->m_plp[id]<<1 : 8;
m->plp[id] = realloc(m->plp[id], sizeof(bam_pileup1_t) * m->m_plp[id]);
extern void *bcf_call_add_rg(void *rghash, const char *hdtext, const char *list);
extern void bcf_call_del_rghash(void *rghash);
mplp_aux_t **data;
extern void *bcf_call_add_rg(void *rghash, const char *hdtext, const char *list);
extern void bcf_call_del_rghash(void *rghash);
mplp_aux_t **data;
- int i, tid, pos, *n_plp, beg0 = 0, end0 = 1u<<29, ref_len, ref_tid, max_depth;
+ int i, tid, pos, *n_plp, beg0 = 0, end0 = 1u<<29, ref_len, ref_tid, max_depth, max_indel_depth;
data[i]->fp = strcmp(fn[i], "-") == 0? bam_dopen(fileno(stdin), "r") : bam_open(fn[i], "r");
data[i]->conf = conf;
h_tmp = bam_header_read(data[i]->fp);
data[i]->fp = strcmp(fn[i], "-") == 0? bam_dopen(fileno(stdin), "r") : bam_open(fn[i], "r");
data[i]->conf = conf;
h_tmp = bam_header_read(data[i]->fp);
bam_smpl_add(sm, fn[i], h_tmp->text);
rghash = bcf_call_add_rg(rghash, h_tmp->text, conf->pl_list);
if (conf->reg) {
bam_smpl_add(sm, fn[i], h_tmp->text);
rghash = bcf_call_add_rg(rghash, h_tmp->text, conf->pl_list);
if (conf->reg) {
gplp.plp = calloc(sm->n, sizeof(void*));
fprintf(stderr, "[%s] %d samples in %d input files\n", __func__, sm->n, n);
gplp.plp = calloc(sm->n, sizeof(void*));
fprintf(stderr, "[%s] %d samples in %d input files\n", __func__, sm->n, n);
max_depth = 8000 / sm->n;
fprintf(stderr, "<%s> Set max per-sample depth to %d\n", __func__, max_depth);
}
max_depth = 8000 / sm->n;
fprintf(stderr, "<%s> Set max per-sample depth to %d\n", __func__, max_depth);
}
bam_mplp_set_maxcnt(iter, max_depth);
while (bam_mplp_auto(iter, &tid, &pos, n_plp, plp) > 0) {
if (conf->reg && (pos < beg0 || pos >= end0)) continue; // out of the region requested
bam_mplp_set_maxcnt(iter, max_depth);
while (bam_mplp_auto(iter, &tid, &pos, n_plp, plp) > 0) {
if (conf->reg && (pos < beg0 || pos >= end0)) continue; // out of the region requested
- if (hash) {
- khint_t k;
- k = kh_get(64, hash, (uint64_t)tid<<32 | pos);
- if (k == kh_end(hash)) continue;
- }
+ if (conf->bed && tid >= 0 && !bed_overlap(conf->bed, h->target_name[tid], pos, pos+1)) continue;
group_smpl(&gplp, sm, &buf, n, fn, n_plp, plp);
_ref0 = (ref && pos < ref_len)? ref[pos] : 'N';
ref16 = bam_nt16_table[_ref0];
group_smpl(&gplp, sm, &buf, n, fn, n_plp, plp);
_ref0 = (ref && pos < ref_len)? ref[pos] : 'N';
ref16 = bam_nt16_table[_ref0];
- if (!(conf->flag&MPLP_NO_INDEL) && bcf_call_gap_prep(gplp.n, gplp.n_plp, gplp.plp, pos, bca, ref, rghash) >= 0) {
+ if (!(conf->flag&MPLP_NO_INDEL) && total_depth < max_indel_depth && bcf_call_gap_prep(gplp.n, gplp.n_plp, gplp.plp, pos, bca, ref, rghash) >= 0) {
for (i = 0; i < gplp.n; ++i)
bcf_call_glfgen(gplp.n_plp[i], gplp.plp[i], -1, bca, bcr + i);
if (bcf_call_combine(gplp.n, bcr, -1, &bc) >= 0) {
for (i = 0; i < gplp.n; ++i)
bcf_call_glfgen(gplp.n_plp[i], gplp.plp[i], -1, bca, bcr + i);
if (bcf_call_combine(gplp.n, bcr, -1, &bc) >= 0) {
for (i = 0; i < gplp.n; ++i) free(gplp.plp[i]);
free(gplp.plp); free(gplp.n_plp); free(gplp.m_plp);
bcf_call_del_rghash(rghash);
for (i = 0; i < gplp.n; ++i) free(gplp.plp[i]);
free(gplp.plp); free(gplp.n_plp); free(gplp.m_plp);
bcf_call_del_rghash(rghash);
mplp.openQ = 40; mplp.extQ = 20; mplp.tandemQ = 100;
mplp.min_frac = 0.002; mplp.min_support = 1;
mplp.flag = MPLP_NO_ORPHAN | MPLP_REALN;
mplp.openQ = 40; mplp.extQ = 20; mplp.tandemQ = 100;
mplp.min_frac = 0.002; mplp.min_support = 1;
mplp.flag = MPLP_NO_ORPHAN | MPLP_REALN;
