int main(int argc, char **argv) {
dlib::BamHandle in = dlib::BamHandle("bed_test.bam");
dlib::ParsedBed bed = dlib::ParsedBed("bed_test.bed", in.header);
bam1_t *b = bam_init1();
size_t diffs = 0;
void *lh3bed = bed_read("bed_test.bed");
samFile *so = sam_open("disagreed.bam", "wb9");
sam_hdr_write(so, in.header);
size_t disagrees = 0, agrees = 0;
int dbr = 0, lh3r = 0;
while(in.read(b) != -1) {
if(b->core.flag & (BAM_FUNMAP)) continue;
if((dbr = bed.bam1_test(b)) != (lh3r = bed_overlap(lh3bed, in.header->target_name[b->core.tid], b->core.pos, bam_endpos(b)))) {
LOG_EXIT("dbr: %i. lh3r: %i. Contig: %s. Position: %i. endpos; %i\n", dbr, lh3r, in.header->target_name[b->core.tid], b->core.pos, bam_endpos(b));
if(++disagrees % 100 == 0) LOG_DEBUG("disagrees: %lu.\n", disagrees);
sam_write1(so, in.header, b);
} else {
if(++agrees % 500000 == 0) LOG_DEBUG("agrees: %lu.\n", agrees);
}
}
sam_close(so);
bam_destroy1(b);
bed_destroy(lh3bed);
return EXIT_SUCCESS;
}
int bam_mpileup(int argc, char *argv[])
{
int c;
const char *file_list = NULL;
char **fn = NULL;
int nfiles = 0, use_orphan = 0;
mplp_conf_t mplp;
memset(&mplp, 0, sizeof(mplp_conf_t));
#define MPLP_PRINT_POS 0x4000
mplp.max_mq = 60;
mplp.min_baseQ = 13;
mplp.capQ_thres = 0;
mplp.max_depth = 250; mplp.max_indel_depth = 250;
mplp.openQ = 40; mplp.extQ = 20; mplp.tandemQ = 100;
mplp.min_frac = 0.002; mplp.min_support = 1;
mplp.flag = MPLP_NO_ORPHAN | MPLP_REALN;
while ((c = getopt(argc, argv, "Agf:r:l:M:q:Q:uaRC:BDSd:L:b:P:o:e:h:Im:F:EG:6Os")) >= 0) {
switch (c) {
case 'f':
mplp.fai = fai_load(optarg);
if (mplp.fai == 0) return 1;
break;
case 'd': mplp.max_depth = atoi(optarg); break;
case 'r': mplp.reg = strdup(optarg); break;
case 'l': mplp.bed = bed_read(optarg); break;
case 'P': mplp.pl_list = strdup(optarg); break;
case 'g': mplp.flag |= MPLP_GLF; break;
case 'u': mplp.flag |= MPLP_NO_COMP | MPLP_GLF; break;
case 'a': mplp.flag |= MPLP_NO_ORPHAN | MPLP_REALN; break;
case 'B': mplp.flag &= ~MPLP_REALN; break;
case 'D': mplp.flag |= MPLP_FMT_DP; break;
case 'S': mplp.flag |= MPLP_FMT_SP; break;
case 'I': mplp.flag |= MPLP_NO_INDEL; break;
case 'E': mplp.flag |= MPLP_EXT_BAQ; break;
case '6': mplp.flag |= MPLP_ILLUMINA13; break;
case 'R': mplp.flag |= MPLP_IGNORE_RG; break;
case 's': mplp.flag |= MPLP_PRINT_MAPQ; break;
case 'O': mplp.flag |= MPLP_PRINT_POS; break;
case 'C': mplp.capQ_thres = atoi(optarg); break;
case 'M': mplp.max_mq = atoi(optarg); break;
case 'q': mplp.min_mq = atoi(optarg); break;
case 'Q': mplp.min_baseQ = atoi(optarg); break;
case 'b': file_list = optarg; break;
case 'o': mplp.openQ = atoi(optarg); break;
case 'e': mplp.extQ = atoi(optarg); break;
case 'h': mplp.tandemQ = atoi(optarg); break;
case 'A': use_orphan = 1; break;
case 'F': mplp.min_frac = atof(optarg); break;
case 'm': mplp.min_support = atoi(optarg); break;
case 'L': mplp.max_indel_depth = atoi(optarg); break;
case 'G': {
FILE *fp_rg;
char buf[1024];
mplp.rghash = bcf_str2id_init();
if ((fp_rg = fopen(optarg, "r")) == 0)
fprintf(stderr, "(%s) Fail to open file %s. Continue anyway.\n", __func__, optarg);
while (!feof(fp_rg) && fscanf(fp_rg, "%s", buf) > 0) // this is not a good style, but forgive me...
bcf_str2id_add(mplp.rghash, strdup(buf));
fclose(fp_rg);
}
break;
}
}
if (use_orphan) mplp.flag &= ~MPLP_NO_ORPHAN;
if (argc == 1) {
fprintf(stderr, "\n");
fprintf(stderr, "Usage: samtools mpileup [options] in1.bam [in2.bam [...]]\n\n");
fprintf(stderr, "Input options:\n\n");
fprintf(stderr, " -6 assume the quality is in the Illumina-1.3+ encoding\n");
fprintf(stderr, " -A count anomalous read pairs\n");
fprintf(stderr, " -B disable BAQ computation\n");
fprintf(stderr, " -b FILE list of input BAM files [null]\n");
fprintf(stderr, " -C INT parameter for adjusting mapQ; 0 to disable [0]\n");
fprintf(stderr, " -d INT max per-BAM depth to avoid excessive memory usage [%d]\n", mplp.max_depth);
fprintf(stderr, " -E extended BAQ for higher sensitivity but lower specificity\n");
fprintf(stderr, " -f FILE faidx indexed reference sequence file [null]\n");
fprintf(stderr, " -G FILE exclude read groups listed in FILE [null]\n");
fprintf(stderr, " -l FILE list of positions (chr pos) or regions (BED) [null]\n");
fprintf(stderr, " -M INT cap mapping quality at INT [%d]\n", mplp.max_mq);
fprintf(stderr, " -r STR region in which pileup is generated [null]\n");
fprintf(stderr, " -R ignore RG tags\n");
fprintf(stderr, " -q INT skip alignments with mapQ smaller than INT [%d]\n", mplp.min_mq);
fprintf(stderr, " -Q INT skip bases with baseQ/BAQ smaller than INT [%d]\n", mplp.min_baseQ);
fprintf(stderr, "\nOutput options:\n\n");
fprintf(stderr, " -D output per-sample DP in BCF (require -g/-u)\n");
fprintf(stderr, " -g generate BCF output (genotype likelihoods)\n");
fprintf(stderr, " -O output base positions on reads (disabled by -g/-u)\n");
fprintf(stderr, " -s output mapping quality (disabled by -g/-u)\n");
fprintf(stderr, " -S output per-sample strand bias P-value in BCF (require -g/-u)\n");
fprintf(stderr, " -u generate uncompress BCF output\n");
fprintf(stderr, "\nSNP/INDEL genotype likelihoods options (effective with `-g' or `-u'):\n\n");
fprintf(stderr, " -e INT Phred-scaled gap extension seq error probability [%d]\n", mplp.extQ);
fprintf(stderr, " -F FLOAT minimum fraction of gapped reads for candidates [%g]\n", mplp.min_frac);
fprintf(stderr, " -h INT coefficient for homopolymer errors [%d]\n", mplp.tandemQ);
fprintf(stderr, " -I do not perform indel calling\n");
fprintf(stderr, " -L INT max per-sample depth for INDEL calling [%d]\n", mplp.max_indel_depth);
fprintf(stderr, " -m INT minimum gapped reads for indel candidates [%d]\n", mplp.min_support);
fprintf(stderr, " -o INT Phred-scaled gap open sequencing error probability [%d]\n", mplp.openQ);
fprintf(stderr, " -P STR comma separated list of platforms for indels [all]\n");
fprintf(stderr, "\n");
fprintf(stderr, "Notes: Assuming diploid individuals.\n\n");
return 1;
}
if (file_list) {
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
mpileup(&mplp,nfiles,fn);
for (c=0; c<nfiles; c++) free(fn[c]);
free(fn);
} else mpileup(&mplp, argc - optind, argv + optind);
if (mplp.rghash) bcf_str2id_thorough_destroy(mplp.rghash);
free(mplp.reg); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);
if (mplp.bed) bed_destroy(mplp.bed);
return 0;
}
int main(int argc, char *argv[])
{
