extern "C" void bwa_seed2genome_destroy()
{
if (bwt_bwt[0])
bwt_destroy(bwt_bwt[0]);
if (bwt_bwt[1])
bwt_destroy(bwt_bwt[1]);
if (bwt_bns)
bns_destroy(bwt_bns) ;
}
int64_t bns_fasta2bntseq(gzFile fp_fa, const char *prefix, int for_only)
{
extern void seq_reverse(int len, ubyte_t *seq, int is_comp); // in bwaseqio.c
kseq_t *seq;
char name[1024];
bntseq_t *bns;
uint8_t *pac = 0;
int32_t m_seqs, m_holes;
int64_t ret = -1, m_pac, l;
bntamb1_t *q;
FILE *fp;
// initialization
seq = kseq_init(fp_fa);
bns = (bntseq_t*)calloc(1, sizeof(bntseq_t));
bns->seed = 11; // fixed seed for random generator
srand48(bns->seed);
m_seqs = m_holes = 8; m_pac = 0x10000;
bns->anns = (bntann1_t*)calloc(m_seqs, sizeof(bntann1_t));
bns->ambs = (bntamb1_t*)calloc(m_holes, sizeof(bntamb1_t));
pac = calloc(m_pac/4, 1);
q = bns->ambs;
strcpy(name, prefix); strcat(name, ".pac");
fp = xopen(name, "wb");
// read sequences
while (kseq_read(seq) >= 0) pac = add1(seq, bns, pac, &m_pac, &m_seqs, &m_holes, &q);
if (!for_only) { // add the reverse complemented sequence
m_pac = (bns->l_pac * 2 + 3) / 4 * 4;
pac = realloc(pac, m_pac/4);
memset(pac + (bns->l_pac+3)/4, 0, (m_pac - (bns->l_pac+3)/4*4) / 4);
for (l = bns->l_pac - 1; l >= 0; --l, ++bns->l_pac)
_set_pac(pac, bns->l_pac, 3-_get_pac(pac, l));
}
ret = bns->l_pac;
{ // finalize .pac file
ubyte_t ct;
err_fwrite(pac, 1, (bns->l_pac>>2) + ((bns->l_pac&3) == 0? 0 : 1), fp);
// the following codes make the pac file size always (l_pac/4+1+1)
if (bns->l_pac % 4 == 0) {
ct = 0;
err_fwrite(&ct, 1, 1, fp);
}
ct = bns->l_pac % 4;
err_fwrite(&ct, 1, 1, fp);
// close .pac file
err_fflush(fp);
err_fclose(fp);
}
bns_dump(bns, prefix);
bns_destroy(bns);
kseq_destroy(seq);
free(pac);
return ret;
}
int64_t dump_forward_pac(gzFile fp_fa, const char *prefix)
{
extern void seq_reverse(int len, ubyte_t *seq, int is_comp); // in bwaseqio.c
kseq_t *seq;
char name[1024];
bntseq_t *bns;
uint8_t *pac = 0;
int32_t m_seqs, m_holes;
int64_t ret = -1, m_pac;
bntamb1_t *q;
FILE *fp;
// initialization
seq = kseq_init(fp_fa);
bns = (bntseq_t*)calloc(1, sizeof(bntseq_t));
bns->seed = 11; // fixed seed for random generator
srand48(bns->seed);
m_seqs = m_holes = 8; m_pac = 0x10000;
bns->anns = (bntann1_t*)calloc(m_seqs, sizeof(bntann1_t));
bns->ambs = (bntamb1_t*)calloc(m_holes, sizeof(bntamb1_t));
pac = calloc(m_pac/4, 1);
q = bns->ambs;
strcpy(name, prefix); strcat(name, ".bis.pac");
fp = xopen(name, "wb");
// read sequences
while (kseq_read(seq) >= 0) pac = add1(seq, bns, pac, &m_pac, &m_seqs, &m_holes, &q);
ret = bns->l_pac;
{ // finalize .pac file
ubyte_t ct;
err_fwrite(pac, 1, (bns->l_pac>>2) + ((bns->l_pac&3) == 0? 0 : 1), fp);
// the following codes make the pac file size always (l_pac/4+1+1)
if (bns->l_pac % 4 == 0) {
ct = 0;
err_fwrite(&ct, 1, 1, fp);
}
ct = bns->l_pac % 4;
err_fwrite(&ct, 1, 1, fp);
// close .pac file
err_fflush(fp);
err_fclose(fp);
