removed color-space support

which has been broken since 0.6.x

Author: lh3

Makefile

@@ -6,8 +6,7 @@ DFLAGS=		-DHAVE_PTHREAD #-D_NO_SSE2 #-D_FILE_OFFSET_BITS=64
 LOBJS=		bamlite.o utils.o bwt.o bwtio.o bwtaln.o bwtgap.o bntseq.o bwamem.o bwamem_pair.o stdaln.o \
 			bseq.o bwaseqio.o bwase.o kstring.o
 AOBJS=		QSufSort.o bwt_gen.o \
-			is.o bwtmisc.o bwtindex.o ksw.o simple_dp.o \
-			bwape.o cs2nt.o \
+			is.o bwtmisc.o bwtindex.o ksw.o bwape.o \
 			bwtsw2_core.o bwtsw2_main.o bwtsw2_aux.o bwt_lite.o \
 			bwtsw2_chain.o fastmap.o bwtsw2_pair.o
 PROG=		bwa

bwape.c

@@ -212,19 +212,6 @@ static int pairing(bwa_seq_t *p[2], pe_data_t *d, const pe_opt_t *opt, int s_mm,
 				last_pos[x.y&1][1] = x;
 			}
 		}
-	} else if (opt->type == BWA_PET_SOLID) {
-		for (i = 0; i < d->arr.n; ++i) {
-			pair64_t x = d->arr.a[i];
-			int strand = x.y>>1&1;
-			if ((strand^x.y)&1) { // push
-				int y = 1 - (x.y&1);
-				__pairing_aux(last_pos[y][1], x);
-				__pairing_aux(last_pos[y][0], x);
-			} else { // check
-				last_pos[x.y&1][0] = last_pos[x.y&1][1];
-				last_pos[x.y&1][1] = x;
-			}
-		}
 	} else {
 		fprintf(stderr, "[paring] not implemented yet!\n");
 		exit(1);
@@ -567,11 +554,11 @@ ubyte_t *bwa_paired_sw(const bntseq_t *bns, const ubyte_t *_pacseq, int n_seqs,
 			++n_tot[is_singleton];
 			cigar[0] = cigar[1] = 0;
 			n_cigar[0] = n_cigar[1] = 0;
-			if (popt->type != BWA_PET_STD && popt->type != BWA_PET_SOLID) continue; // other types of pairing is not considered
+			if (popt->type != BWA_PET_STD) continue; // other types of pairing is not considered
 			for (k = 0; k < 2; ++k) { // p[1-k] is the reference read and p[k] is the read considered to be modified
 				ubyte_t *seq;
 				if (p[1-k]->type == BWA_TYPE_NO_MATCH) continue; // if p[1-k] is unmapped, skip
-				if (popt->type == BWA_PET_STD) {
+				{ // note that popt->type == BWA_PET_STD always true; in older versions, there was a branch for color-space FF/RR reads
 					if (p[1-k]->strand == 0) { // then the mate is on the reverse strand and has larger coordinate
 						__set_rght_coor(beg[k], end[k], p[1-k], p[k]);
 						seq = p[k]->rseq;
@@ -580,17 +567,6 @@ ubyte_t *bwa_paired_sw(const bntseq_t *bns, const ubyte_t *_pacseq, int n_seqs,
 						seq = p[k]->seq;
 						seq_reverse(p[k]->len, seq, 0); // because ->seq is reversed; this will reversed back shortly
 					}
-				} else { // BWA_PET_SOLID
-					if (p[1-k]->strand == 0) { // R3-F3 pairing
-						if (k == 0) __set_left_coor(beg[k], end[k], p[1-k], p[k]); // p[k] is R3
-						else __set_rght_coor(beg[k], end[k], p[1-k], p[k]); // p[k] is F3
-						seq = p[k]->rseq;
-						seq_reverse(p[k]->len, seq, 0); // because ->seq is reversed
-					} else { // F3-R3 pairing
-						if (k == 0) __set_rght_coor(beg[k], end[k], p[1-k], p[k]); // p[k] is R3
-						else __set_left_coor(beg[k], end[k], p[1-k], p[k]); // p[k] is F3
-						seq = p[k]->seq;
-					}
 				}
 				// perform SW alignment
 				cigar[k] = bwa_sw_core(bns->l_pac, pacseq, p[k]->len, seq, &beg[k], end[k] - beg[k], &n_cigar[k], &cnt[k]);
@@ -654,7 +630,7 @@ void bwa_sai2sam_pe_core(const char *prefix, char *const fn_sa[2], char *const f
 	bwa_seq_t *seqs[2];
 	bwa_seqio_t *ks[2];
 	clock_t t;
-	bntseq_t *bns, *ntbns = 0;
+	bntseq_t *bns;
 	FILE *fp_sa[2];
 	gap_opt_t opt, opt0;
 	khint_t iter;
@@ -679,10 +655,7 @@ void bwa_sai2sam_pe_core(const char *prefix, char *const fn_sa[2], char *const f
 	opt0 = opt;
 	fread(&opt, sizeof(gap_opt_t), 1, fp_sa[1]); // overwritten!
 	ks[1] = bwa_open_reads(opt.mode, fn_fa[1]);
-	if (!(opt.mode & BWA_MODE_COMPREAD)) {
-		popt->type = BWA_PET_SOLID;
-		ntbns = bwa_open_nt(prefix);
-	} else { // for Illumina alignment only
+	{ // for Illumina alignment only
 		if (popt->is_preload) {
 			strcpy(str, prefix); strcat(str, ".bwt");  bwt = bwt_restore_bwt(str);
 			strcpy(str, prefix); strcat(str, ".sa"); bwt_restore_sa(str, bwt);
@@ -715,7 +688,7 @@ void bwa_sai2sam_pe_core(const char *prefix, char *const fn_sa[2], char *const f
 
