#define MPLP_NO_COMP 0x20
#define MPLP_NO_ORPHAN 0x40
#define MPLP_REALN 0x80
-#define MPLP_FMT_DP 0x100
-#define MPLP_FMT_SP 0x200
#define MPLP_NO_INDEL 0x400
-#define MPLP_EXT_BAQ 0x800
+#define MPLP_REDO_BAQ 0x800
#define MPLP_ILLUMINA13 0x1000
#define MPLP_IGNORE_RG 0x2000
#define MPLP_PRINT_POS 0x4000
#define MPLP_PRINT_MAPQ 0x8000
+#define MPLP_PER_SAMPLE 0x10000
void *bed_read(const char *fn);
void bed_destroy(void *_h);
int bed_overlap(const void *_h, const char *chr, int beg, int end);
typedef struct {
- int max_mq, min_mq, flag, min_baseQ, capQ_thres, max_depth, max_indel_depth;
+ int max_mq, min_mq, flag, min_baseQ, capQ_thres, max_depth, max_indel_depth, fmt_flag;
int openQ, extQ, tandemQ, min_support; // for indels
double min_frac; // for indels
char *reg, *pl_list;
}
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);
+ if (has_ref && (ma->conf->flag&MPLP_REALN)) bam_prob_realn_core(b, ma->ref, (ma->conf->flag & MPLP_REDO_BAQ)? 7 : 3);
if (has_ref && ma->conf->capQ_thres > 10) {
int q = bam_cap_mapQ(b, ma->ref, ma->conf->capQ_thres);
if (q < 0) skip = 1;
bam_header_t *h_tmp;
data[i] = calloc(1, sizeof(mplp_aux_t));
data[i]->fp = strcmp(fn[i], "-") == 0? bam_dopen(fileno(stdin), "r") : bam_open(fn[i], "r");
+ if ( !data[i]->fp )
+ {
+ fprintf(stderr, "[%s] failed to open %s: %s\n", __func__, fn[i], strerror(errno));
+ exit(1);
+ }
data[i]->conf = conf;
h_tmp = bam_header_read(data[i]->fp);
data[i]->h = i? h : h_tmp; // for i==0, "h" has not been set yet
bam_index_t *idx;
idx = bam_index_load(fn[i]);
if (idx == 0) {
- fprintf(stderr, "[%s] fail to load index for %d-th input.\n", __func__, i+1);
+ fprintf(stderr, "[%s] fail to load index for %s\n", __func__, fn[i]);
exit(1);
}
if (bam_parse_region(h_tmp, conf->reg, &tid, &beg, &end) < 0) {
- fprintf(stderr, "[%s] malformatted region or wrong seqname for %d-th input.\n", __func__, i+1);
+ fprintf(stderr, "[%s] malformatted region or wrong seqname for %s\n", __func__, fn[i]);
exit(1);
}
if (i == 0) tid0 = tid, beg0 = beg, end0 = end;
bca->openQ = conf->openQ, bca->extQ = conf->extQ, bca->tandemQ = conf->tandemQ;
bca->min_frac = conf->min_frac;
bca->min_support = conf->min_support;
+ bca->per_sample_flt = conf->flag & MPLP_PER_SAMPLE;
}
if (tid0 >= 0 && conf->fai) { // region is set
ref = faidx_fetch_seq(conf->fai, h->target_name[tid0], 0, 0x7fffffff, &ref_len);
for (i = 0; i < gplp.n; ++i)
bcf_call_glfgen(gplp.n_plp[i], gplp.plp[i], ref16, bca, bcr + i);
bcf_call_combine(gplp.n, bcr, ref16, &bc);
- bcf_call2bcf(tid, pos, &bc, b, (conf->flag&(MPLP_FMT_DP|MPLP_FMT_SP))? bcr : 0,
- (conf->flag&MPLP_FMT_SP), 0, 0);
+ bcf_call2bcf(tid, pos, &bc, b, bcr, conf->fmt_flag, 0, 0);
bcf_write(bp, bh, b);
bcf_destroy(b);
// call indels
bcf_call_glfgen(gplp.n_plp[i], gplp.plp[i], -1, bca, bcr + i);
if (bcf_call_combine(gplp.n, bcr, -1, &bc) >= 0) {
b = calloc(1, sizeof(bcf1_t));
- bcf_call2bcf(tid, pos, &bc, b, (conf->flag&(MPLP_FMT_DP|MPLP_FMT_SP))? bcr : 0,
- (conf->flag&MPLP_FMT_SP), bca, ref);
+ bcf_call2bcf(tid, pos, &bc, b, bcr, conf->fmt_flag, bca, ref);
bcf_write(bp, bh, b);
bcf_destroy(b);
}
} else {
printf("%s\t%d\t%c", h->target_name[tid], pos + 1, (ref && pos < ref_len)? ref[pos] : 'N');
for (i = 0; i < n; ++i) {
- int j;
- printf("\t%d\t", n_plp[i]);
+ int j, cnt;
+ for (j = cnt = 0; j < n_plp[i]; ++j) {
+ const bam_pileup1_t *p = plp[i] + j;
+ if (bam1_qual(p->b)[p->qpos] >= conf->min_baseQ) ++cnt;
+ }
+ printf("\t%d\t", cnt);
if (n_plp[i] == 0) {
printf("*\t*"); // FIXME: printf() is very slow...
