ruby-minigraph 0.0.20.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/LICENSE.txt +21 -0
- data/README.md +62 -0
- data/ext/Rakefile +56 -0
- data/ext/cmappy/cmappy.c +7 -0
- data/ext/cmappy/cmappy.h +8 -0
- data/ext/minigraph/LICENSE.txt +23 -0
- data/ext/minigraph/Makefile +66 -0
- data/ext/minigraph/NEWS.md +317 -0
- data/ext/minigraph/README.md +207 -0
- data/ext/minigraph/algo.c +194 -0
- data/ext/minigraph/algo.h +33 -0
- data/ext/minigraph/asm-call.c +147 -0
- data/ext/minigraph/bseq.c +133 -0
- data/ext/minigraph/bseq.h +76 -0
- data/ext/minigraph/cal_cov.c +139 -0
- data/ext/minigraph/doc/example1.png +0 -0
- data/ext/minigraph/doc/example2.png +0 -0
- data/ext/minigraph/doc/examples.graffle +0 -0
- data/ext/minigraph/format.c +241 -0
- data/ext/minigraph/galign.c +140 -0
- data/ext/minigraph/gchain1.c +532 -0
- data/ext/minigraph/gcmisc.c +223 -0
- data/ext/minigraph/gfa-aug.c +260 -0
- data/ext/minigraph/gfa-base.c +526 -0
- data/ext/minigraph/gfa-bbl.c +372 -0
- data/ext/minigraph/gfa-ed.c +617 -0
- data/ext/minigraph/gfa-io.c +395 -0
- data/ext/minigraph/gfa-priv.h +154 -0
- data/ext/minigraph/gfa.h +166 -0
- data/ext/minigraph/ggen.c +182 -0
- data/ext/minigraph/ggen.h +21 -0
- data/ext/minigraph/ggsimple.c +570 -0
- data/ext/minigraph/gmap.c +211 -0
- data/ext/minigraph/index.c +230 -0
- data/ext/minigraph/kalloc.c +224 -0
- data/ext/minigraph/kalloc.h +82 -0
- data/ext/minigraph/kavl.h +414 -0
- data/ext/minigraph/kdq.h +134 -0
- data/ext/minigraph/ketopt.h +116 -0
- data/ext/minigraph/khashl.h +348 -0
- data/ext/minigraph/krmq.h +474 -0
- data/ext/minigraph/kseq.h +256 -0
- data/ext/minigraph/ksort.h +164 -0
- data/ext/minigraph/kstring.h +165 -0
- data/ext/minigraph/kthread.c +159 -0
- data/ext/minigraph/kthread.h +15 -0
- data/ext/minigraph/kvec-km.h +105 -0
- data/ext/minigraph/kvec.h +110 -0
- data/ext/minigraph/lchain.c +441 -0
- data/ext/minigraph/main.c +301 -0
- data/ext/minigraph/map-algo.c +500 -0
- data/ext/minigraph/mgpriv.h +128 -0
- data/ext/minigraph/minigraph.1 +359 -0
- data/ext/minigraph/minigraph.h +176 -0
- data/ext/minigraph/miniwfa.c +834 -0
- data/ext/minigraph/miniwfa.h +95 -0
- data/ext/minigraph/misc/mgutils.js +1451 -0
- data/ext/minigraph/misc.c +12 -0
- data/ext/minigraph/options.c +134 -0
- data/ext/minigraph/shortk.c +251 -0
- data/ext/minigraph/sketch.c +109 -0
- data/ext/minigraph/sys.c +147 -0
- data/ext/minigraph/sys.h +20 -0
- data/ext/minigraph/test/MT-chimp.fa +277 -0
- data/ext/minigraph/test/MT-human.fa +239 -0
- data/ext/minigraph/test/MT-orangA.fa +276 -0
- data/ext/minigraph/test/MT.gfa +19 -0
- data/ext/minigraph/tex/Makefile +13 -0
- data/ext/minigraph/tex/minigraph.bib +676 -0
- data/ext/minigraph/tex/minigraph.tex +986 -0
- data/ext/minigraph/tex/plots/CHM13-f1-90.bb.anno.gp +42 -0
- data/ext/minigraph/tex/plots/CHM13-f1-90.bb.anno.tbl +13 -0
- data/ext/minigraph/tex/plots/CHM13-f1-90.bb.mini-inter-none.win.gp +269 -0
- data/ext/minigraph/tex/plots/CHM13-f1-90.bb.mini-inter-none.win.sh +7 -0
- data/ext/minigraph/tex/plots/CHM13v1.cen.bed +23 -0
- data/ext/minigraph/tex/plots/CHM13v1.size +23 -0
- data/ext/minigraph/tex/plots/anno2tbl.js +40 -0
- data/ext/minigraph/tex/plots/bedutils.js +367 -0
- data/ext/minigraph/tex/plots/chr-plot.js +130 -0
- data/ext/minigraph/tex/plots/gen-anno.mak +24 -0
- data/ext/minigraph.patch +21 -0
- data/lib/minigraph/ffi/constants.rb +230 -0
- data/lib/minigraph/ffi/functions.rb +70 -0
- data/lib/minigraph/ffi/mappy.rb +8 -0
- data/lib/minigraph/ffi.rb +27 -0
- data/lib/minigraph/version.rb +5 -0
- data/lib/minigraph.rb +72 -0
- metadata +159 -0
@@ -0,0 +1,570 @@
|
|
1
|
+
#include <assert.h>
|
2
|
+
#include "mgpriv.h"
|
3
|
+
#include "gfa-priv.h"
|
4
|
+
#include "kalloc.h"
|
5
|
+
#include "bseq.h"
|
6
|