 		fprintf(stderr, "[bwa_sai2sam_pe_core] refine gapped alignments... ");
 		for (j = 0; j < 2; ++j)
-			bwa_refine_gapped(bns, n_seqs, seqs[j], pacseq, ntbns);
+			bwa_refine_gapped(bns, n_seqs, seqs[j], pacseq);
 		fprintf(stderr, "%.2f sec\n", (float)(clock() - t) / CLOCKS_PER_SEC); t = clock();
 		if (pac == 0) free(pacseq);
 
@@ -740,7 +713,6 @@ void bwa_sai2sam_pe_core(const char *prefix, char *const fn_sa[2], char *const f
 
 	// destroy
 	bns_destroy(bns);
-	if (ntbns) bns_destroy(ntbns);
 	for (i = 0; i < 2; ++i) {
 		bwa_seq_close(ks[i]);
 		fclose(fp_sa[i]);

bwase.c

@@ -296,18 +296,12 @@ void bwa_correct_trimmed(bwa_seq_t *s)
 	s->len = s->full_len;
 }
 
-void bwa_refine_gapped(const bntseq_t *bns, int n_seqs, bwa_seq_t *seqs, ubyte_t *_pacseq, bntseq_t *ntbns)
+void bwa_refine_gapped(const bntseq_t *bns, int n_seqs, bwa_seq_t *seqs, ubyte_t *_pacseq)
 {
-	ubyte_t *pacseq, *ntpac = 0;
+	ubyte_t *pacseq;
 	int i, j;
 	kstring_t *str;
 