int c, skip = -1, meta = -1, list_chrms = 0, force = 0, print_header = 0, bed_reg = 0;
ti_conf_t conf = ti_conf_gff;
const char *reheader = NULL;
while ((c = getopt(argc, argv, "p:s:b:e:0S:c:lhfBr:")) >= 0) {
switch (c) {
case 'B': bed_reg = 1; break;
case '0': conf.preset |= TI_FLAG_UCSC; break;
case 'S': skip = atoi(optarg); break;
case 'c': meta = optarg[0]; break;
case 'p':
if (strcmp(optarg, "gff") == 0) conf = ti_conf_gff;
else if (strcmp(optarg, "bed") == 0) conf = ti_conf_bed;
else if (strcmp(optarg, "sam") == 0) conf = ti_conf_sam;
else if (strcmp(optarg, "vcf") == 0 || strcmp(optarg, "vcf4") == 0) conf = ti_conf_vcf;
else if (strcmp(optarg, "psltbl") == 0) conf = ti_conf_psltbl;
else {
fprintf(stderr, "[main] unrecognized preset '%s'\n", optarg);
return 1;
}
break;
case 's': conf.sc = atoi(optarg); break;
case 'b': conf.bc = atoi(optarg); break;
case 'e': conf.ec = atoi(optarg); break;
case 'l': list_chrms = 1; break;
case 'h': print_header = 1; break;
case 'f': force = 1; break;
case 'r': reheader = optarg; break;
}
}
if (skip >= 0) conf.line_skip = skip;
if (meta >= 0) conf.meta_char = meta;
if (optind == argc) {
fprintf(stderr, "\n");
fprintf(stderr, "Program: tabix (TAB-delimited file InderXer)\n");
fprintf(stderr, "Version: %s\n\n", PACKAGE_VERSION);
fprintf(stderr, "Usage: tabix <in.tab.bgz> [region1 [region2 [...]]]\n\n");
fprintf(stderr, "Options: -p STR preset: gff, bed, sam, vcf, psltbl [gff]\n");
fprintf(stderr, " -s INT sequence name column [1]\n");
fprintf(stderr, " -b INT start column [4]\n");
fprintf(stderr, " -e INT end column; can be identical to '-b' [5]\n");
fprintf(stderr, " -S INT skip first INT lines [0]\n");
fprintf(stderr, " -c CHAR symbol for comment/meta lines [#]\n");
fprintf(stderr, " -r FILE replace the header with the content of FILE [null]\n");
fprintf(stderr, " -B region1 is a BED file (entire file will be read)\n");
fprintf(stderr, " -0 zero-based coordinate\n");
fprintf(stderr, " -h print the header lines\n");
fprintf(stderr, " -l list chromosome names\n");
fprintf(stderr, " -f force to overwrite the index\n");
fprintf(stderr, "\n");
return 1;
}
if (list_chrms) {
ti_index_t *idx;
int i, n;
const char **names;
idx = ti_index_load(argv[optind]);
if (idx == 0) {
fprintf(stderr, "[main] fail to load the index file.\n");
return 1;
}
names = ti_seqname(idx, &n);
for (i = 0; i < n; ++i) printf("%s\n", names[i]);
free(names);
ti_index_destroy(idx);
return 0;
}
if (reheader)
return reheader_file(reheader,argv[optind],conf.meta_char);
struct stat stat_tbi,stat_vcf;
char *fnidx = calloc(strlen(argv[optind]) + 5, 1);
strcat(strcpy(fnidx, argv[optind]), ".tbi");
if (optind + 1 == argc) {
if (force == 0) {
if (stat(fnidx, &stat_tbi) == 0)
{
// Before complaining, check if the VCF file isn't newer. This is a common source of errors,
// people tend not to notice that tabix failed
stat(argv[optind], &stat_vcf);
if ( stat_vcf.st_mtime <= stat_tbi.st_mtime )
{
fprintf(stderr, "[tabix] the index file exists. Please use '-f' to overwrite.\n");
free(fnidx);
return 1;
}
}
}
if ( bgzf_check_bgzf(argv[optind])!=1 )
{
fprintf(stderr,"[tabix] was bgzip used to compress this file? %s\n", argv[optind]);
free(fnidx);
return 1;
}
return ti_index_build(argv[optind], &conf);
}
{ // retrieve
tabix_t *t;
// Common source of errors: new VCF is used with an old index
stat(fnidx, &stat_tbi);
stat(argv[optind], &stat_vcf);
if ( force==0 && stat_vcf.st_mtime > stat_tbi.st_mtime )
{
fprintf(stderr, "[tabix] the index file is older than the vcf file. Please use '-f' to overwrite or reindex.\n");
free(fnidx);
return 1;
}
free(fnidx);
if ((t = ti_open(argv[optind], 0)) == 0) {
fprintf(stderr, "[main] fail to open the data file.\n");
return 1;
}
if (strcmp(argv[optind+1], ".") == 0) { // retrieve all
ti_iter_t iter;
const char *s;
int len;
iter = ti_query(t, 0, 0, 0);
while ((s = ti_read(t, iter, &len)) != 0) {
fputs(s, stdout); fputc('\n', stdout);
}
ti_iter_destroy(iter);
} else { // retrieve from specified regions
int i, len;
ti_iter_t iter;
const char *s;
const ti_conf_t *idxconf;
if (ti_lazy_index_load(t) < 0 && bed_reg == 0) {
fprintf(stderr,"[tabix] failed to load the index file.\n");
return 1;
}
idxconf = ti_get_conf(t->idx);
if ( print_header )
{
// If requested, print the header lines here
iter = ti_query(t, 0, 0, 0);
while ((s = ti_read(t, iter, &len)) != 0) {
if ((int)(*s) != idxconf->meta_char) break;
fputs(s, stdout); fputc('\n', stdout);
}
ti_iter_destroy(iter);
}
if (bed_reg) {
extern int bed_overlap(const void *_h, const char *chr, int beg, int end);
extern void *bed_read(const char *fn);
extern void bed_destroy(void *_h);
const ti_conf_t *conf_ = idxconf? idxconf : &conf; // use the index file if available
void *bed = bed_read(argv[optind+1]); // load the BED file
ti_interval_t intv;
if (bed == 0) {
fprintf(stderr, "[main] fail to read the BED file.\n");
return 1;
}
iter = ti_query(t, 0, 0, 0);
while ((s = ti_read(t, iter, &len)) != 0) {
int c;
ti_get_intv(conf_, len, (char*)s, &intv);
c = *intv.se; *intv.se = '\0';
if (bed_overlap(bed, intv.ss, intv.beg, intv.end)) {
*intv.se = c;
puts(s);
}
*intv.se = c;
}
ti_iter_destroy(iter);
bed_destroy(bed);
} else {
for (i = optind + 1; i < argc; ++i) {
int tid, beg, end;
if (ti_parse_region(t->idx, argv[i], &tid, &beg, &end) == 0) {
iter = ti_queryi(t, tid, beg, end);
while ((s = ti_read(t, iter, &len)) != 0) {
fputs(s, stdout); fputc('\n', stdout);
}
ti_iter_destroy(iter);
}
// else fprintf(stderr, "[main] invalid region: unknown target name or minus interval.\n");
}
}
}
ti_close(t);
}
return 0;
}
int main_depth(int argc, char *argv[])
#endif
{
int i, n, tid, beg, end, pos, *n_plp, baseQ = 0, mapQ = 0;
const bam_pileup1_t **plp;
char *reg = 0; // specified region
void *bed = 0; // BED data structure
bam_header_t *h = 0; // BAM header of the 1st input
aux_t **data;
bam_mplp_t mplp;
// parse the command line
while ((n = getopt(argc, argv, "r:b:q:Q:")) >= 0) {
switch (n) {
case 'r': reg = strdup(optarg); break; // parsing a region requires a BAM header
case 'b': bed = bed_read(optarg); break; // BED or position list file can be parsed now
case 'q': baseQ = atoi(optarg); break; // base quality threshold
case 'Q': mapQ = atoi(optarg); break; // mapping quality threshold
}
}
if (optind == argc) {
fprintf(stderr, "Usage: bam2depth [-r reg] [-q baseQthres] [-Q mapQthres] [-b in.bed] <in1.bam> [...]\n");
return 1;
}
// initialize the auxiliary data structures