+
#include "algo.h"
|
7
|
+
#include "sys.h"
|
8
|
+
#include "ggen.h"
|
9
|
+
#include "kvec-km.h"
|
10
|
+
|
11
|
+
int32_t mg_gc_index(void *km, int min_mapq, int min_map_len, int min_depth_len, const gfa_t *g, int32_t n_seq, mg_gchains_t *const* gcs,
|
12
|
+
double *a_dens, int32_t **soff_, int32_t **qoff_, mg_intv_t **sintv_, mg_intv_t **qintv_)
|
13
|
+
{
|
14
|
+
int32_t t, i, j, max_acnt, *scnt, *soff, *qcnt, *qoff;
|
15
|
+
int64_t sum_acnt, sum_alen;
|
16
|
+
mg_intv_t *sintv, *qintv;
|
17
|
+
|
18
|
+
// count the number of intervals on each segment
|
19
|
+
KCALLOC(km, scnt, g->n_seg);
|
20
|
+
KCALLOC(km, qcnt, n_seq);
|
21
|
+
for (t = 0, max_acnt = 0; t < n_seq; ++t) {
|
22
|
+
const mg_gchains_t *gt = gcs[t];
|
23
|
+
for (i = 0; i < gt->n_gc; ++i) {
|
24
|
+
const mg_gchain_t *gc = >->gc[i];
|
25
|
+
if (gc->id != gc->parent) continue;
|
26
|
+
if (gc->blen < min_depth_len || gc->mapq < min_mapq) continue;
|
27
|
+
if (gc->n_anchor > max_acnt) max_acnt = gc->n_anchor;
|
28
|
+
++qcnt[t];
|
29
|
+
for (j = 0; j < gc->cnt; ++j)
|
30
|
+
++scnt[gt->lc[gc->off + j].v>>1];
|
31
|
+
}
|
32
|
+
}
|
33
|
+
if (max_acnt == 0) { // no gchain
|
34
|
+
kfree(km, scnt); kfree(km, qcnt);
|
35
|
+
return 0;
|
36
|
+
}
|
37
|
+
|
38
|
+
// compute soff[] and qoff[]
|
39
|
+
KMALLOC(km, soff, g->n_seg + 1);
|
40
|
+
KMALLOC(km, qoff, n_seq + 1);
|
41
|
+
for (soff[0] = 0, i = 1; i <= g->n_seg; ++i)
|
42
|
+
soff[i] = soff[i - 1] + scnt[i - 1];
|
43
|
+
for (qoff[0] = 0, i = 1; i <= n_seq; ++i)
|
44
|
+
qoff[i] = qoff[i - 1] + qcnt[i - 1];
|
45
|
+
|
46
|
+
// populate the interval list
|
47
|
+
memset(scnt, 0, 4 * g->n_seg);
|
48
|
+
memset(qcnt, 0, 4 * n_seq);
|
49
|
+
KMALLOC(km, sintv, soff[g->n_seg]);
|
50
|
+
KMALLOC(km, qintv, qoff[n_seq]);
|
51
|
+
sum_acnt = sum_alen = 0;
|
52
|
+
for (t = 0; t < n_seq; ++t) {
|
53
|
+
const mg_gchains_t *gt = gcs[t];
|
54
|
+
for (i = 0; i < gt->n_gc; ++i) {
|
55
|
+
const mg_gchain_t *gc = >->gc[i];
|
56
|
+
mg_intv_t *p;
|
57
|
+
if (gc->id != gc->parent) continue;
|
58
|
+
if (gc->blen < min_depth_len || gc->mapq < min_mapq) continue;
|
59
|
+
p = &qintv[qoff[t] + qcnt[t]];
|
60
|
+
++qcnt[t];
|
61
|
+
p->st = gc->qs, p->en = gc->qe, p->rev = 0, p->far = -1, p->i = -1;
|
62
|
+
for (j = 0; j < gc->cnt; ++j) {
|
63
|
+
const mg_llchain_t *lc = >->lc[gc->off + j];
|
64
|
+
int32_t rs, re, tmp;
|
65
|
+
if (lc->cnt > 0) { // compute start and end on the forward strand on the segment
|
66
|
+
const mg128_t *qs = >->a[lc->off];
|
67
|
+
const mg128_t *qe = >->a[lc->off + lc->cnt - 1];
|
68
|
+
int32_t rs0 = (int32_t)qs->x + 1 - (int32_t)(qs->y>>32&0xff);
|
69
|
+
int32_t re0 = (int32_t)qe->x;
|
70
|
+
assert(rs0 >= 0 && re0 > rs0 && re0 < g->seg[lc->v>>1].len);
|
71
|
+
sum_alen += re0 - rs0, sum_acnt += (qe->x>>32) - (qs->x>>32) + 1;
|
72
|
+
rs = 0, re = g->seg[lc->v>>1].len;
|
73
|
+
if (j == 0) rs = gc->p? gc->p->ss : rs0;
|
74
|
+
if (j == gc->cnt - 1) re = gc->p? gc->p->ee : re0;
|
75
|
+
if (lc->v&1) // swap rs and re
|
76
|
+
tmp = rs, rs = g->seg[lc->v>>1].len - re, re = g->seg[lc->v>>1].len - tmp;
|
77
|
+
} else rs = 0, re = g->seg[lc->v>>1].len;
|
78
|
+
p = &sintv[soff[lc->v>>1] + scnt[lc->v>>1]];
|
79
|
+
++scnt[lc->v>>1];
|
80
|
+
p->st = rs, p->en = re, p->rev = lc->v&1, p->far = -1, p->i = -1;
|
81
|
+
}
|
82
|
+
}
|
83
|
+
}
|
84
|
+
*a_dens = (double)sum_acnt / sum_alen;
|
85
|
+
|
86
|
+
// sort and index intervals
|
87
|
+
for (i = 0; i < g->n_seg; ++i) {
|
88
|
+
assert(soff[i+1] - soff[i] == scnt[i]);
|
89
|
+
mg_intv_index(soff[i+1] - soff[i], &sintv[soff[i]]);
|
90
|
+
}
|
91
|
+
kfree(km, scnt);
|
92
|
+
for (i = 0; i < n_seq; ++i) {
|
93
|
+
assert(qoff[i+1] - qoff[i] == qcnt[i]);
|
94
|
+
mg_intv_index(qoff[i+1] - qoff[i], &qintv[qoff[i]]);