if (conf->flag & MPLP_PRINT_POS) printf("\t*");
} else {
- for (j = 0; j < n_plp[i]; ++j)
- pileup_seq(plp[i] + j, pos, ref_len, ref);
+ for (j = 0; j < n_plp[i]; ++j) {
+ const bam_pileup1_t *p = plp[i] + j;
+ if (bam1_qual(p->b)[p->qpos] >= conf->min_baseQ)
+ pileup_seq(plp[i] + j, pos, ref_len, ref);
+ }
putchar('\t');
for (j = 0; j < n_plp[i]; ++j) {
const bam_pileup1_t *p = plp[i] + j;
- int c = bam1_qual(p->b)[p->qpos] + 33;
- if (c > 126) c = 126;
- putchar(c);
+ int c = bam1_qual(p->b)[p->qpos];
+ if (c >= conf->min_baseQ) {
+ c = c + 33 < 126? c + 33 : 126;
+ putchar(c);
+ }
}
if (conf->flag & MPLP_PRINT_MAPQ) {
putchar('\t');
int nfiles = 0, use_orphan = 0;
mplp_conf_t mplp;
memset(&mplp, 0, sizeof(mplp_conf_t));
- #define MPLP_PRINT_POS 0x4000
mplp.max_mq = 60;
mplp.min_baseQ = 13;
mplp.capQ_thres = 0;
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;
- while ((c = getopt(argc, argv, "Agf:r:l:M:q:Q:uaRC:BDSd:L:b:P:o:e:h:Im:F:EG:6Os")) >= 0) {
+ while ((c = getopt(argc, argv, "Agf:r:l:M:q:Q:uaRC:BDSd:L:b:P:po:e:h:Im:F:EG:6OsV")) >= 0) {
switch (c) {
case 'f':
mplp.fai = fai_load(optarg);
case 'r': mplp.reg = strdup(optarg); break;
case 'l': mplp.bed = bed_read(optarg); break;
case 'P': mplp.pl_list = strdup(optarg); break;
+ case 'p': mplp.flag |= MPLP_PER_SAMPLE; break;
case 'g': mplp.flag |= MPLP_GLF; break;
case 'u': mplp.flag |= MPLP_NO_COMP | MPLP_GLF; break;
case 'a': mplp.flag |= MPLP_NO_ORPHAN | MPLP_REALN; break;
case 'B': mplp.flag &= ~MPLP_REALN; break;
- case 'D': mplp.flag |= MPLP_FMT_DP; break;
- case 'S': mplp.flag |= MPLP_FMT_SP; break;
+ case 'D': mplp.fmt_flag |= B2B_FMT_DP; break;
+ case 'S': mplp.fmt_flag |= B2B_FMT_SP; break;
+ case 'V': mplp.fmt_flag |= B2B_FMT_DV; break;
case 'I': mplp.flag |= MPLP_NO_INDEL; break;
- case 'E': mplp.flag |= MPLP_EXT_BAQ; break;
+ case 'E': mplp.flag |= MPLP_REDO_BAQ; break;
case '6': mplp.flag |= MPLP_ILLUMINA13; break;
case 'R': mplp.flag |= MPLP_IGNORE_RG; break;
case 's': mplp.flag |= MPLP_PRINT_MAPQ; break;
fprintf(stderr, " -b FILE list of input BAM files [null]\n");
fprintf(stderr, " -C INT parameter for adjusting mapQ; 0 to disable [0]\n");
fprintf(stderr, " -d INT max per-BAM depth to avoid excessive memory usage [%d]\n", mplp.max_depth);
- fprintf(stderr, " -E extended BAQ for higher sensitivity but lower specificity\n");
+ fprintf(stderr, " -E recalculate extended BAQ on the fly thus ignoring existing BQs\n");
fprintf(stderr, " -f FILE faidx indexed reference sequence file [null]\n");
fprintf(stderr, " -G FILE exclude read groups listed in FILE [null]\n");
fprintf(stderr, " -l FILE list of positions (chr pos) or regions (BED) [null]\n");
fprintf(stderr, " -L INT max per-sample depth for INDEL calling [%d]\n", mplp.max_indel_depth);
fprintf(stderr, " -m INT minimum gapped reads for indel candidates [%d]\n", mplp.min_support);
fprintf(stderr, " -o INT Phred-scaled gap open sequencing error probability [%d]\n", mplp.openQ);
+ fprintf(stderr, " -p apply -m and -F per-sample to increase sensitivity\n");
fprintf(stderr, " -P STR comma separated list of platforms for indels [all]\n");
fprintf(stderr, "\n");
fprintf(stderr, "Notes: Assuming diploid individuals.\n\n");
return 1;
}
+ bam_no_B = 1;
if (file_list) {
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
mpileup(&mplp,nfiles,fn);