n = argc - optind; // the number of BAMs on the command line
data = (aux_t **) calloc(n, sizeof(void*)); // data[i] for the i-th input
beg = 0; end = 1<<30; tid = -1; // set the default region
for (i = 0; i < n; ++i) {
bam_header_t *htmp;
data[i] = (aux_t *) calloc(1, sizeof(aux_t));
data[i]->fp = bam_open(argv[optind+i], "r"); // open BAM
data[i]->min_mapQ = mapQ; // set the mapQ filter
htmp = bam_header_read(data[i]->fp); // read the BAM header
if (i == 0) {
h = htmp; // keep the header of the 1st BAM
if (reg) bam_parse_region(h, reg, &tid, &beg, &end); // also parse the region
} else bam_header_destroy(htmp); // if not the 1st BAM, trash the header
if (tid >= 0) { // if a region is specified and parsed successfully
bam_index_t *idx = bam_index_load(argv[optind+i]); // load the index
data[i]->iter = bam_iter_query(idx, tid, beg, end); // set the iterator
bam_index_destroy(idx); // the index is not needed any more; phase out of the memory
}
}
// the core multi-pileup loop
mplp = bam_mplp_init(n, read_bam, (void**)data); // initialization
n_plp = (int*) calloc(n, sizeof(int)); // n_plp[i] is the number of covering reads from the i-th BAM
plp = (bam_pileup1_t **) calloc(n, sizeof(void*)); // plp[i] points to the array of covering reads (internal in mplp)
while (bam_mplp_auto(mplp, &tid, &pos, n_plp, plp) > 0) { // come to the next covered position
if (pos < beg || pos >= end) continue; // out of range; skip
if (bed && bed_overlap(bed, h->target_name[tid], pos, pos + 1) == 0) continue; // not in BED; skip
fputs(h->target_name[tid], stdout); printf("\t%d", pos+1); // a customized printf() would be faster
for (i = 0; i < n; ++i) { // base level filters have to go here
int j, m = 0;
for (j = 0; j < n_plp[i]; ++j) {
const bam_pileup1_t *p = plp[i] + j; // DON'T modfity plp[][] unless you really know
if (p->is_del || p->is_refskip) ++m; // having dels or refskips at tid:pos
else if (bam1_qual(p->b)[p->qpos] < baseQ) ++m; // low base quality
}
printf("\t%d", n_plp[i] - m); // this the depth to output
}
putchar('\n');
}
free(n_plp); free(plp);
bam_mplp_destroy(mplp);
bam_header_destroy(h);
for (i = 0; i < n; ++i) {
bam_close(data[i]->fp);
if (data[i]->iter) bam_iter_destroy(data[i]->iter);
free(data[i]);
}
free(data); free(reg);
if (bed) bed_destroy(bed);
return 0;
}
int main_depth(int argc, char *argv[])
{
int i, n, tid, reg_tid, beg, end, pos, *n_plp, baseQ = 0, mapQ = 0, min_len = 0;
int all = 0, status = EXIT_SUCCESS, nfiles, max_depth = -1;
const bam_pileup1_t **plp;
char *reg = 0; // specified region
void *bed = 0; // BED data structure
char *file_list = NULL, **fn = NULL;
bam_hdr_t *h = NULL; // BAM header of the 1st input
aux_t **data;
bam_mplp_t mplp;
int last_pos = -1, last_tid = -1, ret;
sam_global_args ga = SAM_GLOBAL_ARGS_INIT;
static const struct option lopts[] = {
SAM_OPT_GLOBAL_OPTIONS('-', 0, '-', '-', 0),
{ NULL, 0, NULL, 0 }
};
// parse the command line
while ((n = getopt_long(argc, argv, "r:b:q:Q:l:f:am:d:", lopts, NULL)) >= 0) {
switch (n) {
case 'l':
min_len = atoi(optarg);
break; // minimum query length
case 'r':
reg = strdup(optarg);
break; // parsing a region requires a BAM header
case 'b':
bed = bed_read(optarg); // BED or position list file can be parsed now
if (!bed) {
print_error_errno("depth", "Could not read file \"%s\"", optarg);
return 1;
}
break;
case 'q':
baseQ = atoi(optarg);
break; // base quality threshold
case 'Q':
mapQ = atoi(optarg);
break; // mapping quality threshold
case 'f':
file_list = optarg;
break;
case 'a':
all++;
break;
case 'd':
case 'm':
max_depth = atoi(optarg);
break; // maximum coverage depth
default:
if (parse_sam_global_opt(n, optarg, lopts, &ga) == 0) break;
/* else fall-through */
case '?':
return usage();
}
}
if (optind == argc && !file_list)
return usage();
// initialize the auxiliary data structures
if (file_list)
{
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
n = nfiles;
argv = fn;
optind = 0;
}
else
n = argc - optind; // the number of BAMs on the command line
data = calloc(n, sizeof(aux_t*)); // data[i] for the i-th input
reg_tid = 0;
beg = 0;
end = INT_MAX; // set the default region
for (i = 0; i < n; ++i) {
int rf;
data[i] = calloc(1, sizeof(aux_t));
data[i]->fp = sam_open_format(argv[optind+i], "r", &ga.in); // open BAM
if (data[i]->fp == NULL) {
print_error_errno("depth", "Could not open \"%s\"", argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
rf = SAM_FLAG | SAM_RNAME | SAM_POS | SAM_MAPQ | SAM_CIGAR | SAM_SEQ;
if (baseQ) rf |= SAM_QUAL;
if (hts_set_opt(data[i]->fp, CRAM_OPT_REQUIRED_FIELDS, rf)) {
fprintf(stderr, "Failed to set CRAM_OPT_REQUIRED_FIELDS value\n");
return 1;
}
if (hts_set_opt(data[i]->fp, CRAM_OPT_DECODE_MD, 0)) {
fprintf(stderr, "Failed to set CRAM_OPT_DECODE_MD value\n");
return 1;
}
data[i]->min_mapQ = mapQ; // set the mapQ filter
data[i]->min_len = min_len; // set the qlen filter
data[i]->hdr = sam_hdr_read(data[i]->fp); // read the BAM header
if (data[i]->hdr == NULL) {
fprintf(stderr, "Couldn't read header for \"%s\"\n",
argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
if (reg) { // if a region is specified
hts_idx_t *idx = sam_index_load(data[i]->fp, argv[optind+i]); // load the index
if (idx == NULL) {
print_error("depth", "can't load index for \"%s\"", argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
data[i]->iter = sam_itr_querys(idx, data[i]->hdr, reg); // set the iterator
hts_idx_destroy(idx); // the index is not needed any more; free the memory
if (data[i]->iter == NULL) {
print_error("depth", "can't parse region \"%s\"", reg);
status = EXIT_FAILURE;
goto depth_end;
}
}
}
h = data[0]->hdr; // easy access to the header of the 1st BAM
if (reg) {
beg = data[0]->iter->beg; // and to the parsed region coordinates
end = data[0]->iter->end;
reg_tid = data[0]->iter->tid;
}
// the core multi-pileup loop
mplp = bam_mplp_init(n, read_bam, (void**)data); // initialization
if (0 < max_depth)
bam_mplp_set_maxcnt(mplp,max_depth); // set maximum coverage depth
n_plp = calloc(n, sizeof(int)); // n_plp[i] is the number of covering reads from the i-th BAM
plp = calloc(n, sizeof(bam_pileup1_t*)); // plp[i] points to the array of covering reads (internal in mplp)
while ((ret=bam_mplp_auto(mplp, &tid, &pos, n_plp, plp)) > 0) { // come to the next covered position
if (pos < beg || pos >= end) continue; // out of range; skip
if (tid >= h->n_targets) continue; // diff number of @SQ lines per file?