|
95
|
+
}
|
96
|
+
kfree(km, qcnt);
|
97
|
+
|
98
|
+
*sintv_ = sintv, *qintv_ = qintv;
|
99
|
+
*soff_ = soff, *qoff_ = qoff;
|
100
|
+
return max_acnt;
|
101
|
+
}
|
102
|
+
|
103
|
+
/**********************
|
104
|
+
* Graph augmentation *
|
105
|
+
**********************/
|
106
|
+
|
107
|
+
void mg_ggsimple(void *km, const mg_ggopt_t *opt, gfa_t *g, int32_t n_seq, const mg_bseq1_t *seq, mg_gchains_t *const* gcs)
|
108
|
+
{
|
109
|
+
int32_t t, i, j, *soff, *qoff, max_acnt, *sc, m_ovlp = 0, *ovlp = 0, n_ins, m_ins, n_inv;
|
110
|
+
int32_t l_pseq, m_pseq;
|
111
|
+
uint64_t *meta;
|
112
|
+
mg_intv_t *sintv, *qintv;
|
113
|
+
double a_dens;
|
114
|
+
gfa_ins_t *ins;
|
115
|
+
char *pseq;
|
116
|
+
|
117
|
+
max_acnt = mg_gc_index(km, opt->min_mapq, opt->min_map_len, opt->min_depth_len, g, n_seq, gcs, &a_dens, &soff, &qoff, &sintv, &qintv);
|
118
|
+
if (max_acnt == 0) return;
|
119
|
+
|
120
|
+
// extract poorly regions
|
121
|
+
m_pseq = l_pseq = 0, pseq = 0;
|
122
|
+
m_ins = n_ins = 0, ins = 0;
|
123
|
+
n_inv = 0;
|
124
|
+
KMALLOC(km, sc, max_acnt);
|
125
|
+
KMALLOC(km, meta, max_acnt);
|
126
|
+
for (t = 0; t < n_seq; ++t) {
|
127
|
+
const mg_gchains_t *gt = gcs[t];
|
128
|
+
for (i = 0; i < gt->n_gc; ++i) {
|
129
|
+
const mg_gchain_t *gc = >->gc[i];
|
130
|
+
int32_t off_a, off_l, n_ss, far_q;
|
131
|
+
mg_msseg_t *ss;
|
132
|
+
if (gc->id != gc->parent) continue;
|
133
|
+
if (gc->blen < opt->min_map_len || gc->mapq < opt->min_mapq) continue;
|
134
|
+
assert(gc->cnt > 0);
|
135
|
+
|
136
|
+
// fill sc[]. This part achieves a similar goal to the one in mg_gchain_extra(). It makes more assumptions, but is logically simpler.
|
137
|
+
off_l = gc->off;
|
138
|
+
off_a = gt->lc[off_l].off + 1;
|
139
|
+
far_q = 0;
|
140
|
+
for (j = 1; j < gc->n_anchor; ++j, ++off_a) {
|
141
|
+
const mg128_t *q = >->a[off_a - 1], *p = >->a[off_a];
|
142
|
+
const mg_llchain_t *lc = >->lc[off_l];
|
143
|
+
int32_t s, ed = -1, off_l0 = off_l, pd, qd = (int32_t)p->y - (int32_t)q->y, c = (int32_t)(p->x>>32) - (int32_t)(q->x>>32) - 1;
|
144
|
+
if ((int32_t)q->y > far_q) far_q = (int32_t)q->y; // far_q keeps the rightmost query position seen so far
|
145
|
+
if (off_a == lc->off + lc->cnt) { // we are at the end of the current lchain
|
146
|
+
pd = g->seg[lc->v>>1].len - (int32_t)q->x - 1;
|
147
|
+
for (++off_l; off_l < gc->off + gc->cnt && gt->lc[off_l].cnt == 0; ++off_l)
|
148
|
+
pd += g->seg[gt->lc[off_l].v>>1].len;
|
149
|
+
assert(off_l < gc->off + gc->cnt);
|
150
|
+
if (gt->lc[off_l].ed >= 0) ed = gt->lc[off_l].ed;
|
151
|
+
pd += (int32_t)p->x + 1;
|
152
|
+
} else pd = (int32_t)p->x - (int32_t)q->x;
|
153
|
+
if ((opt->flag&MG_G_NO_QOVLP) && (int32_t)p->y < far_q) s = 1; // query overlap
|
154
|
+
else if (pd == qd && c == 0) s = -opt->match_pen;
|
155
|
+
else if (ed >= 0) {
|
156
|
+
int32_t min_d = pd < qd? pd : qd;
|
157
|
+
double t = 1. / (1.01 - opt->ggs_max_iden);
|
158
|
+
if (t > 10.) t = 10.;
|
159
|
+
s = (int32_t)(ed * t - min_d);
|
160
|
+
} else if (pd > qd) {
|
161
|
+
double x = qd * a_dens;
|
162
|
+
x = x > c? x : c;
|
163
|
+
s = (int32_t)(x + (pd - qd) * a_dens + .499);
|
164
|
+
} else {
|
165
|
+
s = (int32_t)(qd * a_dens + .499);
|
166
|
+
s = s > c? s : c;
|
167
|
+
}
|
168
|
+
sc[j - 1] = s;
|
169
|
+
meta[j-1] = (uint64_t)pd<<32 | off_l0;
|
170
|
+
}
|
171
|
+
|
172
|
+
// get regions to insert
|
173
|
+
ss = mg_mss_all(0, gc->n_anchor - 1, sc, 10, 0, &n_ss);
|
174
|
+
off_a = gt->lc[gc->off].off;
|
175
|
+
for (j = 0; j < n_ss; ++j) {
|
176
|
+
const mg128_t *p, *q;
|
177
|
+
int32_t st, en, ls, le, span, pd, k, n_ovlp, min_len, is_inv = 0;
|
178
|
+
gfa_ins_t I;
|
179
|
+
|
180
|
+
// find the initial positions
|
181
|
+
min_len = opt->ggs_min_end_cnt > 0? opt->ggs_min_end_cnt : 0;