-	if (ntbns) { // in color space
-		ntpac = (ubyte_t*)calloc(ntbns->l_pac/4+1, 1);
-		rewind(ntbns->fp_pac);
-		fread(ntpac, 1, ntbns->l_pac/4 + 1, ntbns->fp_pac);
-	}
-
 	if (!_pacseq) {
 		pacseq = (ubyte_t*)calloc(bns->l_pac/4+1, 1);
 		rewind(bns->fp_pac);
@@ -328,47 +322,23 @@ void bwa_refine_gapped(const bntseq_t *bns, int n_seqs, bwa_seq_t *seqs, ubyte_t
 		s->cigar = bwa_refine_gapped_core(bns->l_pac, pacseq, s->len, s->strand? s->rseq : s->seq, &s->pos,
 									  (s->strand? 1 : -1) * (s->n_gapo + s->n_gape), &s->n_cigar, 1);
 	}
-#if 0
-	if (ntbns) { // in color space
-		for (i = 0; i < n_seqs; ++i) {
-			bwa_seq_t *s = seqs + i;
-			bwa_cs2nt_core(s, bns->l_pac, ntpac);
-			for (j = 0; j < s->n_multi; ++j) {
-				bwt_multi1_t *q = s->multi + j;
-				int n_cigar;
-				if (q->gap == 0) continue;
-				free(q->cigar);
-				q->cigar = bwa_refine_gapped_core(bns->l_pac, ntpac, s->len, q->strand? s->rseq : s->seq, &q->pos,
-											  (q->strand? 1 : -1) * q->gap, &n_cigar, 0);
-				q->n_cigar = n_cigar;
-			}
-			if (s->type != BWA_TYPE_NO_MATCH && s->cigar) { // update cigar again
-				free(s->cigar);
-				s->cigar = bwa_refine_gapped_core(bns->l_pac, ntpac, s->len, s->strand? s->rseq : s->seq, &s->pos,
-											  (s->strand? 1 : -1) * (s->n_gapo + s->n_gape), &s->n_cigar, 0);
-			}
-		}
-	}
-#endif
 	// generate MD tag
 	str = (kstring_t*)calloc(1, sizeof(kstring_t));
 	for (i = 0; i != n_seqs; ++i) {
 		bwa_seq_t *s = seqs + i;
 		if (s->type != BWA_TYPE_NO_MATCH) {
 			int nm;
 			s->md = bwa_cal_md1(s->n_cigar, s->cigar, s->len, s->pos, s->strand? s->rseq : s->seq,
-								bns->l_pac, ntbns? ntpac : pacseq, str, &nm);
+								bns->l_pac, pacseq, str, &nm);
 			s->nm = nm;
 		}
 	}
 	free(str->s); free(str);
 
 	// correct for trimmed reads
-	if (!ntbns) // trimming is only enabled for Illumina reads
-		for (i = 0; i < n_seqs; ++i) bwa_correct_trimmed(seqs + i);
+	for (i = 0; i < n_seqs; ++i) bwa_correct_trimmed(seqs + i);
 
 	if (!_pacseq) free(pacseq);
-	free(ntpac);
 }
 
 int64_t pos_end(const bwa_seq_t *p)
@@ -587,7 +557,7 @@ void bwa_sai2sam_se_core(const char *prefix, const char *fn_sa, const char *fn_f
 	bwa_seq_t *seqs;
 	bwa_seqio_t *ks;
 	clock_t t;
-	bntseq_t *bns, *ntbns = 0;
+	bntseq_t *bns;
 	FILE *fp_sa;
 	gap_opt_t opt;
 
@@ -599,8 +569,6 @@ void bwa_sai2sam_se_core(const char *prefix, const char *fn_sa, const char *fn_f
 
 	m_aln = 0;
 	fread(&opt, sizeof(gap_opt_t), 1, fp_sa);
-	if (!(opt.mode & BWA_MODE_COMPREAD)) // in color space; initialize ntpac
-		ntbns = bwa_open_nt(prefix);
 	bwa_print_sam_SQ(bns);
 	//bwa_print_sam_PG();
 	// set ks
@@ -628,7 +596,7 @@ void bwa_sai2sam_se_core(const char *prefix, const char *fn_sa, const char *fn_f
 		fprintf(stderr, "%.2f sec\n", (float)(clock() - t) / CLOCKS_PER_SEC); t = clock();
 