if (all) {
while (tid > last_tid) {
if (last_tid >= 0 && !reg) {
// Deal with remainder or entirety of last tid.
while (++last_pos < h->target_len[last_tid]) {
// Horribly inefficient, but the bed API is an obfuscated black box.
if (bed && bed_overlap(bed, h->target_name[last_tid], last_pos, last_pos + 1) == 0)
continue;
fputs(h->target_name[last_tid], stdout);
printf("\t%d", last_pos+1);
for (i = 0; i < n; i++)
putchar('\t'), putchar('0');
putchar('\n');
}
}
last_tid++;
last_pos = -1;
if (all < 2)
break;
}
// Deal with missing portion of current tid
while (++last_pos < pos) {
if (last_pos < beg) continue; // out of range; skip
if (bed && bed_overlap(bed, h->target_name[tid], last_pos, last_pos + 1) == 0)
continue;
fputs(h->target_name[tid], stdout);
printf("\t%d", last_pos+1);
for (i = 0; i < n; i++)
putchar('\t'), putchar('0');
putchar('\n');
}
last_tid = tid;
last_pos = pos;
}
if (bed && bed_overlap(bed, h->target_name[tid], pos, pos + 1) == 0) continue;
fputs(h->target_name[tid], stdout);
printf("\t%d", pos+1); // a customized printf() would be faster
for (i = 0; i < n; ++i) { // base level filters have to go here
int j, m = 0;
for (j = 0; j < n_plp[i]; ++j) {
const bam_pileup1_t *p = plp[i] + j; // DON'T modfity plp[][] unless you really know
if (p->is_del || p->is_refskip) ++m; // having dels or refskips at tid:pos
else if (bam_get_qual(p->b)[p->qpos] < baseQ) ++m; // low base quality
}
printf("\t%d", n_plp[i] - m); // this the depth to output
}
putchar('\n');
}
if (ret < 0) status = EXIT_FAILURE;
free(n_plp);
free(plp);
bam_mplp_destroy(mplp);
if (all) {
// Handle terminating region
if (last_tid < 0 && reg && all > 1) {
last_tid = reg_tid;
last_pos = beg-1;
}
while (last_tid >= 0 && last_tid < h->n_targets) {
while (++last_pos < h->target_len[last_tid]) {
if (last_pos >= end) break;
if (bed && bed_overlap(bed, h->target_name[last_tid], last_pos, last_pos + 1) == 0)
continue;
fputs(h->target_name[last_tid], stdout);
printf("\t%d", last_pos+1);
for (i = 0; i < n; i++)
putchar('\t'), putchar('0');
putchar('\n');
}
last_tid++;
last_pos = -1;
if (all < 2 || reg)
break;
}
}
depth_end:
for (i = 0; i < n && data[i]; ++i) {
bam_hdr_destroy(data[i]->hdr);
if (data[i]->fp) sam_close(data[i]->fp);
hts_itr_destroy(data[i]->iter);
free(data[i]);
}
free(data);
free(reg);
if (bed) bed_destroy(bed);
if ( file_list )
{
for (i=0; i<n; i++) free(fn[i]);
free(fn);
}
sam_global_args_free(&ga);
return status;
}
int main_depth(int argc, char *argv[])
#endif
{
int i, n, tid, beg, end, pos, *n_plp, baseQ = 0, mapQ = 0, min_len = 0, nfiles;
const bam_pileup1_t **plp;
char *reg = 0; // specified region
void *bed = 0; // BED data structure
char *file_list = NULL, **fn = NULL;
bam_header_t *h = 0; // BAM header of the 1st input
aux_t **data;
bam_mplp_t mplp;
// parse the command line
while ((n = getopt(argc, argv, "r:b:q:Q:l:f:")) >= 0) {
switch (n) {
case 'l': min_len = atoi(optarg); break; // minimum query length
case 'r': reg = strdup(optarg); break; // parsing a region requires a BAM header
case 'b': bed = bed_read(optarg); break; // BED or position list file can be parsed now
case 'q': baseQ = atoi(optarg); break; // base quality threshold
case 'Q': mapQ = atoi(optarg); break; // mapping quality threshold
case 'f': file_list = optarg; break;
}
}
if (optind == argc && !file_list) {
fprintf(stderr, "\n");
fprintf(stderr, "Usage: samtools depth [options] in1.bam [in2.bam [...]]\n");
fprintf(stderr, "Options:\n");
fprintf(stderr, " -b <bed> list of positions or regions\n");
fprintf(stderr, " -f <list> list of input BAM filenames, one per line [null]\n");
fprintf(stderr, " -l <int> minQLen\n");
fprintf(stderr, " -q <int> base quality threshold\n");
fprintf(stderr, " -Q <int> mapping quality threshold\n");
fprintf(stderr, " -r <chr:from-to> region\n");
fprintf(stderr, "\n");
return 1;
}
// initialize the auxiliary data structures
if (file_list)
{
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
n = nfiles;
argv = fn;
optind = 0;
}
else
n = argc - optind; // the number of BAMs on the command line
data = calloc(n, sizeof(void*)); // data[i] for the i-th input
beg = 0; end = 1<<30; tid = -1; // set the default region
for (i = 0; i < n; ++i) {
bam_header_t *htmp;
data[i] = calloc(1, sizeof(aux_t));
data[i]->fp = bam_open(argv[optind+i], "r"); // open BAM
data[i]->min_mapQ = mapQ; // set the mapQ filter
data[i]->min_len = min_len; // set the qlen filter
htmp = bam_header_read(data[i]->fp); // read the BAM header
if (i == 0) {
h = htmp; // keep the header of the 1st BAM
if (reg) bam_parse_region(h, reg, &tid, &beg, &end); // also parse the region
} else bam_header_destroy(htmp); // if not the 1st BAM, trash the header
if (tid >= 0) { // if a region is specified and parsed successfully
bam_index_t *idx = bam_index_load(argv[optind+i]); // load the index
data[i]->iter = bam_iter_query(idx, tid, beg, end); // set the iterator
bam_index_destroy(idx); // the index is not needed any more; phase out of the memory
}
}
// the core multi-pileup loop
mplp = bam_mplp_init(n, read_bam, (void**)data); // initialization
bam_mplp_set_maxcnt(mplp,2147483647); // set max_depth to int max
n_plp = calloc(n, sizeof(int)); // n_plp[i] is the number of covering reads from the i-th BAM
plp = calloc(n, sizeof(void*)); // plp[i] points to the array of covering reads (internal in mplp)
while (bam_mplp_auto(mplp, &tid, &pos, n_plp, plp) > 0) { // come to the next covered position
if (pos < beg || pos >= end) continue; // out of range; skip
if (bed && bed_overlap(bed, h->target_name[tid], pos, pos + 1) == 0) continue; // not in BED; skip
fputs(h->target_name[tid], stdout); printf("\t%d", pos+1); // a customized printf() would be faster
for (i = 0; i < n; ++i) { // base level filters have to go here
int j, m = 0;
for (j = 0; j < n_plp[i]; ++j) {
const bam_pileup1_t *p = plp[i] + j; // DON'T modfity plp[][] unless you really know
if (p->is_del || p->is_refskip) ++m; // having dels or refskips at tid:pos
else if (bam1_qual(p->b)[p->qpos] < baseQ) ++m; // low base quality
}
printf("\t%d", n_plp[i] - m); // this the depth to output
}
putchar('\n');
}
free(n_plp); free(plp);
bam_mplp_destroy(mplp);
bam_header_destroy(h);
for (i = 0; i < n; ++i) {
bam_close(data[i]->fp);
if (data[i]->iter) bam_iter_destroy(data[i]->iter);
free(data[i]);
}
free(data); free(reg);
if (bed) bed_destroy(bed);
if ( file_list )
{
for (i=0; i<n; i++) free(fn[i]);
free(fn);
}
return 0;
}
int main_samview(int argc, char *argv[])
{
int index;
for(index = 0; index < argc; index++) {
printf("The %d is %s\n",index,argv[index]);
}
getchar();return 0;
int c, is_header = 0, is_header_only = 0, ret = 0, compress_level = -1, is_count = 0;
int is_long_help = 0, n_threads = 0;
int64_t count = 0;
samFile *in = 0, *out = 0, *un_out=0;
bam_hdr_t *header = NULL;
char out_mode[5], out_un_mode[5], *out_format = "";
char *fn_in = 0, *fn_out = 0, *fn_list = 0, *q, *fn_un_out = 0;
sam_global_args ga = SAM_GLOBAL_ARGS_INIT;
samview_settings_t settings = {
.rghash = NULL,
.min_mapQ = 0,
.flag_on = 0,
.flag_off = 0,
.min_qlen = 0,
.remove_B = 0,
.subsam_seed = 0,
.subsam_frac = -1.,
.library = NULL,
.bed = NULL,
};
static const struct option lopts[] = {
SAM_OPT_GLOBAL_OPTIONS('-', 0, 'O', 0, 'T'),
{ "threads", required_argument, NULL, '@' },
{ NULL, 0, NULL, 0 }
};