|
182
|
+
if (min_len < ss[j].sc * opt->ggs_min_end_frac) min_len = ss[j].sc * opt->ggs_min_end_frac;
|
183
|
+
if (ss[j].st <= min_len || ss[j].en >= gc->n_anchor - 1 - min_len) continue; // too close to ends
|
184
|
+
st = ss[j].st, en = ss[j].en;
|
185
|
+
q = >->a[off_a + st];
|
186
|
+
p = >->a[off_a + en];
|
187
|
+
span = p->y>>32&0xff;
|
188
|
+
I.ctg = t;
|
189
|
+
ls = (int32_t)meta[st], le = (int32_t)meta[en]; // first and last lchain; CLOSED interval
|
190
|
+
assert(ls <= le);
|
191
|
+
I.v[0] = gt->lc[ls].v;
|
192
|
+
I.v[1] = gt->lc[le].v;
|
193
|
+
I.voff[0] = (int32_t)q->x + 1 - span;
|
194
|
+
I.voff[1] = (int32_t)p->x + 1;
|
195
|
+
I.coff[0] = (int32_t)q->y + 1 - span;
|
196
|
+
I.coff[1] = (int32_t)p->y + 1;
|
197
|
+
assert(I.voff[0] <= g->seg[I.v[0]>>1].len);
|
198
|
+
assert(I.voff[1] <= g->seg[I.v[1]>>1].len);
|
199
|
+
for (k = st, pd = span; k < en; ++k)
|
200
|
+
pd += meta[k]>>32;
|
201
|
+
|
202
|
+
if (I.coff[0] > I.coff[1]) {
|
203
|
+
if (mg_verbose >= 2 && pd + (I.coff[0] - I.coff[1]) >= opt->min_var_len)
|
204
|
+
fprintf(stderr, "[W::%s] query overlap on gchain %d: [%c%s:%d,%c%s:%d|%d] <=> %s:[%d,%d|%d]\n", __func__, t, "><"[I.v[0]&1], g->seg[I.v[0]>>1].name, I.voff[0], "><"[I.v[1]&1], g->seg[I.v[1]>>1].name, I.voff[1], pd, seq[t].name, I.coff[0], I.coff[1], I.coff[1] - I.coff[0]);
|
205
|
+
continue; // such overlap can't be properly resolved
|
206
|
+
}
|
207
|
+
pd -= gfa_ins_adj(g, opt->ggs_shrink_pen, &I, seq[t].seq);
|
208
|
+
|
209
|
+
min_len = pd > I.coff[1] - I.coff[0]? pd : I.coff[1] - I.coff[0];
|
210
|
+
if (I.coff[0] <= min_len || I.coff[1] >= seq[t].l_seq - min_len) continue; // test if the event is close to ends again
|
211
|
+
|
212
|
+
// filtering
|
213
|
+
if (I.coff[1] - I.coff[0] < opt->min_var_len && pd < opt->min_var_len)
|
214
|
+
continue;
|
215
|
+
for (k = I.coff[0]; k < I.coff[1]; ++k) { // test ambiguous bases
|
216
|
+
int c = seq[t].seq[k];
|
217
|
+
if (c == 'n' || c == 'N') break;
|
218
|
+
}
|
219
|
+
if (k != I.coff[1]) continue; // no ambiguous bases on the insert
|
220
|
+
n_ovlp = mg_intv_overlap(km, qoff[t+1] - qoff[t], &qintv[qoff[t]], I.coff[0], I.coff[1], &ovlp, &m_ovlp); // test overlapping on the query
|
221
|
+
if (n_ovlp == 0) fprintf(stderr, "[W::%s] query interval %s:%d-%d is not covered\n", __func__, seq[t].name, I.coff[0], I.coff[1]);
|
222
|
+
if (n_ovlp != 1) continue;
|
223
|
+
for (k = ls; k <= le; ++k) { // find other mappings overlapping with the insert on the graph
|
224
|
+
uint32_t v = gt->lc[k].v, len = g->seg[v>>1].len;
|
225
|
+
int32_t s = 0, e = len, tmp;
|
226
|
+
if (k == ls) s = (int32_t)gt->a[off_a+st].x + 1 - (int32_t)(gt->a[off_a+st].y>>32&0xff);
|
227
|
+
if (k == le) e = (int32_t)gt->a[off_a+en].x + 1;
|
228
|
+
if (v&1) tmp = s, s = len - e, e = len - tmp;
|
229
|
+
n_ovlp = mg_intv_overlap(km, soff[(v>>1)+1] - soff[v>>1], &sintv[soff[v>>1]], s, e, &ovlp, &m_ovlp);
|
230
|
+
if (n_ovlp == 0) fprintf(stderr, "[W::%s] graph interval %s:%d-%d is not covered by %s:%d-%d\n", __func__, g->seg[v>>1].name, s, e, seq[t].name, I.coff[0], I.coff[1]); // this should be an assert()
|
231
|
+
if (n_ovlp != 1) break;
|
232
|
+
}
|
233
|
+
if (k <= le) continue;
|
234
|
+
if (pd - (I.coff[1] - I.coff[0]) < opt->min_var_len && (I.coff[1] - I.coff[0]) - pd < opt->min_var_len) { // if length difference > min_var_len, just insert
|
235
|
+
int32_t qd = I.coff[1] - I.coff[0], mlen, blen, score;
|
236
|
+
l_pseq = mg_path2seq(km, g, gt, ls, le, I.voff, &pseq, &m_pseq);
|
237
|
+
score = mg_wfa_cmp(km, l_pseq, pseq, qd, &seq[t].seq[I.coff[0]], 5000, &mlen, &blen);
|
238
|
+
if (score > 0) {
|
239
|
+
if (mlen > blen * opt->ggs_max_iden) continue; // make sure k-mer identity is small enough
|
240
|
+
