minibwa 0.0.0 → 0.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5d54a100aaa8439c2dc2fb9b686ca60f4898ea0873370f3dc5b58bfd5af1b7fa
-  data.tar.gz: 17911037009111f945b2f2621ccdc1e6b51928748585cc3753d6341250cca51e
+  metadata.gz: d3a36a760e0ac6885ee19e57ee672350abc63ee679541f757f9aca1bf2d1336b
+  data.tar.gz: 85da4f46eceba36f840a89207b91ae34a5bac5825a8f02cdb0e92bce0d4a38cc
 SHA512:
-  metadata.gz: 0e3a98d1f7e2652f16d3537bbbe7a2e3b89ff05dad6b5f1c166db505a465cdb1f854bbd2970bca0c1c0152f8f5fa4bc10bf7ef4a716a4c9001663bf35ec77743
-  data.tar.gz: 0c99b08294dcaaff5f896da4516c4c52df07cb74e879e58c7eea40f795235ac95b92535e7a86dec93cd900ee6257d4ea4e5e42b199fe4705fe1faa851d0cdb39
+  metadata.gz: 955d5ad73d6016241507b1f1b8b72aaa36cf7fd4004c5586c3ec943aa3d24a3bb47b6a333c84efdf1394c55a9f3bccc0726ef74d7c59e6811d5f44430db79dad
+  data.tar.gz: c3a4359e47da6b147e65c23afe0efb7407a04960c493cd8883dac9ca4a8b4edad7a766a1b225cf6aadf2f02f55643c1cd7f24f861f8f7294a8d6f4cb70c667c6

data/README.md

@@ -3,6 +3,7 @@
 [![Gem Version](https://img.shields.io/gem/v/minibwa?color=brightgreen)](https://rubygems.org/gems/minibwa)
 [![Test](https://github.com/kojix2/ruby-minibwa/actions/workflows/test.yml/badge.svg)](https://github.com/kojix2/ruby-minibwa/actions/workflows/test.yml)
 [![Lines of Code](https://img.shields.io/endpoint?url=https%3A%2F%2Ftokei.kojix2.net%2Fbadge%2Fgithub%2Fkojix2%2Fruby-minibwa%2Flines)](https://tokei.kojix2.net/github/kojix2/ruby-minibwa)
+[![DOI](https://zenodo.org/badge/1319741104.svg)](https://doi.org/10.5281/zenodo.21753646)
 
 Ruby bindings for [minibwa](https://github.com/lh3/minibwa), a short-read
 aligner combining bwa-mem seeding with minimap2 chaining and alignment.

data/ext/minibwa/minibwa/format.c

@@ -254,8 +254,13 @@ void mb_fmt_sam(void *km, kstring_t *s, const l2b_t *l2b, const mb_bseq1_t *t, i
 			kom_sprintf_lite(s, "%d\t", r_next->ts + 1);
 		} else if (r_next) { // && this_tid < 0
 			kom_sprintf_lite(s, "\t%s\t%d\t", l2b->ctg[r_next->tid].name, r_next->ts + 1);
-		} else if (this_tid >= 0) { // && r_next == NULL
-			kom_sprintf_lite(s, "\t=\t%d\t", this_pos + 1); // next segment will take r's coordinate
+		} else if (this_tid >= 0) { // && r_next == NULL. In this case, mate is unmapped and placed at this segment's primary alignment
+			const mb_hit_t *r_pri;
+			r_pri = get_sam_pri(n_h, h);
+			if (r->tid == r_pri->tid)
+				kom_sprintf_lite(s, "\t=\t%d\t", r_pri->ts + 1);
+			else
+				kom_sprintf_lite(s, "\t%s\t%d\t", l2b->ctg[r_pri->tid].name, r_pri->ts + 1);
 		} else kom_sprintf_lite(s, "\t*\t0\t"); // neither has coordinates
 		if (tlen > 0) ++tlen;
 		else if (tlen < 0) --tlen;

data/ext/minibwa/minibwa/ketopt.h

@@ -12,6 +12,7 @@ typedef struct {
 	int opt;   /* equivalent to optopt */
 	char *arg; /* equivalent to optarg */
 	int longidx; /* index of a long option; or -1 if short */
+	int erri; /* offset of the wrong option */
 	/* private variables not intended for external uses */
 	int i, pos, n_args;
 } ketopt_t;
@@ -22,7 +23,7 @@ typedef struct {
 	int val;
 } ko_longopt_t;
 