/* parse command-line options */
strcpy(out_mode, "w");
strcpy(out_un_mode, "w");
while ((c = getopt_long(argc, argv,
"SbBcCt:h1Ho:O:q:f:F:ul:r:?T:R:L:s:@:m:x:U:",
lopts, NULL)) >= 0) {
switch (c) {
case 's':
if ((settings.subsam_seed = strtol(optarg, &q, 10)) != 0) {
srand(settings.subsam_seed);
settings.subsam_seed = rand();
}
settings.subsam_frac = strtod(q, &q);
break;
case 'm': settings.min_qlen = atoi(optarg); break;
case 'c': is_count = 1; break;
case 'S': break;
case 'b': out_format = "b"; break;
case 'C': out_format = "c"; break;
case 't': fn_list = strdup(optarg); break;
case 'h': is_header = 1; break;
case 'H': is_header_only = 1; break;
case 'o': fn_out = strdup(optarg); break;
case 'U': fn_un_out = strdup(optarg); break;
case 'f': settings.flag_on |= strtol(optarg, 0, 0); break;
case 'F': settings.flag_off |= strtol(optarg, 0, 0); break;
case 'q': settings.min_mapQ = atoi(optarg); break;
case 'u': compress_level = 0; break;
case '1': compress_level = 1; break;
case 'l': settings.library = strdup(optarg); break;
case 'L':
if ((settings.bed = bed_read(optarg)) == NULL) {
print_error_errno("view", "Could not read file \"%s\"", optarg);
ret = 1;
goto view_end;
}
break;
case 'r':
if (add_read_group_single("view", &settings, optarg) != 0) {
ret = 1;
goto view_end;
}
break;
case 'R':
if (add_read_groups_file("view", &settings, optarg) != 0) {
ret = 1;
goto view_end;
}
break;
/* REMOVED as htslib doesn't support this
//case 'x': out_format = "x"; break;
//case 'X': out_format = "X"; break;
*/
case '?': is_long_help = 1; break;
case 'B': settings.remove_B = 1; break;
case '@': n_threads = strtol(optarg, 0, 0); break;
case 'x':
{
if (strlen(optarg) != 2) {
fprintf(stderr, "main_samview: Error parsing -x auxiliary tags should be exactly two characters long.\n");
return usage(stderr, EXIT_FAILURE, is_long_help);
}
settings.remove_aux = (char**)realloc(settings.remove_aux, sizeof(char*) * (++settings.remove_aux_len));
settings.remove_aux[settings.remove_aux_len-1] = optarg;
}
break;
default:
if (parse_sam_global_opt(c, optarg, lopts, &ga) != 0)
return usage(stderr, EXIT_FAILURE, is_long_help);
break;
}
}
if (compress_level >= 0 && !*out_format) out_format = "b";
if (is_header_only) is_header = 1;
// File format auto-detection first
if (fn_out) sam_open_mode(out_mode+1, fn_out, NULL);
if (fn_un_out) sam_open_mode(out_un_mode+1, fn_un_out, NULL);
// Overridden by manual -b, -C
if (*out_format)
out_mode[1] = out_un_mode[1] = *out_format;
out_mode[2] = out_un_mode[2] = '\0';
// out_(un_)mode now 1 or 2 bytes long, followed by nul.
if (compress_level >= 0) {
char tmp[2];
tmp[0] = compress_level + '0'; tmp[1] = '\0';
strcat(out_mode, tmp);
strcat(out_un_mode, tmp);
}
if (argc == optind && isatty(STDIN_FILENO)) return usage(stdout, EXIT_SUCCESS, is_long_help); // potential memory leak...
fn_in = (optind < argc)? argv[optind] : "-";
// generate the fn_list if necessary
if (fn_list == 0 && ga.reference) fn_list = samfaipath(ga.reference);
// open file handlers
if ((in = sam_open_format(fn_in, "r", &ga.in)) == 0) {
print_error_errno("view", "failed to open \"%s\" for reading", fn_in);
ret = 1;
goto view_end;
}
if (fn_list) {
if (hts_set_fai_filename(in, fn_list) != 0) {
fprintf(stderr, "[main_samview] failed to use reference \"%s\".\n", fn_list);
ret = 1;
goto view_end;
}
}
if ((header = sam_hdr_read(in)) == 0) {
fprintf(stderr, "[main_samview] fail to read the header from \"%s\".\n", fn_in);
ret = 1;
goto view_end;
}
if (settings.rghash) { // FIXME: I do not know what "bam_header_t::n_text" is for...
char *tmp;
int l;
tmp = drop_rg(header->text, settings.rghash, &l);
free(header->text);
header->text = tmp;
header->l_text = l;
}
if (!is_count) {
if ((out = sam_open_format(fn_out? fn_out : "-", out_mode, &ga.out)) == 0) {
print_error_errno("view", "failed to open \"%s\" for writing", fn_out? fn_out : "standard output");
ret = 1;
goto view_end;
}
if (fn_list) {
if (hts_set_fai_filename(out, fn_list) != 0) {
fprintf(stderr, "[main_samview] failed to use reference \"%s\".\n", fn_list);
ret = 1;
goto view_end;
}
}
if (*out_format || is_header ||
out_mode[1] == 'b' || out_mode[1] == 'c' ||
(ga.out.format != sam && ga.out.format != unknown_format)) {
if (sam_hdr_write(out, header) != 0) {
fprintf(stderr, "[main_samview] failed to write the SAM header\n");
ret = 1;
goto view_end;
}
}
if (fn_un_out) {
if ((un_out = sam_open_format(fn_un_out, out_un_mode, &ga.out)) == 0) {
print_error_errno("view", "failed to open \"%s\" for writing", fn_un_out);
ret = 1;
goto view_end;
}
if (fn_list) {
if (hts_set_fai_filename(un_out, fn_list) != 0) {
fprintf(stderr, "[main_samview] failed to use reference \"%s\".\n", fn_list);
ret = 1;
goto view_end;
}
}
if (*out_format || is_header ||
out_un_mode[1] == 'b' || out_un_mode[1] == 'c' ||
(ga.out.format != sam && ga.out.format != unknown_format)) {
if (sam_hdr_write(un_out, header) != 0) {
fprintf(stderr, "[main_samview] failed to write the SAM header\n");
ret = 1;
goto view_end;
}
}
}
}
if (n_threads > 1) { if (out) hts_set_threads(out, n_threads); }
if (is_header_only) goto view_end; // no need to print alignments
if (optind + 1 >= argc) { // convert/print the entire file
bam1_t *b = bam_init1();
int r;
while ((r = sam_read1(in, header, b)) >= 0) { // read one alignment from `in'
if (!process_aln(header, b, &settings)) {
if (!is_count) { if (check_sam_write1(out, header, b, fn_out, &ret) < 0) break; }
count++;
} else {
if (un_out) { if (check_sam_write1(un_out, header, b, fn_un_out, &ret) < 0) break; }
}
}
if (r < -1) {
fprintf(stderr, "[main_samview] truncated file.\n");
ret = 1;
}
bam_destroy1(b);
} else { // retrieve alignments in specified regions
int i;
bam1_t *b;
hts_idx_t *idx = sam_index_load(in, fn_in); // load index
if (idx == 0) { // index is unavailable
fprintf(stderr, "[main_samview] random alignment retrieval only works for indexed BAM or CRAM files.\n");
ret = 1;
goto view_end;
}
b = bam_init1();
for (i = optind + 1; i < argc; ++i) {
int result;
hts_itr_t *iter = sam_itr_querys(idx, header, argv[i]); // parse a region in the format like `chr2:100-200'
if (iter == NULL) { // region invalid or reference name not found
int beg, end;
if (hts_parse_reg(argv[i], &beg, &end))
fprintf(stderr, "[main_samview] region \"%s\" specifies an unknown reference name. Continue anyway.\n", argv[i]);
else
fprintf(stderr, "[main_samview] region \"%s\" could not be parsed. Continue anyway.\n", argv[i]);
continue;
}
// fetch alignments
while ((result = sam_itr_next(in, iter, b)) >= 0) {
if (!process_aln(header, b, &settings)) {
if (!is_count) { if (check_sam_write1(out, header, b, fn_out, &ret) < 0) break; }
count++;
} else {
if (un_out) { if (check_sam_write1(un_out, header, b, fn_un_out, &ret) < 0) break; }
}
}
hts_itr_destroy(iter);
if (result < -1) {
fprintf(stderr, "[main_samview] retrieval of region \"%s\" failed due to truncated file or corrupt BAM index file\n", argv[i]);
ret = 1;
break;
}
}
bam_destroy1(b);
hts_idx_destroy(idx); // destroy the BAM index
}
view_end:
if (is_count && ret == 0)
printf("%" PRId64 "\n", count);
// close files, free and return
if (in) check_sam_close("view", in, fn_in, "standard input", &ret);
if (out) check_sam_close("view", out, fn_out, "standard output", &ret);
if (un_out) check_sam_close("view", un_out, fn_un_out, "file", &ret);