if (blen - mlen < opt->min_var_len * opt->ggs_max_iden) continue;
|
241
|
+
} else if (!(opt->flag & MG_G_NO_INV)) {
|
242
|
+
mg_revcomp_seq(l_pseq, pseq);
|
243
|
+
score = mg_wfa_cmp(km, l_pseq, pseq, qd, &seq[t].seq[I.coff[0]], 5000, &mlen, &blen);
|
244
|
+
if (score > 0 && mlen > blen * opt->ggs_min_inv_iden) is_inv = 1;
|
245
|
+
}
|
246
|
+
}
|
247
|
+
if (mg_dbg_flag & MG_DBG_INSERT) {
|
248
|
+
int32_t mlen, blen, score, qd = I.coff[1] - I.coff[0];
|
249
|
+
l_pseq = mg_path2seq(km, g, gt, ls, le, I.voff, &pseq, &m_pseq);
|
250
|
+
fprintf(stderr, "IN\t[%c%s:%d,%c%s:%d|%d] <=> %s:[%d,%d|%d] inv:%d\n", "><"[I.v[0]&1], g->seg[I.v[0]>>1].name, I.voff[0], "><"[I.v[1]&1], g->seg[I.v[1]>>1].name, I.voff[1], pd, seq[t].name, I.coff[0], I.coff[1], I.coff[1] - I.coff[0], is_inv);
|
251
|
+
fprintf(stderr, "IP\t%s\nIQ\t", pseq);
|
252
|
+
fwrite(&seq[t].seq[I.coff[0]], 1, qd, stderr);
|
253
|
+
if (pd - qd < opt->min_var_len && qd - pd < opt->min_var_len) {
|
254
|
+
score = mg_wfa_cmp(km, l_pseq, pseq, qd, &seq[t].seq[I.coff[0]], 5000, &mlen, &blen);
|
255
|
+
} else score = -1, mlen = 0, blen = pd > qd? pd : qd;
|
256
|
+
fprintf(stderr, "\nIS\t%d==%d\tnwcmp:%d\tmlen:%d\tblen:%d\n", pd, l_pseq, score, mlen, blen);
|
257
|
+
}
|
258
|
+
if (is_inv) { // turn one inversion to two events
|
259
|
+
gfa_ins_t I_inv[2];
|
260
|
+
I_inv[0].ctg = I_inv[1].ctg = I.ctg;
|
261
|
+
// the first event
|
262
|
+
I_inv[0].coff[0] = I_inv[0].coff[1] = I.coff[0];
|
263
|
+
I_inv[0].v[0] = I.v[0];
|
264
|
+
I_inv[0].voff[0] = I.voff[0];
|
265
|
+
I_inv[0].v[1] = I.v[1]^1;
|
266
|
+
I_inv[0].voff[1] = g->seg[I.v[1]>>1].len - I.voff[1];
|
267
|
+
// the second event
|
268
|
+
I_inv[1].coff[0] = I_inv[1].coff[1] = I.coff[1];
|
269
|
+
I_inv[1].v[0] = I.v[0]^1;
|
270
|
+
I_inv[1].voff[0] = g->seg[I.v[0]>>1].len - I.voff[0];
|
271
|
+
I_inv[1].v[1] = I.v[1];
|
272
|
+
I_inv[1].voff[1] = I.voff[1];
|
273
|
+
// insert
|
274
|
+
if (n_ins == m_ins) KEXPAND(km, ins, m_ins);
|
275
|
+
ins[n_ins++] = I_inv[0];
|
276
|
+
if (n_ins == m_ins) KEXPAND(km, ins, m_ins);
|
277
|
+
ins[n_ins++] = I_inv[1];
|
278
|
+
++n_inv;
|
279
|
+
} else {
|
280
|
+
if (n_ins == m_ins) KEXPAND(km, ins, m_ins);
|
281
|
+
ins[n_ins++] = I;
|
282
|
+
}
|
283
|
+
}
|
284
|
+
kfree(0, ss);
|
285
|
+
}
|
286
|
+
}
|
287
|
+
kfree(km, pseq);
|
288
|
+
kfree(km, ovlp);
|
289
|
+
kfree(km, sc);
|
290
|
+
kfree(km, meta);
|
291
|
+
kfree(km, soff); kfree(km, qoff);
|
292
|
+
kfree(km, sintv); kfree(km, qintv);
|
293
|
+
|
294
|
+
if (n_ins > 0) {
|
295
|
+
char **names, **seqs;
|
296
|
+
KMALLOC(km, names, n_seq);
|
297
|
+
KMALLOC(km, seqs, n_seq);
|
298
|
+
for (i = 0; i < n_seq; ++i)
|
299
|
+
names[i] = seq[i].name, seqs[i] = seq[i].seq;
|
300
|
+
n_ins = gfa_ins_filter(g, n_ins, ins);
|
301
|
+
gfa_augment(g, n_ins, ins, n_seq, (const char*const*)names, (const char*const*)seqs);
|
302
|
+
kfree(km, ins);
|
303
|
+
kfree(km, names);
|
304
|
+
kfree(km, seqs);
|
305
|
+
}
|
306
|
+
if (mg_verbose >= 3)
|
307
|
+
fprintf(stderr, "[M::%s::%.3f*%.2f] inserted %d events, including %d inversions\n", __func__,
|
308
|
+
realtime() - mg_realtime0, cputime() / (realtime() - mg_realtime0), n_ins, n_inv);
|
309
|
+
}
|
310
|
+
|
311
|
+
/**********************
|
312
|
+
* Graph augmentation *
|
313
|
+
**********************/
|
314
|
+
|
315
|
+
typedef struct {
|
316
|
+
int32_t lc, vo, qo, po, len, op, sc;
|
317
|
+
} ed_intv_t;
|
318
|
+
|
319
|
+
static int32_t gg_count_intv(const gfa_t *g, const mg_gchains_t *gt, int32_t i)
|
320
|
+
{
|
321
|
+
const mg_gchain_t *gc = >->gc[i];
|
322
|
+
int32_t j, l = gc->off, x = gc->ps, n = 0;
|
323
|
+
assert(gc->p);
|
324
|
+
for (j = 0; j < gc->p->n_cigar; ++j) {
|
325
|
+
int32_t op = gc->p->cigar[j]&0xf, len = gc->p->cigar[j]>>4, rl = len;