}
/* re-dump forward bis bns, otherwise the .bis.ann and .bis.amb have twice as long pac */
/* strcpy(name, prefix); strcat(name, ".bis"); */
/* bis_bns_dump(bns, prefix); */
bns_destroy(bns);
kseq_destroy(seq);
free(pac);
return ret;
}
int main(int argc, char *argv[])
{
bntseq_t *bns;
bns = bns_restore(argv[1]);
uint8_t *pac;
pac = calloc(bns->l_pac/2+2, 1);
fread(pac, 1, bns->l_pac/2+2, bns->fp_pac);
int i;
for(i = 0; i < bns->l_pac; ++i){
putchar( "ACGT#"[_get_pac(pac, i)]);
}
bns_destroy(bns);
}
BWA::~BWA() {
delete[] reference;
bns_destroy(bns);
bwt_destroy(bwts[0]);
bwt_destroy(bwts[1]);
}
void bns_fasta2bntseq(gzFile fp_fa, const char *prefix)
{
kseq_t *seq;
char name[1024];
bntseq_t *bns;
bntamb1_t *q;
int l_buf;
unsigned char buf[0x10000];
int32_t m_seqs, m_holes, l, i;
FILE *fp;
// initialization
seq = kseq_init(fp_fa);
bns = (bntseq_t*)calloc(1, sizeof(bntseq_t));
bns->seed = 11; // fixed seed for random generator
srand48(bns->seed);
m_seqs = m_holes = 8;
bns->anns = (bntann1_t*)calloc(m_seqs, sizeof(bntann1_t));
bns->ambs = (bntamb1_t*)calloc(m_holes, sizeof(bntamb1_t));
q = bns->ambs;
l_buf = 0;
strcpy(name, prefix); strcat(name, ".pac");
fp = xopen(name, "wb");
memset(buf, 0, 0x10000);
// read sequences
while ((l = kseq_read(seq)) >= 0) {
bntann1_t *p;
int lasts;
if (bns->n_seqs == m_seqs) {
m_seqs <<= 1;
bns->anns = (bntann1_t*)realloc(bns->anns, m_seqs * sizeof(bntann1_t));
}
p = bns->anns + bns->n_seqs;
p->name = strdup((char*)seq->name.s);
p->anno = seq->comment.s? strdup((char*)seq->comment.s) : strdup("(null)");
p->gi = 0; p->len = l;
p->offset = (bns->n_seqs == 0)? 0 : (p-1)->offset + (p-1)->len;
p->n_ambs = 0;
for (i = 0, lasts = 0; i < l; ++i) {
int c = nst_nt4_table[(int)seq->seq.s[i]];
if (c >= 4) { // N
if (lasts == seq->seq.s[i]) { // contiguous N
++q->len;
} else {
if (bns->n_holes == m_holes) {
m_holes <<= 1;
bns->ambs = (bntamb1_t*)realloc(bns->ambs, m_holes * sizeof(bntamb1_t));
}
q = bns->ambs + bns->n_holes;
q->len = 1;
q->offset = p->offset + i;
q->amb = seq->seq.s[i];
++p->n_ambs;
++bns->n_holes;
}
}
lasts = seq->seq.s[i];
{ // fill buffer
if (c >= 4) c = lrand48()&0x3;
if (l_buf == 0x40000) {
fwrite(buf, 1, 0x10000, fp);
memset(buf, 0, 0x10000);
l_buf = 0;
}
buf[l_buf>>2] |= c << ((3 - (l_buf&3)) << 1);
++l_buf;
}
}
++bns->n_seqs;
bns->l_pac += seq->seq.l;
}
xassert(bns->l_pac, "zero length sequence.");
{ // finalize .pac file
ubyte_t ct;
fwrite(buf, 1, (l_buf>>2) + ((l_buf&3) == 0? 0 : 1), fp);
// the following codes make the pac file size always (l_pac/4+1+1)
if (bns->l_pac % 4 == 0) {
ct = 0;
fwrite(&ct, 1, 1, fp);
}
ct = bns->l_pac % 4;
fwrite(&ct, 1, 1, fp);
// close .pac file
fclose(fp);
}
bns_dump(bns, prefix);
bns_destroy(bns);
kseq_destroy(seq);
}
int bwa_bwtsw2(int argc, char *argv[])
{
bsw2opt_t *opt;
bwt_t *target[2];
char buf[1024];
bntseq_t *bns;
int c;
opt = bsw2_init_opt();