free(fn_list); free(fn_out); free(settings.library); free(fn_un_out);
sam_global_args_free(&ga);
if ( header ) bam_hdr_destroy(header);
if (settings.bed) bed_destroy(settings.bed);
if (settings.rghash) {
khint_t k;
for (k = 0; k < kh_end(settings.rghash); ++k)
if (kh_exist(settings.rghash, k)) free((char*)kh_key(settings.rghash, k));
kh_destroy(rg, settings.rghash);
}
if (settings.remove_aux_len) {
free(settings.remove_aux);
}
return ret;
}
static int usage(FILE *fp, int exit_status, int is_long_help)
{
fprintf(fp,
"\n"
"Usage: samtools view [options] <in.bam>|<in.sam>|<in.cram> [region ...]\n"
"\n"
"Options:\n"
// output options
" -b output BAM\n"
" -C output CRAM (requires -T)\n"
" -1 use fast BAM compression (implies -b)\n"
" -u uncompressed BAM output (implies -b)\n"
" -h include header in SAM output\n"
" -H print SAM header only (no alignments)\n"
" -c print only the count of matching records\n"
" -o FILE output file name [stdout]\n"
" -U FILE output reads not selected by filters to FILE [null]\n"
// extra input
" -t FILE FILE listing reference names and lengths (see long help) [null]\n"
// read filters
" -L FILE only include reads overlapping this BED FILE [null]\n"
" -r STR only include reads in read group STR [null]\n"
" -R FILE only include reads with read group listed in FILE [null]\n"
" -q INT only include reads with mapping quality >= INT [0]\n"
" -l STR only include reads in library STR [null]\n"
" -m INT only include reads with number of CIGAR operations consuming\n"
" query sequence >= INT [0]\n"
" -f INT only include reads with all bits set in INT set in FLAG [0]\n"
" -F INT only include reads with none of the bits set in INT set in FLAG [0]\n"
// read processing
" -x STR read tag to strip (repeatable) [null]\n"
" -B collapse the backward CIGAR operation\n"
" -s FLOAT integer part sets seed of random number generator [0];\n"
" rest sets fraction of templates to subsample [no subsampling]\n"
// general options
" -@, --threads INT\n"
" number of BAM/CRAM compression threads [0]\n"
" -? print long help, including note about region specification\n"
" -S ignored (input format is auto-detected)\n");
sam_global_opt_help(fp, "-.O.T");
fprintf(fp, "\n");
if (is_long_help)
fprintf(fp,
"Notes:\n"
"\n"
"1. This command now auto-detects the input format (BAM/CRAM/SAM).\n"
" Further control over the CRAM format can be specified by using the\n"
" --output-fmt-option, e.g. to specify the number of sequences per slice\n"
" and to use avoid reference based compression:\n"
"\n"
"\tsamtools view -C --output-fmt-option seqs_per_slice=5000 \\\n"
"\t --output-fmt-option no_ref -o out.cram in.bam\n"
"\n"
" Options can also be specified as a comma separated list within the\n"
" --output-fmt value too. For example this is equivalent to the above\n"
"\n"
"\tsamtools view --output-fmt cram,seqs_per_slice=5000,no_ref \\\n"
"\t -o out.cram in.bam\n"
"\n"
"2. The file supplied with `-t' is SPACE/TAB delimited with the first\n"
" two fields of each line consisting of the reference name and the\n"
" corresponding sequence length. The `.fai' file generated by \n"
" `samtools faidx' is suitable for use as this file. This may be an\n"
" empty file if reads are unaligned.\n"
"\n"
"3. SAM->BAM conversion: samtools view -bT ref.fa in.sam.gz\n"
"\n"
"4. BAM->SAM conversion: samtools view -h in.bam\n"
"\n"
"5. A region should be presented in one of the following formats:\n"
" `chr1', `chr2:1,000' and `chr3:1000-2,000'. When a region is\n"
" specified, the input alignment file must be a sorted and indexed\n"
" alignment (BAM/CRAM) file.\n"
"\n"
"6. Option `-u' is preferred over `-b' when the output is piped to\n"
" another samtools command.\n"
"\n");
return exit_status;
}
int bam_mpileup(int argc, char *argv[])
{
int c;
const char *file_list = NULL;
char **fn = NULL;
int nfiles = 0, use_orphan = 0;
mplp_conf_t mplp;
memset(&mplp, 0, sizeof(mplp_conf_t));
mplp.min_baseQ = 13;
mplp.capQ_thres = 0;
mplp.max_depth = 250; mplp.max_indel_depth = 250;
mplp.openQ = 40; mplp.extQ = 20; mplp.tandemQ = 100;
mplp.min_frac = 0.002; mplp.min_support = 1;
mplp.flag = MPLP_NO_ORPHAN | MPLP_REALN | MPLP_SMART_OVERLAPS;
mplp.argc = argc; mplp.argv = argv;
mplp.rflag_filter = BAM_FUNMAP | BAM_FSECONDARY | BAM_FQCFAIL | BAM_FDUP;
mplp.output_fname = NULL;
static const struct option lopts[] =
{
{"rf", required_argument, NULL, 1}, // require flag
{"ff", required_argument, NULL, 2}, // filter flag
{"incl-flags", required_argument, NULL, 1},
{"excl-flags", required_argument, NULL, 2},
{"output", required_argument, NULL, 3},
{"open-prob", required_argument, NULL, 4},
{"illumina1.3+", no_argument, NULL, '6'},
{"count-orphans", no_argument, NULL, 'A'},
{"bam-list", required_argument, NULL, 'b'},
{"no-BAQ", no_argument, NULL, 'B'},
{"no-baq", no_argument, NULL, 'B'},
{"adjust-MQ", required_argument, NULL, 'C'},
{"adjust-mq", required_argument, NULL, 'C'},
{"max-depth", required_argument, NULL, 'd'},
{"redo-BAQ", no_argument, NULL, 'E'},
{"redo-baq", no_argument, NULL, 'E'},
{"fasta-ref", required_argument, NULL, 'f'},
{"exclude-RG", required_argument, NULL, 'G'},
{"exclude-rg", required_argument, NULL, 'G'},
{"positions", required_argument, NULL, 'l'},
{"region", required_argument, NULL, 'r'},
{"ignore-RG", no_argument, NULL, 'R'},
{"ignore-rg", no_argument, NULL, 'R'},
{"min-MQ", required_argument, NULL, 'q'},
{"min-mq", required_argument, NULL, 'q'},
{"min-BQ", required_argument, NULL, 'Q'},
{"min-bq", required_argument, NULL, 'Q'},
{"ignore-overlaps", no_argument, NULL, 'x'},
{"BCF", no_argument, NULL, 'g'},
{"bcf", no_argument, NULL, 'g'},
{"VCF", no_argument, NULL, 'v'},
{"vcf", no_argument, NULL, 'v'},
{"output-BP", no_argument, NULL, 'O'},
{"output-bp", no_argument, NULL, 'O'},
{"output-MQ", no_argument, NULL, 's'},
{"output-mq", no_argument, NULL, 's'},
{"output-tags", required_argument, NULL, 't'},
{"uncompressed", no_argument, NULL, 'u'},
{"ext-prob", required_argument, NULL, 'e'},
{"gap-frac", required_argument, NULL, 'F'},
{"tandem-qual", required_argument, NULL, 'h'},
{"skip-indels", no_argument, NULL, 'I'},
{"max-idepth", required_argument, NULL, 'L'},
{"min-ireads ", required_argument, NULL, 'm'},
{"per-sample-mF", no_argument, NULL, 'p'},
{"per-sample-mf", no_argument, NULL, 'p'},
{"platforms", required_argument, NULL, 'P'},
{NULL, 0, NULL, 0}
};
while ((c = getopt_long(argc, argv, "Agf:r:l:q:Q:uRC:BDSd:L:b:P:po:e:h:Im:F:EG:6OsVvxt:",lopts,NULL)) >= 0) {
switch (c) {
case 'x': mplp.flag &= ~MPLP_SMART_OVERLAPS; break;
case 1 :
mplp.rflag_require = bam_str2flag(optarg);
if ( mplp.rflag_require<0 ) { fprintf(stderr,"Could not parse --rf %s\n", optarg); return 1; }
break;
case 2 :
mplp.rflag_filter = bam_str2flag(optarg);
if ( mplp.rflag_filter<0 ) { fprintf(stderr,"Could not parse --ff %s\n", optarg); return 1; }
break;
case 3 : mplp.output_fname = optarg; break;
case 4 : mplp.openQ = atoi(optarg); break;
case 'f':
mplp.fai = fai_load(optarg);
if (mplp.fai == 0) return 1;
mplp.fai_fname = optarg;
break;
case 'd': mplp.max_depth = atoi(optarg); break;
case 'r': mplp.reg = strdup(optarg); break;
case 'l':
// In the original version the whole BAM was streamed which is inefficient
// with few BED intervals and big BAMs. Todo: devise a heuristic to determine
// best strategy, that is streaming or jumping.