-static ketopt_t KETOPT_INIT = { 1, 0, 0, -1, 1, 0, 0 };
+static ketopt_t KETOPT_INIT = { 1, 0, 0, -1, -1, 1, 0, 0 };
 
 static void ketopt_permute(char *argv[], int j, int n) /* move argv[j] over n elements to the left */
 {
@@ -60,7 +61,7 @@ static int ketopt(ketopt_t *s, int argc, char *argv[], int permute, const char *
 		while (s->i < argc && (argv[s->i][0] != '-' || argv[s->i][1] == '\0'))
 			++s->i, ++s->n_args;
 	}
-	s->arg = 0, s->longidx = -1, i0 = s->i;
+	s->arg = 0, s->longidx = -1, s->erri = -1, i0 = s->i;
 	if (s->i >= argc || argv[s->i][0] != '-' || argv[s->i][1] == '\0') {
 		s->ind = s->i - s->n_args;
 		return -1;
@@ -82,7 +83,7 @@ static int ketopt(ketopt_t *s, int argc, char *argv[], int permute, const char *
 					else ++n_partial, o_partial = &longopts[k];
 				}
 			if (n_exact > 1 || (n_exact == 0 && n_partial > 1)) {
-				s->i++;
+				s->erri = s->i++;
 				return '?';
 			}
 			o = n_exact == 1? o_exact : n_partial == 1? o_partial : 0;
@@ -93,7 +94,7 @@ static int ketopt(ketopt_t *s, int argc, char *argv[], int permute, const char *
 					if (s->i < argc - 1) s->arg = argv[++s->i];
 					else opt = ':'; /* missing option argument */
 				}
-			}
+			} else s->erri = s->i;
 		}
 	} else { /* a short option */
 		char *p;
@@ -101,7 +102,7 @@ static int ketopt(ketopt_t *s, int argc, char *argv[], int permute, const char *
 		opt = s->opt = argv[s->i][s->pos++];
 		p = strchr((char*)ostr, opt);
 		if (p == 0) {
-			opt = '?'; /* unknown option */
+			s->erri = s->i, opt = '?'; /* unknown option */
 		} else if (p[1] == ':') {
 			if (argv[s->i][s->pos] == 0) {
 				if (s->i < argc - 1) s->arg = argv[++s->i];
@@ -115,6 +116,7 @@ static int ketopt(ketopt_t *s, int argc, char *argv[], int permute, const char *
 		if (s->n_args > 0) /* permute */
 			for (j = i0; j < s->i; ++j)
 				ketopt_permute(argv, j, s->n_args);
+		if (s->erri >= 0) s->erri -= s->n_args; /* permutation moved the option left by n_args */
 	}
 	s->ind = s->i - s->n_args;
 	return opt;

data/ext/minibwa/minibwa/map-algo.c

@@ -363,10 +363,10 @@ add_primary:
 	kfree(km, w);
 }
 
-void mb_set_sam_pri(int32_t n, mb_hit_t *r, int32_t is_primary5)
+int32_t mb_set_sam_pri(int32_t n, mb_hit_t *r, int32_t is_primary5)
 {
-	int32_t i, n_pri = 0, min_i = -1, min_qs = -1, first_i = -1;
-	if (n <= 0) return;
+	int32_t i, new_pri, n_pri = 0, min_i = -1, min_qs = -1, first_i = -1;
+	if (n <= 0) return -1;
 	for (i = 0; i < n; ++i) {
 		r[i].sam_pri = 0;
 		if (r[i].id != r[i].parent) continue;
@@ -375,8 +375,9 @@ void mb_set_sam_pri(int32_t n, mb_hit_t *r, int32_t is_primary5)
 			min_i = i, min_qs = r[i].qs;
 	}
 	assert(n_pri > 0);
-	if (is_primary5) r[min_i].sam_pri = 1;
-	else r[first_i].sam_pri = 1;
+	new_pri = is_primary5? min_i : first_i;
+	r[new_pri].sam_pri = 1;
+	return new_pri;
 }
 
