r1251: read junctions for jumps

This commit is contained in:
Heng Li
2025-04-04 16:23:19 -04:00
parent d9d3c0cc3f
commit a8c567b5e9
3 changed files with 50 additions and 6 deletions
+46 -3
View File
@@ -42,6 +42,11 @@ typedef struct mm_idx_intv_s {
mm_idx_intv1_t *a;
} mm_idx_intv_t;
typedef struct mm_idx_jjump_s {
int32_t n, m;
mm_idx_jjump1_t *a;
} mm_idx_jjump_t;
mm_idx_t *mm_idx_init(int w, int k, int b, int flag)
{
mm_idx_t *mi;
@@ -72,6 +77,11 @@ void mm_idx_destroy(mm_idx_t *mi)
free(mi->I[i].a);
free(mi->I);
}
if (mi->J) {
for (i = 0; i < mi->n_seq; ++i)
free(mi->J[i].a);
free(mi->J);
}
if (!mi->km) {
for (i = 0; i < mi->n_seq; ++i)
free(mi->seq[i].name);
@@ -669,6 +679,9 @@ KRADIX_SORT_INIT(bed, mm_idx_intv1_t, sort_key_bed, 4)
#define sort_key_end(a) ((a).en)
KRADIX_SORT_INIT(end, mm_idx_intv1_t, sort_key_bed, 4)
#define sort_key_jj(a) ((a).off)
KRADIX_SORT_INIT(jj, mm_idx_jjump1_t, sort_key_jj, 4)
mm_idx_intv_t *mm_idx_read_bed(const mm_idx_t *mi, const char *fn, int read_junc)
{
gzFile fp;
@@ -780,11 +793,41 @@ static mm_idx_intv_t *mm_idx_bed_read_core(const mm_idx_t *mi, const char *fn, i
return I;
}
int mm_idx_bed_read2(mm_idx_t *mi, const char *fn, int read_junc, int for_score, int for_jump)
{
int32_t i;
mm_idx_intv_t *I;
if (mi->h == 0) mm_idx_index_name(mi);
I = mm_idx_bed_read_core(mi, fn, read_junc);
if (I == 0) return 0;
if (for_jump) {
mm_idx_jjump_t *J;
J = CALLOC(mm_idx_jjump_t, mi->n_seq);
for (i = 0; i < mi->n_seq; ++i) {
int32_t j, k;
mm_idx_intv_t *intv = &I[i];
mm_idx_jjump_t *jj = &J[i];
jj->n = intv->n * 2;
jj->a = CALLOC(mm_idx_jjump1_t, jj->n);
for (j = k = 0; j < intv->n; ++j) {
jj->a[k].off = intv->a[j].st, jj->a[k].off2 = intv->a[j].en, jj->a[k++].strand = intv->a[j].strand;
jj->a[k].off = intv->a[j].en, jj->a[k].off2 = intv->a[j].st, jj->a[k++].strand = intv->a[j].strand;
}
radix_sort_jj(jj->a, jj->a + jj->n);
}
mi->J = J;
}
if (!for_score) {
for (i = 0; i < mi->n_seq; ++i)
free(I[i].a);
free(I);
} else mi->I = I;
return 0;
}
int mm_idx_bed_read(mm_idx_t *mi, const char *fn, int read_junc)
{
if (mi->h == 0) mm_idx_index_name(mi);
mi->I = mm_idx_bed_read_core(mi, fn, read_junc);
return 0;
return mm_idx_bed_read2(mi, fn, read_junc, 1, 0);
}
int mm_idx_bed_junc(const mm_idx_t *mi, int32_t ctg, int32_t st, int32_t en, uint8_t *s)
+2 -1
View File
@@ -5,7 +5,7 @@
#include <stdio.h>
#include <sys/types.h>
#define MM_VERSION "2.28-r1249-dirty"
#define MM_VERSION "2.28-r1251-dirty"
#define MM_F_NO_DIAG (0x001LL) // no exact diagonal hit
#define MM_F_NO_DUAL (0x002LL) // skip pairs where query name is lexicographically larger than target name
@@ -93,6 +93,7 @@ typedef struct {
struct mm_idx_bucket_s *B; // index (hidden)
struct mm_idx_intv_s *I; // intervals (hidden)
struct mm_idx_spsc_s *spsc;// splice score (hidden)
struct mm_idx_jjump_s *J; // junctions to create jumps (hidden)
void *km, *h;
} mm_idx_t;
+2 -2
View File
@@ -54,8 +54,8 @@ typedef struct {
typedef struct {
int32_t off, off2;
int32_t is_rev;
} mm_jjump1_t;
int32_t strand;
} mm_idx_jjump1_t;
double cputime(void);
double realtime(void);