mplp.bed = bed_read(optarg);
if (!mplp.bed) { print_error_errno("Could not read file \"%s\"", optarg); return 1; }
break;
case 'P': mplp.pl_list = strdup(optarg); break;
case 'p': mplp.flag |= MPLP_PER_SAMPLE; break;
case 'g': mplp.flag |= MPLP_BCF; break;
case 'v': mplp.flag |= MPLP_BCF | MPLP_VCF; break;
case 'u': mplp.flag |= MPLP_NO_COMP | MPLP_BCF; break;
case 'B': mplp.flag &= ~MPLP_REALN; break;
case 'D': mplp.fmt_flag |= B2B_FMT_DP; fprintf(stderr, "[warning] samtools mpileup option `-D` is functional, but deprecated. Please switch to `-t DP` in future.\n"); break;
case 'S': mplp.fmt_flag |= B2B_FMT_SP; fprintf(stderr, "[warning] samtools mpileup option `-S` is functional, but deprecated. Please switch to `-t SP` in future.\n"); break;
case 'V': mplp.fmt_flag |= B2B_FMT_DV; fprintf(stderr, "[warning] samtools mpileup option `-V` is functional, but deprecated. Please switch to `-t DV` in future.\n"); break;
case 'I': mplp.flag |= MPLP_NO_INDEL; break;
case 'E': mplp.flag |= MPLP_REDO_BAQ; break;
case '6': mplp.flag |= MPLP_ILLUMINA13; break;
case 'R': mplp.flag |= MPLP_IGNORE_RG; break;
case 's': mplp.flag |= MPLP_PRINT_MAPQ; break;
case 'O': mplp.flag |= MPLP_PRINT_POS; break;
case 'C': mplp.capQ_thres = atoi(optarg); break;
case 'q': mplp.min_mq = atoi(optarg); break;
case 'Q': mplp.min_baseQ = atoi(optarg); break;
case 'b': file_list = optarg; break;
case 'o': {
char *end;
long value = strtol(optarg, &end, 10);
// Distinguish between -o INT and -o FILE (a bit of a hack!)
if (*end == '\0') mplp.openQ = value;
else mplp.output_fname = optarg;
}
break;
case 'e': mplp.extQ = atoi(optarg); break;
case 'h': mplp.tandemQ = atoi(optarg); break;
case 'A': use_orphan = 1; break;
case 'F': mplp.min_frac = atof(optarg); break;
case 'm': mplp.min_support = atoi(optarg); break;
case 'L': mplp.max_indel_depth = atoi(optarg); break;
case 'G': {
FILE *fp_rg;
char buf[1024];
mplp.rghash = khash_str2int_init();
if ((fp_rg = fopen(optarg, "r")) == 0)
fprintf(stderr, "(%s) Fail to open file %s. Continue anyway.\n", __func__, optarg);
while (!feof(fp_rg) && fscanf(fp_rg, "%s", buf) > 0) // this is not a good style, but forgive me...
khash_str2int_inc(mplp.rghash, strdup(buf));
fclose(fp_rg);
}
break;
case 't': mplp.fmt_flag |= parse_format_flag(optarg); break;
default:
fprintf(stderr,"Invalid option: '%c'\n", c);
return 1;
}
}
if ( !(mplp.flag&MPLP_REALN) && mplp.flag&MPLP_REDO_BAQ )
{
fprintf(stderr,"Error: The -B option cannot be combined with -E\n");
return 1;
}
if (use_orphan) mplp.flag &= ~MPLP_NO_ORPHAN;
if (argc == 1)
{
print_usage(stderr, &mplp);
return 1;
}
int ret;
if (file_list) {
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
ret = mpileup(&mplp,nfiles,fn);
for (c=0; c<nfiles; c++) free(fn[c]);
free(fn);
}
else
ret = mpileup(&mplp, argc - optind, argv + optind);
if (mplp.rghash) khash_str2int_destroy_free(mplp.rghash);
free(mplp.reg); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);
if (mplp.bed) bed_destroy(mplp.bed);
return ret;
}
int main_depth(int argc, char *argv[])
{
int i, n, tid, beg, end, pos, *n_plp, baseQ = 0, mapQ = 0, min_len = 0, status = EXIT_SUCCESS, nfiles;
const bam_pileup1_t **plp;
char *reg = 0; // specified region
void *bed = 0; // BED data structure
char *file_list = NULL, **fn = NULL;
bam_hdr_t *h = NULL; // BAM header of the 1st input
aux_t **data;
bam_mplp_t mplp;
// parse the command line
while ((n = getopt(argc, argv, "r:b:q:Q:l:f:")) >= 0) {
switch (n) {
case 'l': min_len = atoi(optarg); break; // minimum query length
case 'r': reg = strdup(optarg); break; // parsing a region requires a BAM header
case 'b':
bed = bed_read(optarg); // BED or position list file can be parsed now
if (!bed) { print_error_errno("Could not read file \"%s\"", optarg); return 1; }
break;
case 'q': baseQ = atoi(optarg); break; // base quality threshold
case 'Q': mapQ = atoi(optarg); break; // mapping quality threshold
case 'f': file_list = optarg; break;
}
}
if (optind == argc && !file_list) {
fprintf(pysamerr, "\n");
fprintf(pysamerr, "Usage: samtools depth [options] in1.bam [in2.bam [...]]\n");
fprintf(pysamerr, "Options:\n");
fprintf(pysamerr, " -b <bed> list of positions or regions\n");
fprintf(pysamerr, " -f <list> list of input BAM filenames, one per line [null]\n");
fprintf(pysamerr, " -l <int> read length threshold (ignore reads shorter than <int>)\n");
fprintf(pysamerr, " -q <int> base quality threshold\n");
fprintf(pysamerr, " -Q <int> mapping quality threshold\n");
fprintf(pysamerr, " -r <chr:from-to> region\n");
fprintf(pysamerr, "\n");
return 1;
}
// initialize the auxiliary data structures
if (file_list)
{
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
n = nfiles;
argv = fn;
optind = 0;
}
else
n = argc - optind; // the number of BAMs on the command line
data = calloc(n, sizeof(aux_t*)); // data[i] for the i-th input
beg = 0; end = 1<<30; // set the default region
for (i = 0; i < n; ++i) {
data[i] = calloc(1, sizeof(aux_t));
data[i]->fp = sam_open(argv[optind+i], "r"); // open BAM
if (data[i]->fp == NULL) {
print_error_errno("Could not open \"%s\"", argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
if (hts_set_opt(data[i]->fp, CRAM_OPT_REQUIRED_FIELDS,
SAM_FLAG | SAM_RNAME | SAM_POS | SAM_MAPQ | SAM_CIGAR |
SAM_SEQ)) {
fprintf(pysamerr, "Failed to set CRAM_OPT_REQUIRED_FIELDS value\n");
return 1;
}
if (hts_set_opt(data[i]->fp, CRAM_OPT_DECODE_MD, 0)) {