|
326
|
+
assert(op == 1 || op == 2 || op == 7 || op == 8);
|
327
|
+
if (op == 2 || op == 7 || op == 8) {
|
328
|
+
while (x + rl > g->seg[gt->lc[l].v>>1].len) {
|
329
|
+
rl -= g->seg[gt->lc[l].v>>1].len - x;
|
330
|
+
++n, ++l, x = 0;
|
331
|
+
}
|
332
|
+
x += rl;
|
333
|
+
}
|
334
|
+
++n;
|
335
|
+
}
|
336
|
+
return n;
|
337
|
+
}
|
338
|
+
|
339
|
+
static void gg_write_intv(const gfa_t *g, const mg_gchains_t *gt, int32_t i, ed_intv_t *intv)
|
340
|
+
{
|
341
|
+
const mg_gchain_t *gc = >->gc[i];
|
342
|
+
int32_t j, l = gc->off, pl = 0, x = gc->ps, y = gc->qs, n = 0;
|
343
|
+
ed_intv_t *p;
|
344
|
+
assert(gc->p);
|
345
|
+
for (j = 0; j < gc->p->n_cigar; ++j) {
|
346
|
+
int32_t op = gc->p->cigar[j]&0xf, len = gc->p->cigar[j]>>4, rl = len;
|
347
|
+
if (op == 2 || op == 7 || op == 8) {
|
348
|
+
while (x + rl > g->seg[gt->lc[l].v>>1].len) {
|
349
|
+
p = &intv[n++];
|
350
|
+
p->lc = l, p->vo = x, p->qo = y, p->po = pl, p->len = g->seg[gt->lc[l].v>>1].len - x, p->op = op;
|
351
|
+
if (op == 7 || op == 8) y += p->len;
|
352
|
+
rl -= p->len, pl += p->len, ++l, x = 0;
|
353
|
+
}
|
354
|
+
}
|
355
|
+
p = &intv[n++];
|
356
|
+
p->lc = l, p->vo = x, p->qo = y, p->po = pl, p->len = rl, p->op = op;
|
357
|
+
if (op == 7 || op == 8) x += rl, y += rl, pl += rl;
|
358
|
+
else if (op == 1) y += rl;
|
359
|
+
else if (op == 2) x += rl, pl += rl;
|
360
|
+
}
|
361
|
+
assert(y == gc->qe && pl == gc->pe - gc->ps);
|
362
|
+
}
|
363
|
+
|
364
|
+
static void gg_score_intv(int32_t n_intv, ed_intv_t *intv)
|
365
|
+
{
|
366
|
+
int32_t j;
|
367
|
+
for (j = 0; j < n_intv; ++j) {
|
368
|
+
int32_t s;
|
369
|
+
if (intv[j].op == 7)
|
370
|
+
s = intv[j].len >= 10? -intv[j].len : 0;
|
371
|
+
else s = intv[j].len;
|
372
|
+
intv[j].sc = s;
|
373
|
+
}
|
374
|
+
}
|
375
|
+
|
376
|
+
static void gg_merge_seg(const ed_intv_t *intv, int32_t n_ss, mg_msseg_t *ss)
|
377
|
+
{
|
378
|
+
int32_t j0, j;
|
379
|
+
for (j0 = 0, j = 1; j < n_ss; ++j) {
|
380
|
+
mg_msseg_t *s0 = &ss[j0], *s1 = &ss[j];
|
381
|
+
int32_t i, mid = 0;
|
382
|
+
for (i = s0->en + 1; i < s1->st; ++i)
|
383
|
+
mid += intv[i].sc;
|
384
|
+
//fprintf(stderr, "XX\t%d\t%d\t%d\t%d\t%d\t%d\n", j, s0->sc, mid, s1->sc, s0->en+1, s1->st);
|
385
|
+
if (-mid < s0->sc * 0.2 && -mid < s1->sc * 0.2) { // FIXME: mid is sometimes 0
|
386
|
+
s0->en = s1->en, s0->sc += s1->sc + mid;
|
387
|
+
s1->st = s1->en, s1->sc = 0;
|
388
|
+
} else j0 = j;
|
389
|
+
}
|
390
|
+
}
|
391
|
+
|
392
|
+
void mg_ggsimple_cigar(void *km, const mg_ggopt_t *opt, gfa_t *g, int32_t n_seq, const mg_bseq1_t *seq, mg_gchains_t *const* gcs)
|
393
|
+
{
|
394
|
+
int32_t t, i, *soff, *qoff, max_acnt, m_ovlp = 0, *ovlp = 0, n_ins = 0, m_ins, n_inv;
|
395
|
+
int32_t l_pseq, m_pseq;
|
396
|
+
mg_intv_t *sintv, *qintv;
|
397
|
+
double a_dens;
|
398
|
+
gfa_ins_t *ins;
|
399
|
+
char *pseq;
|
400
|
+
|
401
|
+
max_acnt = mg_gc_index(km, opt->min_mapq, opt->min_map_len, opt->min_depth_len, g, n_seq, gcs, &a_dens, &soff, &qoff, &sintv, &qintv);
|
402
|
+
if (max_acnt == 0) return;
|
403
|
+
|
404
|
+
// extract poorly regions
|
405
|
+
m_pseq = l_pseq = 0, pseq = 0;
|
406
|
+
m_ins = n_ins = 0, ins = 0;
|
407
|
+
n_inv = 0;
|
408
|
+
for (t = 0; t < n_seq; ++t) {
|
409
|
+
const mg_gchains_t *gt = gcs[t];
|
410
|
+
for (i = 0; i < gt->n_gc; ++i) {
|
411
|
+
const mg_gchain_t *gc = >->gc[i];
|
412
|
+
int32_t j, n_ss, n_intv, *sc;
|
413
|
+
ed_intv_t *intv;
|
414
|
+
mg_msseg_t *ss;
|
415
|
+
if (gc->id != gc->parent) continue;
|
416
|
+
if (gc->p == 0 || gc->blen < opt->min_map_len || gc->mapq < opt->min_mapq) continue;
|
417
|
+
assert(gc->cnt > 0);
|
418
|
+
|
419
|
+
n_intv = gg_count_intv(g, gt, i);
|
420
|
+
KCALLOC(km, intv, n_intv);
|
421
|
+
gg_write_intv(g, gt, i, intv);
|
422