srand48(11);
optind = 1;
while ((c = getopt(argc, argv, "q:r:a:b:t:T:w:d:z:m:y:s:c:N:Hf:")) >= 0) {
switch (c) {
case 'q': opt->q = atoi(optarg); break;
case 'r': opt->r = atoi(optarg); break;
case 'a': opt->a = atoi(optarg); break;
case 'b': opt->b = atoi(optarg); break;
case 'w': opt->bw = atoi(optarg); break;
case 'T': opt->t = atoi(optarg); break;
case 't': opt->n_threads = atoi(optarg); break;
case 'z': opt->z = atoi(optarg); break;
case 'y': opt->yita = atof(optarg); break;
case 's': opt->is = atoi(optarg); break;
case 'm': opt->mask_level = atof(optarg); break;
case 'c': opt->coef = atof(optarg); break;
case 'N': opt->t_seeds = atoi(optarg); break;
case 'H': opt->hard_clip = 1; break;
case 'f': xreopen(optarg, "w", stdout); break;
}
}
opt->qr = opt->q + opt->r;
if (optind + 2 > argc) {
fprintf(stderr, "\n");
fprintf(stderr, "Usage: bwa bwasw [options] <target.prefix> <query.fa>\n\n");
fprintf(stderr, "Options: -a INT score for a match [%d]\n", opt->a);
fprintf(stderr, " -b INT mismatch penalty [%d]\n", opt->b);
fprintf(stderr, " -q INT gap open penalty [%d]\n", opt->q);
fprintf(stderr, " -r INT gap extension penalty [%d]\n", opt->r);
// fprintf(stderr, " -y FLOAT error recurrence coef. (4..16) [%.1f]\n", opt->yita);
fprintf(stderr, "\n");
fprintf(stderr, " -t INT number of threads [%d]\n", opt->n_threads);
fprintf(stderr, " -s INT size of a chunk of reads [%d]\n", opt->chunk_size);
fprintf(stderr, "\n");
fprintf(stderr, " -w INT band width [%d]\n", opt->bw);
fprintf(stderr, " -m FLOAT mask level [%.2f]\n", opt->mask_level);
fprintf(stderr, "\n");
fprintf(stderr, " -T INT score threshold divided by a [%d]\n", opt->t);
fprintf(stderr, " -s INT maximum seeding interval size [%d]\n", opt->is);
fprintf(stderr, " -z INT Z-best [%d]\n", opt->z);
fprintf(stderr, " -N INT # seeds to trigger reverse alignment [%d]\n", opt->t_seeds);
fprintf(stderr, " -c FLOAT coefficient of length-threshold adjustment [%.1f]\n", opt->coef);
fprintf(stderr, " -H in SAM output, use hard clipping rather than soft\n");
fprintf(stderr, " -f FILE file to output results to instead of stdout\n\n");
fprintf(stderr, "Note: For long Illumina, 454 and Sanger reads, assembly contigs, fosmids and\n");
fprintf(stderr, " BACs, the default setting usually works well. For the current PacBio\n");
fprintf(stderr, " reads (end of 2010), '-b5 -q2 -r1 -z10' is recommended. One may also\n");
fprintf(stderr, " increase '-z' for better sensitivity.\n");
fprintf(stderr, "\n");
if (0) {
double c, theta, eps, delta;
c = opt->a / log(opt->yita);
theta = exp(-opt->b / c) / opt->yita;
eps = exp(-opt->q / c);
delta = exp(-opt->r / c);
fprintf(stderr, "mismatch: %lf, gap_open: %lf, gap_ext: %lf\n\n",
theta, eps, delta);
}
return 1;
}
// adjust opt for opt->a
opt->t *= opt->a;
opt->coef *= opt->a;
strcpy(buf, argv[optind]); target[0] = bwt_restore_bwt(strcat(buf, ".bwt"));