 		fprintf(stderr, "[bwa_aln_core] refine gapped alignments... ");
-		bwa_refine_gapped(bns, n_seqs, seqs, 0, ntbns);
+		bwa_refine_gapped(bns, n_seqs, seqs, 0);
 		fprintf(stderr, "%.2f sec\n", (float)(clock() - t) / CLOCKS_PER_SEC); t = clock();
 
 		fprintf(stderr, "[bwa_aln_core] print alignments... ");
@@ -642,7 +610,6 @@ void bwa_sai2sam_se_core(const char *prefix, const char *fn_sa, const char *fn_f
 
 	// destroy
 	bwa_seq_close(ks);
-	if (ntbns) bns_destroy(ntbns);
 	bns_destroy(bns);
 	fclose(fp_sa);
 	free(aln);

bwase.h

@@ -14,7 +14,7 @@ extern "C" {
 	// Calculate the approximate position of the sequence from the specified bwt with loaded suffix array.
 	void bwa_cal_pac_pos_core(const bntseq_t *bns, const bwt_t* bwt, bwa_seq_t* seq, const int max_mm, const float fnr);
 	// Refine the approximate position of the sequence to an actual placement for the sequence.
-	void bwa_refine_gapped(const bntseq_t *bns, int n_seqs, bwa_seq_t *seqs, ubyte_t *_pacseq, bntseq_t *ntbns);
+	void bwa_refine_gapped(const bntseq_t *bns, int n_seqs, bwa_seq_t *seqs, ubyte_t *_pacseq);
 	// Backfill certain alignment properties mainly centering around number of matches.
 	void bwa_aln2seq(int n_aln, const bwt_aln1_t *aln, bwa_seq_t *s);
 	// Calculate the end position of a read given a certain sequence.

bwtaln.c

@@ -252,7 +252,7 @@ int bwa_aln(int argc, char *argv[])
 	char *prefix;
 
 	opt = gap_init_opt();
-	while ((c = getopt(argc, argv, "n:o:e:i:d:l:k:cLR:m:t:NM:O:E:q:f:b012IYB:")) >= 0) {
+	while ((c = getopt(argc, argv, "n:o:e:i:d:l:k:LR:m:t:NM:O:E:q:f:b012IYB:")) >= 0) {
 		switch (c) {
 		case 'n':
 			if (strstr(optarg, ".")) opt->fnr = atof(optarg), opt->max_diff = -1;
@@ -272,7 +272,6 @@ int bwa_aln(int argc, char *argv[])
 		case 'L': opt->mode |= BWA_MODE_LOGGAP; break;
 		case 'R': opt->max_top2 = atoi(optarg); break;
 		case 'q': opt->trim_qual = atoi(optarg); break;
-		case 'c': opt->mode &= ~BWA_MODE_COMPREAD; break;
 		case 'N': opt->mode |= BWA_MODE_NONSTOP; opt->max_top2 = 0x7fffffff; break;
 		case 'f': xreopen(optarg, "wb", stdout); break;
 		case 'b': opt->mode |= BWA_MODE_BAM; break;
@@ -310,7 +309,6 @@ int bwa_aln(int argc, char *argv[])
 		fprintf(stderr, "         -q INT    quality threshold for read trimming down to %dbp [%d]\n", BWA_MIN_RDLEN, opt->trim_qual);
         fprintf(stderr, "         -f FILE   file to write output to instead of stdout\n");
 		fprintf(stderr, "         -B INT    length of barcode\n");
-//		fprintf(stderr, "         -c        input sequences are in the color space\n");
 		fprintf(stderr, "         -L        log-scaled gap penalty for long deletions\n");
 		fprintf(stderr, "         -N        non-iterative mode: search for all n-difference hits (slooow)\n");
 		fprintf(stderr, "         -I        the input is in the Illumina 1.3+ FASTQ-like format\n");

bwtaln.h

@@ -107,7 +107,6 @@ typedef struct {
 } gap_opt_t;
 