|
+
gg_score_intv(n_intv, intv);
|
423
|
+
KCALLOC(km, sc, n_intv);
|
424
|
+
for (j = 0; j < n_intv; ++j) sc[j] = intv[j].sc;
|
425
|
+
ss = mg_mss_all(0, n_intv, sc, opt->min_var_len, 2 * opt->min_var_len, &n_ss);
|
426
|
+
gg_merge_seg(intv, n_ss, ss);
|
427
|
+
|
428
|
+
// get regions to insert
|
429
|
+
for (j = 0; j < n_ss; ++j) {
|
430
|
+
int32_t st, en, pd, k, n_ovlp, min_len, is_inv = 0, ls, le;
|
431
|
+
gfa_ins_t I;
|
432
|
+
ed_intv_t *is, *ie;
|
433
|
+
|
434
|
+
// find the initial positions
|
435
|
+
st = ss[j].st, en = ss[j].en; // this is a CLOSED interval
|
436
|
+
if (st == en) continue;
|
437
|
+
is = &intv[st], ie = &intv[en - 1];
|
438
|
+
assert(is->op != 7 && ie->op != 7);
|
439
|
+
|
440
|
+
ls = is->lc, le = ie->lc;
|
441
|
+
I.ctg = t;
|
442
|
+
I.v[0] = gt->lc[ls].v;
|
443
|
+
I.v[1] = gt->lc[le].v;
|
444
|
+
I.voff[0] = is->vo;
|
445
|
+
I.voff[1] = ie->vo + (ie->op != 1? ie->len : 0);
|
446
|
+
I.coff[0] = is->qo;
|
447
|
+
I.coff[1] = ie->qo + (ie->op != 2? ie->len : 0);
|
448
|
+
assert(I.voff[0] <= g->seg[I.v[0]>>1].len);
|
449
|
+
assert(I.voff[1] <= g->seg[I.v[1]>>1].len);
|
450
|
+
|
451
|
+
if (I.voff[0] == 0) { // if an insert starts at pos 0, make it start at the end of the previous vertex in the chain
|
452
|
+
assert(ls - 1 >= gc->off);
|
453
|
+
I.v[0] = gt->lc[--ls].v;
|
454
|
+
I.voff[0] = g->seg[I.v[0]>>1].len;
|
455
|
+
}
|
456
|
+
if (I.voff[1] == g->seg[I.v[1]>>1].len) { // if an insert ends at the end of the vertex, make it end at the beginning of the next vertex
|
457
|
+
assert(le + 1 < gc->off + gc->cnt);
|
458
|
+
I.v[1] = gt->lc[++le].v;
|
459
|
+
I.voff[1] = 0;
|
460
|
+
}
|
461
|
+
|
462
|
+
pd = ie->po + (ie->op != 1? ie->len : 0) - is->po;
|
463
|
+
pd -= gfa_ins_adj(g, opt->ggs_shrink_pen, &I, seq[t].seq);
|
464
|
+
|
465
|
+
min_len = pd > I.coff[1] - I.coff[0]? pd : I.coff[1] - I.coff[0];
|
466
|
+
if (I.coff[0] <= min_len || I.coff[1] >= seq[t].l_seq - min_len) continue; // test if the event is close to ends again
|
467
|
+
|
468
|
+
// filtering
|
469
|
+
if (I.coff[1] - I.coff[0] < opt->min_var_len && pd < opt->min_var_len)
|
470
|
+
continue;
|
471
|
+
for (k = I.coff[0]; k < I.coff[1]; ++k) { // test ambiguous bases
|
472
|
+
int c = seq[t].seq[k];
|
473
|
+
if (c == 'n' || c == 'N') break;
|
474
|
+
}
|
475
|
+
if (k != I.coff[1]) continue; // no ambiguous bases on the insert
|
476
|
+
n_ovlp = mg_intv_overlap(km, qoff[t+1] - qoff[t], &qintv[qoff[t]], I.coff[0], I.coff[1], &ovlp, &m_ovlp); // test overlapping on the query
|
477
|
+
if (n_ovlp == 0) fprintf(stderr, "[W::%s] query interval %s:%d-%d is not covered\n", __func__, seq[t].name, I.coff[0], I.coff[1]);
|
478
|
+
if (n_ovlp != 1) continue;
|
479
|
+
for (k = is->lc; k <= ie->lc; ++k) { // find other mappings overlapping with the insert on the graph
|
480
|
+
uint32_t v = gt->lc[k].v, len = g->seg[v>>1].len;
|
481
|
+
int32_t s = 0, e = len, tmp;
|
482
|
+
if (k == is->lc) s = is->vo;
|
483
|
+
if (k == ie->lc) e = ie->vo + (ie->op != 1? ie->len : 0);
|
484
|
+
if (v&1) tmp = s, s = len - e, e = len - tmp;
|
485
|
+
if (s == e) {
|
486
|
+
if (s == 0) ++e;
|
487
|
+
else --s;
|
488
|
+
}
|
489
|
+
n_ovlp = mg_intv_overlap(km, soff[(v>>1)+1] - soff[v>>1], &sintv[soff[v>>1]], s, e, &ovlp, &m_ovlp);
|
490
|
+
if (n_ovlp == 0) fprintf(stderr, "[W::%s] graph interval %c%s:%d-%d is not covered by %s:%d-%d\n", __func__, "><"[v&1], g->seg[v>>1].name, s, e, seq[t].name, I.coff[0], I.coff[1]); // this should be an assert()
|
491
|
+
if (n_ovlp != 1) break;
|
492
|
+
}
|
493
|
+
if (k <= ie->lc) continue;
|
494
|
+
if (pd - (I.coff[1] - I.coff[0]) < opt->min_var_len && (I.coff[1] - I.coff[0]) - pd < opt->min_var_len) { // if length difference > min_var_len, just insert
|
495
|
+