strcpy(buf, argv[optind]); bwt_restore_sa(strcat(buf, ".sa"), target[0]);
strcpy(buf, argv[optind]); target[1] = bwt_restore_bwt(strcat(buf, ".rbwt"));
strcpy(buf, argv[optind]); bwt_restore_sa(strcat(buf, ".rsa"), target[1]);
bns = bns_restore(argv[optind]);
bsw2_aln(opt, bns, target, argv[optind+1]);
bns_destroy(bns);
bwt_destroy(target[0]); bwt_destroy(target[1]);
free(opt);
fflush(stdout);
xreopen("/dev/tty","w",stdout);
return 0;
}
int64_t bis_bns_fasta2bntseq(gzFile fp_fa, const char *prefix, uint8_t parent) {
extern void seq_reverse(int len, ubyte_t *seq, int is_comp); // in bwaseqio.c
kseq_t *seq;
char name[1024];
bntseq_t *bns;
uint8_t *pac = 0, *_pac = 0;
int32_t m_seqs, m_holes;
int64_t ret = -1, m_pac;
bntamb1_t *q;
FILE *fp;
// initialization
gzseek(fp_fa, 0, SEEK_SET);
seq = kseq_init(fp_fa);
bns = (bntseq_t*)calloc(1, sizeof(bntseq_t));
bns->seed = 11; // fixed seed for random generator
srand48(bns->seed);
m_seqs = m_holes = 8; m_pac = 0x10000;
bns->anns = (bntann1_t*)calloc(m_seqs, sizeof(bntann1_t));
bns->ambs = (bntamb1_t*)calloc(m_holes, sizeof(bntamb1_t));
_pac = calloc(m_pac/4, 1);
q = bns->ambs;
if (parent) {
strcpy(name, prefix); strcat(name, ".par.pac");
} else {
strcpy(name, prefix); strcat(name, ".dau.pac");
}
fp = xopen(name, "wb");
// read sequences
while (kseq_read(seq) >= 0) {
_pac = bis_add1(seq, bns, _pac, &m_pac, &m_seqs, &m_holes, &q);
}
/* kseq_rewind(seq); */
/* gzseek(seq->f->f, 0, SEEK_SET); */
/* fprintf(stderr, "foward end\n"); */
/* fflush(stderr); */
/* while (kseq_read(seq) >= 0) { */
/* if (parent) nt256char_rev_ip(seq->seq.s, seq->seq.l); */
/* pac = bis_add1(seq, bns, pac, &m_pac, &m_seqs, &m_holes, &q, parent, 1); */
/* } */
int64_t l,k;
m_pac = (bns->l_pac*2+3)/4*4; /* in bit */
pac = calloc(m_pac/4,sizeof(uint8_t));
for (l=0; l<bns->l_pac; ++l) {
uint8_t c = _get_pac(_pac,l);
if (parent && c == 1) c = 3;
if (!parent && c == 2) c = 0;
_set_pac(pac, l, c);
}
for (k=bns->l_pac-1; k>=0; --k,++l) {
uint8_t c = 3-_get_pac(_pac,k);
if (parent && c == 1) c = 3;
if (!parent && c == 2) c = 0;
_set_pac(pac, l, c);
}
free(_pac);
/* int64_t l; */
/* fprintf(stderr, "reverse end\n"); */
/* fflush(stderr); */
/* if (!for_only) { // add the reverse complemented sequence */
/* m_pac = (bns->l_pac * 2 + 3) / 4 * 4; */
/* pac = realloc(pac, m_pac/4); */
/* memset(pac + (bns->l_pac+3)/4, 0, (m_pac - (bns->l_pac+3)/4*4) / 4); */
/* for (l = bns->l_pac - 1; l >= 0; --l, ++bns->l_pac) */
/* _set_pac(pac, bns->l_pac, 3-_get_pac(pac, l)); */
/* } */
assert(bns->l_pac<<1 == l);
{ // finalize .pac file
ubyte_t ct;
err_fwrite(pac, 1, (l>>2) + ((l&3) == 0? 0 : 1), fp);
// the following codes make the pac file size always (l_pac/4+1+1)
if (l % 4 == 0) {
ct = 0;
err_fwrite(&ct, 1, 1, fp);
}
ct = l % 4;
err_fwrite(&ct, 1, 1, fp);
// close .pac file
err_fflush(fp);
err_fclose(fp);
}
if (parent) bis_bns_dump(bns, prefix);
bns_destroy(bns);