case 'P': mplp.pl_list = strdup(optarg); break;
case 'g': mplp.flag |= MPLP_GLF; break;
case 'u': mplp.flag |= MPLP_NO_COMP | MPLP_GLF; break;
case 'P': mplp.pl_list = strdup(optarg); break;
case 'g': mplp.flag |= MPLP_GLF; break;
case 'u': mplp.flag |= MPLP_NO_COMP | MPLP_GLF; break;
case 'S': mplp.flag |= MPLP_FMT_SP; break;
case 'I': mplp.flag |= MPLP_NO_INDEL; break;
case 'E': mplp.flag |= MPLP_EXT_BAQ; break;
case 'S': mplp.flag |= MPLP_FMT_SP; break;
case 'I': mplp.flag |= MPLP_NO_INDEL; break;
case 'E': mplp.flag |= MPLP_EXT_BAQ; break;
case 'C': mplp.capQ_thres = atoi(optarg); break;
case 'M': mplp.max_mq = atoi(optarg); break;
case 'q': mplp.min_mq = atoi(optarg); break;
case 'C': mplp.capQ_thres = atoi(optarg); break;
case 'M': mplp.max_mq = atoi(optarg); break;
case 'q': mplp.min_mq = atoi(optarg); break;
case 'A': use_orphan = 1; break;
case 'F': mplp.min_frac = atof(optarg); break;
case 'm': mplp.min_support = atoi(optarg); break;
case 'A': use_orphan = 1; break;
case 'F': mplp.min_frac = atof(optarg); break;
case 'm': mplp.min_support = atoi(optarg); break;
fprintf(stderr, " -M INT cap mapping quality at INT [%d]\n", mplp.max_mq);
fprintf(stderr, " -Q INT min base quality [%d]\n", mplp.min_baseQ);
fprintf(stderr, " -q INT filter out alignment with MQ smaller than INT [%d]\n", mplp.min_mq);
fprintf(stderr, " -M INT cap mapping quality at INT [%d]\n", mplp.max_mq);
fprintf(stderr, " -Q INT min base quality [%d]\n", mplp.min_baseQ);
fprintf(stderr, " -q INT filter out alignment with MQ smaller than INT [%d]\n", mplp.min_mq);
- fprintf(stderr, " -d INT max per-sample depth [%d]\n", mplp.max_depth);
+ fprintf(stderr, " -d INT max per-BAM depth [%d]\n", mplp.max_depth);
+ fprintf(stderr, " -L INT max per-sample depth for INDEL calling [%d]\n", mplp.max_indel_depth);
fprintf(stderr, " -P STR comma separated list of platforms for indels [all]\n");
fprintf(stderr, " -o INT Phred-scaled gap open sequencing error probability [%d]\n", mplp.openQ);
fprintf(stderr, " -e INT Phred-scaled gap extension seq error probability [%d]\n", mplp.extQ);
fprintf(stderr, " -P STR comma separated list of platforms for indels [all]\n");
fprintf(stderr, " -o INT Phred-scaled gap open sequencing error probability [%d]\n", mplp.openQ);
fprintf(stderr, " -e INT Phred-scaled gap extension seq error probability [%d]\n", mplp.extQ);
fprintf(stderr, " -m INT minimum gapped reads for indel candidates [%d]\n", mplp.min_support);
fprintf(stderr, " -F FLOAT minimum fraction of gapped reads for candidates [%g]\n", mplp.min_frac);
fprintf(stderr, " -G FILE exclude read groups listed in FILE [null]\n");
fprintf(stderr, " -m INT minimum gapped reads for indel candidates [%d]\n", mplp.min_support);
fprintf(stderr, " -F FLOAT minimum fraction of gapped reads for candidates [%g]\n", mplp.min_frac);
fprintf(stderr, " -G FILE exclude read groups listed in FILE [null]\n");
fprintf(stderr, " -A use anomalous read pairs in SNP/INDEL calling\n");
fprintf(stderr, " -g generate BCF output\n");
fprintf(stderr, " -u do not compress BCF output\n");
fprintf(stderr, " -A use anomalous read pairs in SNP/INDEL calling\n");
fprintf(stderr, " -g generate BCF output\n");
fprintf(stderr, " -u do not compress BCF output\n");
if (mplp.rghash) bcf_str2id_thorough_destroy(mplp.rghash);
free(mplp.reg); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);
if (mplp.rghash) bcf_str2id_thorough_destroy(mplp.rghash);
free(mplp.reg); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);