#include "sam.h"
#include "kstring.h"
-void bam_fillmd1(bam1_t *b, char *ref, int is_equal)
+void bam_fillmd1_core(bam1_t *b, char *ref, int is_equal, int max_nm)
{
uint8_t *seq = bam1_seq(b);
uint32_t *cigar = bam1_cigar(b);
uint8_t *old_md, *old_nm;
int32_t old_nm_i = -1, nm = 0;
- old_md = bam_aux_get(b, "MD");
- old_nm = bam_aux_get(b, "NM");
- if (c->flag & BAM_FUNMAP) return;
- if (old_nm) old_nm_i = bam_aux2i(old_nm);
str = (kstring_t*)calloc(1, sizeof(kstring_t));
for (i = y = 0, x = c->pos; i < c->n_cigar; ++i) {
int j, l = cigar[i]>>4, op = cigar[i]&0xf;
}
}
ksprintf(str, "%d", u);
+ // apply max_nm
+ if (max_nm > 0 && nm >= max_nm) {
+ for (i = y = 0, x = c->pos; i < c->n_cigar; ++i) {
+ int j, l = cigar[i]>>4, op = cigar[i]&0xf;
+ if (op == BAM_CMATCH) {
+ for (j = 0; j < l; ++j) {
+ int z = y + j;
+ int c1 = bam1_seqi(seq, z), c2 = bam_nt16_table[(int)ref[x+j]];
+ if (ref[x+j] == 0) break; // out of boundary
+ if ((c1 == c2 && c1 != 15 && c2 != 15) || c1 == 0) { // a match
+ seq[z/2] |= (z&1)? 0x0f : 0xf0;
+ bam1_qual(b)[z] = 0;
+ }
+ }
+ if (j < l) break;
+ x += l; y += l;
+ } else if (op == BAM_CDEL || op == BAM_CREF_SKIP) x += l;
+ else if (op == BAM_CINS || op == BAM_CSOFT_CLIP) y += l;
+ }
+ }
+ // update NM
+ old_nm = bam_aux_get(b, "NM");
+ if (c->flag & BAM_FUNMAP) return;
+ if (old_nm) old_nm_i = bam_aux2i(old_nm);
if (!old_nm) bam_aux_append(b, "NM", 'i', 4, (uint8_t*)&nm);
else if (nm != old_nm_i) {
- fprintf(stderr, "[bam_fillmd1] different NM for read '%s': %d != %d\n", bam1_qname(b), old_nm_i, nm);
+ fprintf(stderr, "[bam_fillmd1] different NM for read '%s': %d -> %d\n", bam1_qname(b), old_nm_i, nm);
bam_aux_del(b, old_nm);
bam_aux_append(b, "NM", 'i', 4, (uint8_t*)&nm);
}
+ // update MD
+ old_md = bam_aux_get(b, "MD");
if (!old_md) bam_aux_append(b, "MD", 'Z', str->l + 1, (uint8_t*)str->s);
else {
int is_diff = 0;
break;
if (i < str->l) is_diff = 1;
} else is_diff = 1;
- if (is_diff)
- fprintf(stderr, "[bam_fillmd1] different MD for read '%s': '%s' != '%s'\n", bam1_qname(b), old_md+1, str->s);
+ if (is_diff) {
+ fprintf(stderr, "[bam_fillmd1] different MD for read '%s': '%s' -> '%s'\n", bam1_qname(b), old_md+1, str->s);
+ bam_aux_del(b, old_md);
+ bam_aux_append(b, "MD", 'Z', str->l + 1, (uint8_t*)str->s);
+ }
}
free(str->s); free(str);
}
+void bam_fillmd1(bam1_t *b, char *ref, int is_equal)
+{
+ bam_fillmd1_core(b, ref, is_equal, 0);
+}
+
int bam_fillmd(int argc, char *argv[])
{
- int c, is_equal = 0, tid = -2, ret, len, is_bam_out, is_sam_in, is_uncompressed;
+ int c, is_equal = 0, tid = -2, ret, len, is_bam_out, is_sam_in, is_uncompressed, max_nm = 0;
samfile_t *fp, *fpout = 0;
faidx_t *fai;
char *ref = 0, mode_w[8], mode_r[8];
is_bam_out = is_sam_in = is_uncompressed = 0;
mode_w[0] = mode_r[0] = 0;
strcpy(mode_r, "r"); strcpy(mode_w, "w");
- while ((c = getopt(argc, argv, "eubS")) >= 0) {
+ while ((c = getopt(argc, argv, "eubSn:")) >= 0) {
switch (c) {
case 'e': is_equal = 1; break;
case 'b': is_bam_out = 1; break;
case 'u': is_uncompressed = is_bam_out = 1; break;
case 'S': is_sam_in = 1; break;
+ case 'n': max_nm = atoi(optarg); break;
default: fprintf(stderr, "[bam_fillmd] unrecognized option '-%c'\n", c); return 1;
}
}
free(ref);
ref = fai_fetch(fai, fp->header->target_name[b->core.tid], &len);
tid = b->core.tid;
+ if (ref == 0)
+ fprintf(stderr, "[bam_fillmd] fail to find sequence '%s' in the reference.\n",
+ fp->header->target_name[tid]);
}
- bam_fillmd1(b, ref, is_equal);
+ if (ref) bam_fillmd1_core(b, ref, is_equal, max_nm);
}
samwrite(fpout, b);
}