kseq_destroy(seq);
free(pac);
return l;
}
int bwa_bwtsw2(int argc, char *argv[])
{
extern char *bwa_infer_prefix(const char *hint);
bsw2opt_t *opt;
bwt_t *target;
char buf[1024], *prefix;
bntseq_t *bns;
int c;
opt = bsw2_init_opt();
srand48(11);
while ((c = getopt(argc, argv, "q:r:a:b:t:T:w:d:z:m:s:c:N:Hf:MI:SG:C")) >= 0) {
switch (c) {
case 'q': opt->q = atoi(optarg); break;
case 'r': opt->r = atoi(optarg); break;
case 'a': opt->a = atoi(optarg); break;
case 'b': opt->b = atoi(optarg); break;
case 'w': opt->bw = atoi(optarg); break;
case 'T': opt->t = atoi(optarg); break;
case 't': opt->n_threads = atoi(optarg); break;
case 'z': opt->z = atoi(optarg); break;
case 's': opt->is = atoi(optarg); break;
case 'm': opt->mask_level = atof(optarg); break;
case 'c': opt->coef = atof(optarg); break;
case 'N': opt->t_seeds = atoi(optarg); break;
case 'M': opt->multi_2nd = 1; break;
case 'H': opt->hard_clip = 1; break;
case 'f': xreopen(optarg, "w", stdout); break;
case 'I': opt->max_ins = atoi(optarg); break;
case 'S': opt->skip_sw = 1; break;
case 'C': opt->cpy_cmt = 1; break;
case 'G': opt->max_chain_gap = atoi(optarg); break;
}
}
opt->qr = opt->q + opt->r;
if (optind + 2 > argc) {
fprintf(stderr, "\n");
fprintf(stderr, "Usage: bwa bwasw [options] <target.prefix> <query.fa> [query2.fa]\n\n");
fprintf(stderr, "Options: -a INT score for a match [%d]\n", opt->a);
fprintf(stderr, " -b INT mismatch penalty [%d]\n", opt->b);
fprintf(stderr, " -q INT gap open penalty [%d]\n", opt->q);
fprintf(stderr, " -r INT gap extension penalty [%d]\n", opt->r);
fprintf(stderr, " -w INT band width [%d]\n", opt->bw);
fprintf(stderr, " -m FLOAT mask level [%.2f]\n", opt->mask_level);
fprintf(stderr, "\n");
fprintf(stderr, " -t INT number of threads [%d]\n", opt->n_threads);
fprintf(stderr, " -f FILE file to output results to instead of stdout\n");
fprintf(stderr, " -H in SAM output, use hard clipping instead of soft clipping\n");
fprintf(stderr, " -C copy FASTA/Q comment to SAM output\n");
fprintf(stderr, " -M mark multi-part alignments as secondary\n");
fprintf(stderr, " -S skip Smith-Waterman read pairing\n");
fprintf(stderr, " -I INT ignore pairs with insert >=INT for inferring the size distr [%d]\n", opt->max_ins);
fprintf(stderr, "\n");
fprintf(stderr, " -T INT score threshold divided by a [%d]\n", opt->t);
fprintf(stderr, " -c FLOAT coefficient of length-threshold adjustment [%.1f]\n", opt->coef);
fprintf(stderr, " -z INT Z-best [%d]\n", opt->z);
fprintf(stderr, " -s INT maximum seeding interval size [%d]\n", opt->is);
fprintf(stderr, " -N INT # seeds to trigger rev aln; 2*INT is also the chaining threshold [%d]\n", opt->t_seeds);
fprintf(stderr, " -G INT maximum gap size during chaining [%d]\n", opt->max_chain_gap);
fprintf(stderr, "\n");
fprintf(stderr, "Note: For long Illumina, 454 and Sanger reads, assembly contigs, fosmids and\n");
