int main_shm(int argc, char *argv[])
{
int c, to_list = 0, to_drop = 0, ret = 0;
char *tmpfn = 0;
while ((c = getopt(argc, argv, "ldf:")) >= 0) {
if (c == 'l') to_list = 1;
else if (c == 'd') to_drop = 1;
else if (c == 'f') tmpfn = optarg;
}
if (optind == argc && !to_list && !to_drop) {
fprintf(stderr, "\nUsage: bwa shm [-d|-l] [-f tmpFile] [idxbase]\n\n");
fprintf(stderr, "Options: -d destroy all indices in shared memory\n");
fprintf(stderr, " -l list names of indices in shared memory\n");
fprintf(stderr, " -f FILE temporary file to reduce peak memory\n\n");
return 1;
}
if (optind < argc && (to_list || to_drop)) {
fprintf(stderr, "[E::%s] open -l or -d cannot be used when 'idxbase' is present\n", __func__);
return 1;
}
if (optind < argc) {
if (bwa_shm_test(argv[optind]) == 0) {
bwaidx_t *idx;
idx = bwa_idx_load_from_disk(argv[optind], BWA_IDX_ALL);
if (bwa_shm_stage(idx, argv[optind], tmpfn) < 0) {
fprintf(stderr, "[E::%s] failed to stage the index in shared memory\n", __func__);
ret = 1;
}
bwa_idx_destroy(idx);
} else fprintf(stderr, "[M::%s] index '%s' is already in shared memory\n", __func__, argv[optind]);
}
if (to_list) bwa_shm_list();
if (to_drop) bwa_shm_destroy();
return ret;
}
int main(int argc, char *argv[])
{
bwaidx_t *idx;
gzFile fp;
kseq_t *ks;
mem_opt_t *opt;
if (argc < 3) {
fprintf(stderr, "Usage: bwamem-lite <idx.base> <reads.fq>\n");
return 1;
}
idx = bwa_idx_load(argv[1], BWA_IDX_ALL); // load the BWA index
if (NULL == idx) {
fprintf(stderr, "Index load failed.\n");
exit(EXIT_FAILURE);
}
fp = strcmp(argv[2], "-")? gzopen(argv[2], "r") : gzdopen(fileno(stdin), "r");
if (NULL == fp) {
fprintf(stderr, "Couldn't open %s : %s\n",
strcmp(argv[2], "-") ? argv[2] : "stdin",
errno ? strerror(errno) : "Out of memory");
exit(EXIT_FAILURE);
}
ks = kseq_init(fp); // initialize the FASTA/Q parser
opt = mem_opt_init(); // initialize the BWA-MEM parameters to the default values
while (kseq_read(ks) >= 0) { // read one sequence
mem_alnreg_v ar;
int i, k;
ar = mem_align1(opt, idx->bwt, idx->bns, idx->pac, ks->seq.l, ks->seq.s); // get all the hits
for (i = 0; i < ar.n; ++i) { // traverse each hit
mem_aln_t a;
if (ar.a[i].secondary >= 0) continue; // skip secondary alignments
a = mem_reg2aln(opt, idx->bns, idx->pac, ks->seq.l, ks->seq.s, &ar.a[i]); // get forward-strand position and CIGAR
// print alignment
err_printf("%s\t%c\t%s\t%ld\t%d\t", ks->name.s, "+-"[a.is_rev], idx->bns->anns[a.rid].name, (long)a.pos, a.mapq);
for (k = 0; k < a.n_cigar; ++k) // print CIGAR
err_printf("%d%c", a.cigar[k]>>4, "MIDSH"[a.cigar[k]&0xf]);
err_printf("\t%d\n", a.NM); // print edit distance
free(a.cigar); // don't forget to deallocate CIGAR
}
free(ar.a); // and deallocate the hit list
}
free(opt);
kseq_destroy(ks);
err_gzclose(fp);
bwa_idx_destroy(idx);
return 0;
}
int bwa_bwtsw2(int argc, char *argv[])
{
bsw2opt_t *opt;
bwaidx_t *idx;
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 ((idx = bwa_idx_load(argv[optind], BWA_IDX_BWT|BWA_IDX_BNS)) == 0) return 0;
bsw2_aln(opt, idx->bns, idx->bwt, argv[optind+1], optind+2 < argc? argv[optind+2] : 0);
bwa_idx_destroy(idx);
free(opt);
return 0;
}