int32_t qd = I.coff[1] - I.coff[0], mlen, blen, score = 0;
|
496
|
+
l_pseq = mg_path2seq(km, g, gt, ls, le, I.voff, &pseq, &m_pseq);
|
497
|
+
score = mg_wfa_cmp(km, l_pseq, pseq, qd, &seq[t].seq[I.coff[0]], 5000, &mlen, &blen);
|
498
|
+
if (score > 0) {
|
499
|
+
if (mlen > blen * opt->ggs_max_iden) continue; // make sure k-mer identity is small enough
|
500
|
+
if (blen - mlen < opt->min_var_len * opt->ggs_max_iden) continue;
|
501
|
+
} else if (!(opt->flag & MG_G_NO_INV)) {
|
502
|
+
mg_revcomp_seq(l_pseq, pseq);
|
503
|
+
score = mg_wfa_cmp(km, l_pseq, pseq, qd, &seq[t].seq[I.coff[0]], 5000, &mlen, &blen);
|
504
|
+
if (score > 0 && mlen > blen * opt->ggs_min_inv_iden) is_inv = 1;
|
505
|
+
}
|
506
|
+
}
|
507
|
+
if (mg_dbg_flag & MG_DBG_INSERT) {
|
508
|
+
int32_t mlen, blen, score, qd = I.coff[1] - I.coff[0];
|
509
|
+
l_pseq = mg_path2seq(km, g, gt, ls, le, I.voff, &pseq, &m_pseq);
|
510
|
+
fprintf(stderr, "IN\t[%c%s:%d,%c%s:%d|%d] <=> %s:[%d,%d|%d] inv:%d\n", "><"[I.v[0]&1], g->seg[I.v[0]>>1].name, I.voff[0], "><"[I.v[1]&1], g->seg[I.v[1]>>1].name, I.voff[1], pd, seq[t].name, I.coff[0], I.coff[1], I.coff[1] - I.coff[0], is_inv);
|
511
|
+
fprintf(stderr, "IP\t%s\nIQ\t", pseq);
|
512
|
+
fwrite(&seq[t].seq[I.coff[0]], 1, qd, stderr);
|
513
|
+
if (pd - qd < opt->min_var_len && qd - pd < opt->min_var_len) {
|
514
|
+
score = mg_wfa_cmp(km, l_pseq, pseq, qd, &seq[t].seq[I.coff[0]], 5000, &mlen, &blen);
|
515
|
+
} else score = -1, mlen = 0, blen = pd > qd? pd : qd;
|
516
|
+
fprintf(stderr, "\nIS\t%d==%d\tnwcmp:%d\tmlen:%d\tblen:%d\n", pd, l_pseq, score, mlen, blen);
|
517
|
+
//if (I.voff[0] == 2305301) { for (k = st; k < en; ++k) fprintf(stderr, "%d%c", intv[k].len, "MIDNSHP=XB"[intv[k].op]); fprintf(stderr, "\n"); }
|
518
|
+
}
|
519
|
+
if (is_inv) { // turn one inversion to two events
|
520
|
+
gfa_ins_t I_inv[2];
|
521
|
+
I_inv[0].ctg = I_inv[1].ctg = I.ctg;
|
522
|
+
// the first event
|
523
|
+
I_inv[0].coff[0] = I_inv[0].coff[1] = I.coff[0];
|
524
|
+
I_inv[0].v[0] = I.v[0];
|
525
|
+
I_inv[0].voff[0] = I.voff[0];
|
526
|
+
I_inv[0].v[1] = I.v[1]^1;
|
527
|
+
I_inv[0].voff[1] = g->seg[I.v[1]>>1].len - I.voff[1];
|
528
|
+
// the second event
|
529
|
+
I_inv[1].coff[0] = I_inv[1].coff[1] = I.coff[1];
|
530
|
+
I_inv[1].v[0] = I.v[0]^1;
|
531
|
+
I_inv[1].voff[0] = g->seg[I.v[0]>>1].len - I.voff[0];
|
532
|
+
I_inv[1].v[1] = I.v[1];
|
533
|
+
I_inv[1].voff[1] = I.voff[1];
|
534
|
+
// insert
|
535
|
+
if (n_ins == m_ins) KEXPAND(km, ins, m_ins);
|
536
|
+
ins[n_ins++] = I_inv[0];
|
537
|
+
if (n_ins == m_ins) KEXPAND(km, ins, m_ins);
|
538
|
+
ins[n_ins++] = I_inv[1];
|
539
|
+
++n_inv;
|
540
|
+
} else {
|
541
|
+
if (n_ins == m_ins) KEXPAND(km, ins, m_ins);
|
542
|
+
ins[n_ins++] = I;
|
543
|
+
}
|
544
|
+
}
|
545
|
+
kfree(0, ss); // this is allocated from malloc() inside mg_mss_all()
|
546
|
+
kfree(km, intv);
|
547
|
+
kfree(km, sc);
|
548
|
+
}
|
549
|
+
}
|
550
|
+
kfree(km, pseq);
|
551
|
+
kfree(km, ovlp);
|
552
|
+
kfree(km, soff); kfree(km, qoff);
|
553
|
+
kfree(km, sintv); kfree(km, qintv);
|
554
|
+
|
555
|
+
if (n_ins > 0) {
|
556
|
+
char **names, **seqs;
|
557
|
+
KMALLOC(km, names, n_seq);
|
558
|
+
KMALLOC(km, seqs, n_seq);
|
559
|
+
for (i = 0; i < n_seq; ++i)
|
560
|
+
names[i] = seq[i].name, seqs[i] = seq[i].seq;
|
561
|
+
n_ins = gfa_ins_filter(g, n_ins, ins);
|
562
|
+
gfa_augment(g, n_ins, ins, n_seq, (const char*const*)names, (const char*const*)seqs);
|
563
|
+
kfree(km, ins);
|
564
|
+
kfree(km, names);
|
565
|
+
kfree(km, seqs);
|
566
|
+
}
|
567
|
+
if (mg_verbose >= 3)
|
568
|
+
fprintf(stderr, "[M::%s::%.3f*%.2f] inserted %d events, including %d inversions\n", __func__,
|
569
|
+
realtime() - mg_realtime0, cputime() / (realtime() - mg_realtime0), n_ins, n_inv);
|
570
|
+
}
|