fprintf(stderr, " BACs, the default setting usually works well. For the current PacBio\n");
fprintf(stderr, " reads (end of 2010), '-b5 -q2 -r1 -z10' is recommended. One may also\n");
fprintf(stderr, " increase '-z' for better sensitivity.\n");
fprintf(stderr, "\n");
return 1;
}
// adjust opt for opt->a
opt->t *= opt->a;
opt->coef *= opt->a;
if ((prefix = bwa_infer_prefix(argv[optind])) == 0) {
fprintf(stderr, "[%s] fail to locate the index\n", __func__);
return 0;
}
strcpy(buf, prefix); target = bwt_restore_bwt(strcat(buf, ".bwt"));
strcpy(buf, prefix); bwt_restore_sa(strcat(buf, ".sa"), target);
bns = bns_restore(prefix);
bsw2_aln(opt, bns, target, argv[optind+1], optind+2 < argc? argv[optind+2] : 0);
bns_destroy(bns);
bwt_destroy(target);
free(opt); free(prefix);
return 0;
}
int bwa_bwtsw2(int argc, char *argv[])
{
bsw2opt_t *opt;
bwt_t *target[2];
char buf[1024];
bntseq_t *bns;
int c;
opt = bsw2_init_opt();
srand48(11);
while ((c = getopt(argc, argv, "q:r:a:b:t:T:w:d:z:m:y:s:c:N:H")) >= 0) {
switch (c) {
case 'q': opt->q = atoi(optarg); break;
case 'r': opt->r = atoi(optarg); break;
case 'a': opt->a = atoi(optarg); break;
case 'b': opt->b = atoi(optarg); break;
case 'w': opt->bw = atoi(optarg); break;
case 'T': opt->t = atoi(optarg); break;
case 't': opt->n_threads = atoi(optarg); break;
case 'z': opt->z = atoi(optarg); break;
case 'y': opt->yita = atof(optarg); break;
case 's': opt->is = atoi(optarg); break;
case 'm': opt->mask_level = atof(optarg); break;
case 'c': opt->coef = atof(optarg); break;
case 'N': opt->t_seeds = atoi(optarg); break;
case 'H': opt->hard_clip = 1; break;
}
}
opt->qr = opt->q + opt->r;
if (optind + 2 > argc) {
fprintf(stderr, "\n");
fprintf(stderr, "Usage: bwa dbwtsw [options] <target.prefix> <query.fa>\n\n");
fprintf(stderr, "Options: -a INT score for a match [%d]\n", opt->a);
fprintf(stderr, " -b INT mismatch penalty [%d]\n", opt->b);
fprintf(stderr, " -q INT gap open penalty [%d]\n", opt->q);
fprintf(stderr, " -r INT gap extension penalty [%d]\n", opt->r);
// fprintf(stderr, " -y FLOAT error recurrence coef. (4..16) [%.1f]\n", opt->yita);
fprintf(stderr, "\n");
fprintf(stderr, " -t INT nmber of threads [%d]\n", opt->n_threads);
fprintf(stderr, " -s INT size of a chunk of reads [%d]\n", opt->chunk_size);
fprintf(stderr, "\n");
fprintf(stderr, " -w INT band width [%d]\n", opt->bw);
fprintf(stderr, " -m FLOAT mask level [%.2f]\n", opt->mask_level);
fprintf(stderr, "\n");
fprintf(stderr, " -T INT score threshold divided by a [%d]\n", opt->t);
fprintf(stderr, " -s INT maximum seeding interval size [%d]\n", opt->is);
fprintf(stderr, " -z INT Z-best [%d]\n", opt->z);
fprintf(stderr, " -N INT # seeds to trigger reverse alignment [%d]\n", opt->t_seeds);
fprintf(stderr, " -c FLOAT coefficient of length-threshold adjustment [%.1f]\n", opt->coef);
fprintf(stderr, " -H in SAM output, use hard clipping rather than soft\n");
fprintf(stderr, "\n");