 void mb_select_sub(void *km, float pri_ratio, int min_diff, int best_n, int *n_, mb_hit_t *r)
@@ -751,7 +752,7 @@ mb_hit_t **mb_map_batch(const mb_opt_t *opt, const mb_idx_t *idx, int32_t n_seq,
 	kfree(km, seq4);
 
 	// paired-end processing
-	if (is_pe && n_seq >= 2) {
+	if (is_pe && n_seq >= 2 && !(opt->flag & (MB_F_NO_PAIRING|MB_F_NO_ALN))) {
 		mb_pestat_t pes[4];
 		for (i = 0; i < 4; ++i) pes[i].failed = 1;
 		pes[1].failed = 0;

data/ext/minibwa/minibwa/mbpriv.h

@@ -68,7 +68,7 @@ int32_t mb_cal_mblen(int32_t n, const mb_anchor_t *a, int32_t *blen_);
 mb_hit_t *mb_gen_hit(void *km, uint32_t hash, int qlen, const l2b_t *l2b, int n_u, uint64_t *u, mb_anchor_t *a);
 void mb_sync_high_cov(int32_t n, mb_hit_t *h);
 void mb_set_parent(void *km, float mask_level, int mask_len, int n, mb_hit_t *r, int sub_diff, int hard_mask_level);
-void mb_set_sam_pri(int32_t n, mb_hit_t *r, int32_t is_primary5);
+int32_t mb_set_sam_pri(int32_t n, mb_hit_t *r, int32_t is_primary5);
 void mb_hit_sort(void *km, int *n_regs, mb_hit_t *r);
 void mb_sync_hits(void *km, int n_regs, mb_hit_t *regs);
 void mb_select_sub(void *km, float pri_ratio, int min_diff, int best_n, int *n_, mb_hit_t *r);

data/ext/minibwa/minibwa/minibwa.h

@@ -3,7 +3,7 @@
 
 #include <stdint.h>
 
-#define MB_VERSION "0.6-r416"
+#define MB_VERSION "0.7-r421"
 
 #define MB_F_PAF              (0x1LL)       // output in the PAF format
 #define MB_F_NO_UNMAP         (0x2LL)       // output unmapped query sequences

data/ext/minibwa/minibwa/pe.c

@@ -461,7 +461,7 @@ static int32_t mb_matesw(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_
 void mb_pair(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_t n_hit[2], mb_hit_t *hit[2], const mb_pestat_t pes[4], int32_t qlen[2], char *const qseq[2])
 {
 	const int32_t pe_bonus = 4;
-	int32_t r, i, dp_max_se[2], score_se, dp_max_se2[2], score_se2, do_matesw, is_meth = !!(opt->flag & MB_F_METH);
+	int32_t r, i, dp_max_se[2], score_se, dp_max_se2[2], score_se2, do_matesw, is_meth = !!(opt->flag & MB_F_METH), reset_sam_pri = 1;
 	mb_pairaux_t paux;
 	int32_t seed_ratio[2], min_seed_ratio;
 
@@ -487,7 +487,7 @@ void mb_pair(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_t n_hit[2],
 			mb_pair_hits(km, opt, l2b, n_hit, hit, pes, &paux); // pair again if new hits rescued
 		}
 	}
-	if (paux.n_pp == 0) goto end_pairing;
+	if (paux.n_pp == 0) goto end_pairing; // skip the rest if there are no properly paired hits
 
 	for (r = 0; r < 2; ++r) {
 		for (dp_max_se[r] = dp_max_se2[r] = 0, i = 0; i < n_hit[r]; ++i) {
@@ -502,6 +502,9 @@ void mb_pair(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_t n_hit[2],
 