 #define BWA_PET_STD   1
-#define BWA_PET_SOLID 2
 
 typedef struct {
 	int max_isize, force_isize;

bwtindex.c

@@ -42,11 +42,11 @@ void bwa_pac_rev_core(const char *fn, const char *fn_rev);
 int bwa_index(int argc, char *argv[])
 {
 	char *prefix = 0, *str, *str2, *str3;
-	int c, algo_type = 0, is_color = 0, is_64 = 0;
+	int c, algo_type = 0, is_64 = 0;
 	clock_t t;
 	int64_t l_pac;
 
-	while ((c = getopt(argc, argv, "6ca:p:")) >= 0) {
+	while ((c = getopt(argc, argv, "6a:p:")) >= 0) {
 		switch (c) {
 		case 'a': // if -a is not set, algo_type will be determined later
 			if (strcmp(optarg, "div") == 0) algo_type = 1;
@@ -55,7 +55,6 @@ int bwa_index(int argc, char *argv[])
 			else err_fatal(__func__, "unknown algorithm: '%s'.", optarg);
 			break;
 		case 'p': prefix = strdup(optarg); break;
-		case 'c': is_color = 1; break;
 		case '6': is_64 = 1; break;
 		default: return 1;
 		}
@@ -67,7 +66,6 @@ int bwa_index(int argc, char *argv[])
 		fprintf(stderr, "Options: -a STR    BWT construction algorithm: bwtsw or is [auto]\n");
 		fprintf(stderr, "         -p STR    prefix of the index [same as fasta name]\n");
 		fprintf(stderr, "         -6        index files named as <in.fasta>.64.* instead of <in.fasta>.* \n");
-//		fprintf(stderr, "         -c        build color-space index\n");
 		fprintf(stderr, "\n");
 		fprintf(stderr,	"Warning: `-a bwtsw' does not work for short genomes, while `-a is' and\n");
 		fprintf(stderr, "         `-a div' do not work not for long genomes. Please choose `-a'\n");
@@ -83,29 +81,13 @@ int bwa_index(int argc, char *argv[])
 	str2 = (char*)calloc(strlen(prefix) + 10, 1);
 	str3 = (char*)calloc(strlen(prefix) + 10, 1);
 
-	if (is_color == 0) { // nucleotide indexing
+	{ // nucleotide indexing
 		gzFile fp = xzopen(argv[optind], "r");
 		t = clock();
 		fprintf(stderr, "[bwa_index] Pack FASTA... ");
 		l_pac = bns_fasta2bntseq(fp, prefix, 0);
 		fprintf(stderr, "%.2f sec\n", (float)(clock() - t) / CLOCKS_PER_SEC);
 		gzclose(fp);
-	} else { // color indexing
-		gzFile fp = xzopen(argv[optind], "r");
-		strcat(strcpy(str, prefix), ".nt");
-		t = clock();
-		fprintf(stderr, "[bwa_index] Pack nucleotide FASTA... ");
-		l_pac = bns_fasta2bntseq(fp, str, 0);
-		fprintf(stderr, "%.2f sec\n", (float)(clock() - t) / CLOCKS_PER_SEC);
-		gzclose(fp);
-		{
-			char *tmp_argv[3];
-			tmp_argv[0] = argv[0]; tmp_argv[1] = str; tmp_argv[2] = prefix;
-			t = clock();
-			fprintf(stderr, "[bwa_index] Convert nucleotide PAC to color PAC... ");
-			bwa_pac2cspac(3, tmp_argv);
-			fprintf(stderr, "%.2f sec\n", (float)(clock() - t) / CLOCKS_PER_SEC);
-		}
 	}
 	if (algo_type == 0) algo_type = l_pac > 50000000? 2 : 3; // set the algorithm for generating BWT
 	{