{
double c, theta, eps, delta;
c = opt->a / log(opt->yita);
theta = exp(-opt->b / c) / opt->yita;
eps = exp(-opt->q / c);
delta = exp(-opt->r / c);
fprintf(stderr, "mismatch: %lf, gap_open: %lf, gap_ext: %lf\n\n",
theta, eps, delta);
}
return 1;
}
// adjust opt for opt->a
opt->t *= opt->a;
opt->coef *= opt->a;
strcpy(buf, argv[optind]); target[0] = bwt_restore_bwt(strcat(buf, ".bwt"));
strcpy(buf, argv[optind]); bwt_restore_sa(strcat(buf, ".sa"), target[0]);
strcpy(buf, argv[optind]); target[1] = bwt_restore_bwt(strcat(buf, ".rbwt"));
strcpy(buf, argv[optind]); bwt_restore_sa(strcat(buf, ".rsa"), target[1]);
bns = bns_restore(argv[optind]);
bsw2_aln(opt, bns, target, argv[optind+1]);
bns_destroy(bns);
bwt_destroy(target[0]); bwt_destroy(target[1]);
free(opt);
return 0;
}
int64_t R_bns_fasta2bntseq(gzFile fp_fa, const char *prefix)
{
// extern void seq_reverse(int len, ubyte_t *seq, int is_comp); // in bwaseqio.c
kseq_t *seq;
char name[1024];
bntseq_t *bns;
uint8_t *pac = 0;
uint8_t *reverse_pac = NULL;
int32_t m_seqs, m_holes;
int64_t ret = -1, m_pac, l, m_r_pac;
bntamb1_t *q;
FILE *fp, *fp_r;
int i;
// initialization
seq = kseq_init(fp_fa);
bns = (bntseq_t*)calloc(1, sizeof(bntseq_t));
bns->seed = 11; // fixed seed for random generator
srand48(bns->seed);
m_seqs = m_holes = 8; m_pac = 0x10000;
bns->anns = (bntann1_t*)calloc(m_seqs, sizeof(bntann1_t));
bns->ambs = (bntamb1_t*)calloc(m_holes, sizeof(bntamb1_t));
pac = calloc(m_pac/NT_PER_BYTE, 1);
q = bns->ambs;
strcpy(name, prefix); strcat(name, ".pac");
fp = xopen(name, "wb");
memset(name, '\0', 1024);strcpy(name, prefix); strcat(name, ".rpac");
fp_r = xopen(name, "wb");
// read sequences
while (kseq_read(seq) >= 0) pac = R_add1(seq, bns, pac, &m_pac, &m_seqs, &m_holes, &q);
// add the reverse complemented sequence
ret = bns->l_pac;
fprintf(stderr, "[R_bns_fasta2bntseq]:reverse_pac!\n");
m_r_pac = (bns->l_pac +NT_PER_BYTE-1)/NT_PER_BYTE *NT_PER_BYTE;
reverse_pac = calloc(m_r_pac/NT_PER_BYTE, sizeof(uint8_t));
for(l = bns->l_pac-1, i =0; l>=0, i < bns->l_pac; --l, ++i)
{
_set_pac(reverse_pac, i, _get_pac(pac, l));
}
ubyte_t ct;
{ // finalize .pac and file
fwrite(pac, 1, (bns->l_pac>>1) + ((bns->l_pac&1) == 0? 0 : 1), fp);
// the following codes make the pac file size always (l_pac/4+1+1)
if (bns->l_pac % NT_PER_BYTE == 0) {
ct = 0;
fwrite(&ct, 1, 1, fp);
}
ct = bns->l_pac % NT_PER_BYTE;
fwrite(&ct, 1, 1, fp);
// close .pac file
fclose(fp);
}
{ // finalize .rpac and file
fwrite(reverse_pac, 1, (bns->l_pac>>1) + ((bns->l_pac&1) == 0? 0 : 1), fp_r);
// the following codes make the pac file size always (l_pac/4+1+1)
if (bns->l_pac % NT_PER_BYTE == 0) {
ct = 0;
fwrite(&ct, 1, 1, fp_r);
}
ct = bns->l_pac % NT_PER_BYTE;
fwrite(&ct, 1, 1, fp_r);
// close .rpac file
fclose(fp_r);
}
bns_dump(bns, prefix);
bns_destroy(bns);
kseq_destroy(seq);
free(pac);
free(reverse_pac);
return ret;
}