 	mb_sync_high_cov(n_hit[0], hit[0]);
 	mb_sync_high_cov(n_hit[1], hit[1]);
+	for (r = 0; r < 2; ++r) // clear 0x2 as these will be re-evaluated and set in the following
+		for (i = 0; i < n_hit[r]; ++i)
+			hit[r][i].proper_pair = 0;
 	if (paux.score >= score_se - opt->pen_unpair * opt->a) { // choose the paired hits
 		int32_t mapq_pe, score2 = paux.sub_sc, diff;
 		double identity;
@@ -534,7 +537,8 @@ void mb_pair(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_t n_hit[2],
 				p->mapq = 0;
 			}
 			for (i = 0; i < n_hit[r]; ++i) { // handle other chimeric hits
-				mb_hit_t *p = &hit[r][i], *q = h[r];
+				const mb_hit_t *q = h[r];
+				mb_hit_t *p = &hit[r][i];
 				if (q != p && p->id == p->parent) { // p is a chimeric hit that is not h[r]
 					int32_t j, ol = p->qe <= q->qs || p->qs >= q->qe? 0 : (p->qe < q->qe? p->qe : q->qe) - (p->qs > q->qs? p->qs : q->qs);
 					if (ol > opt->mask_level * (p->qe - p->qs)) { // if p overlaps with h[r] a lot, make it a secondary hit
@@ -545,7 +549,24 @@ void mb_pair(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_t n_hit[2],
 					}
 				}
 			}
+			for (i = 0; i < n_hit[r]; ++i) {
+				mb_hit_t *p = &hit[r][i];
+				p->sam_pri = p->proper_pair = 0;
+				if (p->id == p->parent)
+					p->proper_pair = 1;
+			}
+			h[r]->sam_pri = 1;
+		}
+		if (opt->flag & MB_F_PRIMARY5) {
+			int32_t pri[2];
+			pri[0] = mb_set_sam_pri(n_hit[0], hit[0], 1);
+			pri[1] = mb_set_sam_pri(n_hit[1], hit[1], 1);
+			if (&hit[0][pri[0]] != h[0] || &hit[1][pri[1]] != h[1]) // if sam_pri is changed, clear flag 0x2
+				for (r = 0; r < 2; ++r)
+					for (i = 0; i < n_hit[r]; ++i)
+						hit[r][i].proper_pair = 0;
 		}
+		reset_sam_pri = 0;
 	} else { // choose the unpaired hits
 		int32_t diff = score_se - opt->pen_unpair * opt->a - paux.score;
 		int32_t mapq_pe = 6 * diff / opt->a;
@@ -554,6 +575,8 @@ void mb_pair(void *km, const mb_opt_t *opt, const l2b_t *l2b, int32_t n_hit[2],
 				hit[r][i].mapq = hit[r][i].mapq < mapq_pe? hit[r][i].mapq : mapq_pe;
 	}
 end_pairing:
-	mb_set_sam_pri(n_hit[0], hit[0], !!(opt->flag & MB_F_PRIMARY5));
-	mb_set_sam_pri(n_hit[1], hit[1], !!(opt->flag & MB_F_PRIMARY5));
+	if (reset_sam_pri) {
+		mb_set_sam_pri(n_hit[0], hit[0], !!(opt->flag & MB_F_PRIMARY5));
+		mb_set_sam_pri(n_hit[1], hit[1], !!(opt->flag & MB_F_PRIMARY5));
+	}
 }

data/lib/minibwa/version.rb

@@ -2,5 +2,5 @@
 
 module Minibwa
   # Gem version.
-  VERSION = '0.0.0'
+  VERSION = '0.0.1'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: minibwa
 version: !ruby/object:Gem::Version
-  version: 0.0.0
+  version: 0.0.1
 platform: ruby
 authors:
 - kojix2
@@ -82,7 +82,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 4.0.16
+rubygems_version: 3.6.9
 specification_version: 4
 summary: Ruby bindings for minibwa, a short-read aligner
 test_files: []