fprintf(pysamerr, "Failed to set CRAM_OPT_DECODE_MD value\n");
return 1;
}
data[i]->min_mapQ = mapQ; // set the mapQ filter
data[i]->min_len = min_len; // set the qlen filter
data[i]->hdr = sam_hdr_read(data[i]->fp); // read the BAM header
if (reg) { // if a region is specified
hts_idx_t *idx = sam_index_load(data[i]->fp, argv[optind+i]); // load the index
if (idx == NULL) {
print_error("can't load index for \"%s\"", argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
data[i]->iter = sam_itr_querys(idx, data[i]->hdr, reg); // set the iterator
hts_idx_destroy(idx); // the index is not needed any more; free the memory
if (data[i]->iter == NULL) {
print_error("can't parse region \"%s\"", reg);
status = EXIT_FAILURE;
goto depth_end;
}
}
}
h = data[0]->hdr; // easy access to the header of the 1st BAM
if (reg) {
beg = data[0]->iter->beg; // and to the parsed region coordinates
end = data[0]->iter->end;
}
// the core multi-pileup loop
mplp = bam_mplp_init(n, read_bam, (void**)data); // initialization
n_plp = calloc(n, sizeof(int)); // n_plp[i] is the number of covering reads from the i-th BAM
plp = calloc(n, sizeof(bam_pileup1_t*)); // plp[i] points to the array of covering reads (internal in mplp)
while (bam_mplp_auto(mplp, &tid, &pos, n_plp, plp) > 0) { // come to the next covered position
if (pos < beg || pos >= end) continue; // out of range; skip
if (bed && bed_overlap(bed, h->target_name[tid], pos, pos + 1) == 0) continue; // not in BED; skip
fputs(h->target_name[tid], stdout); printf("\t%d", pos+1); // a customized printf() would be faster
for (i = 0; i < n; ++i) { // base level filters have to go here
int j, m = 0;
for (j = 0; j < n_plp[i]; ++j) {
const bam_pileup1_t *p = plp[i] + j; // DON'T modfity plp[][] unless you really know
if (p->is_del || p->is_refskip) ++m; // having dels or refskips at tid:pos
else if (bam_get_qual(p->b)[p->qpos] < baseQ) ++m; // low base quality
}
printf("\t%d", n_plp[i] - m); // this the depth to output
}
putchar('\n');
}
free(n_plp); free(plp);
bam_mplp_destroy(mplp);
depth_end:
for (i = 0; i < n && data[i]; ++i) {
bam_hdr_destroy(data[i]->hdr);
if (data[i]->fp) sam_close(data[i]->fp);
hts_itr_destroy(data[i]->iter);
free(data[i]);
}
free(data); free(reg);
if (bed) bed_destroy(bed);
if ( file_list )
{
for (i=0; i<n; i++) free(fn[i]);
free(fn);
}
return status;
}
int main(int argc, char *argv[])
{
int c, i, n, ret, res;
int tid, pos, *n_plp;
cmdopt_t o;
bam_mplp_t mplp;
const bam_pileup1_t **plp;
aux_t **data;
bam_hdr_t *h = 0;
sv_t sv1;
qual_sum_t qual2;
khiter_t k_iter;
khash_t(sv_hash) *sv_h = kh_init(sv_hash);
khash_t(sv_geno) *geno_h = kh_init(sv_geno);
khash_t(colmap) *smp_cols;
khash_t(ped) *ped_h = 0;
mempool_t *mp;
char **samples;
o.min_q = 40; o.min_s = 80; o.min_len = 150; o.min_dp = 10; o.bed = 0, o.fnped = 0, o.mi_prob=0.005;
while ((c = getopt(argc, argv, "hq:s:l:d:b:p:m:")) >= 0) {
if (c == 'h') { usage(stderr, &o); return 0; }
else if (c == 'q') o.min_q = atoi(optarg);
else if (c == 's') o.min_s = atoi(optarg);
else if (c == 'l') o.min_len = atoi(optarg);
else if (c == 'd') o.min_dp = atoi(optarg);
else if (c == 'p') o.fnped = optarg;
else if (c == 'm') o.mi_prob = atof(optarg);
else if (c == 'b') {
if ((o.bed = bed_read(optarg)) == NULL) {
return -1;
}
}
}
if (o.mi_prob < 0.0000000000001 || o.mi_prob > 0.1) {
fprintf(stderr, "Error. Probability of a mendelian inconsistency must be between 0.1 and 0.0000000000001.\n");
}
if (argc - optind < 1) {
usage(stderr, &o);
return 1;
}
// Open files and initalize aux data //
n = argc - optind;
data = calloc(n, sizeof(aux_t*));
samples = (char**)malloc(n * sizeof(char*));
for (i = 0; i < n; ++i) {
data[i] = calloc(1, sizeof (aux_t));
data[i]->fp = sam_open(argv[optind + i], "r");
if (!data[i]->fp) {
fprintf(stderr, "Input file \"%s\" could not be opened.\n", argv[optind + 1]);
return 1;
}
data[i]->min_mapq = o.min_q;
data[i]->min_as = o.min_s;
data[i]->min_len = o.min_len;
data[i]->hdr = sam_hdr_read(data[i]->fp);
if (!data[i]->hdr) {
fprintf(stderr, "Could not read the header for input file \"%s\".\n", argv[optind + 1]);
return 1;
}
samples[i] = find_sample(data[i]->hdr, &res);
if (!samples[i]) {
fprintf(stderr, "Warning. No sample name detected for bam %s. Using filename\n", argv[optind + i]);
samples[i] = argv[optind + i];
}
}
h = data[0]->hdr;
smp_cols = map_samples(samples, n);
if (o.fnped) {
if ((ped_h = read_ped(o.fnped, smp_cols)) == 0) { return -1; }
}
// The core data processing loop //
mplp = bam_mplp_init(n, read_bam, (void**)data);
n_plp = calloc(n, sizeof(int)); // n_plp[i] is the number of covering reads from the i-th BAM
plp = calloc(n, sizeof(bam_pileup1_t*)); // plp[i] points to the array of covering reads in mplp
//quals = (qual_vec_t*)calloc(n, sizeof(qual_vec_t));
mp = mp_init();
while ((ret = bam_mplp_auto(mplp, &tid, &pos, n_plp, plp)) > 0) { // iterate of positions with coverage
int n_sv;
if (o.bed && tid >= 0 && !bed_overlap(o.bed, h->target_name[tid], pos, pos+1)) continue;
n_sv = plp2sv(h, tid, pos, n, n_plp, plp, sv_h);
if (n_sv > 1) { fprintf(stderr, "Warning: more than two alleles detected at %s:%d\n", h->target_name[tid], pos); }
if (n_sv) {
fprintf(stderr, "SV detected at %d:%d\n", tid, pos);
for (k_iter = kh_begin(sv_h); k_iter != kh_end(sv_h); ++k_iter) {
if (kh_exist(sv_h, k_iter)) {
sv1 = kh_value(sv_h, k_iter);
fprintf(stderr, "SV tid1=%d, tid2=%d, pos1=%d, pos2=%d, ori1=%d, ori2=%d, allele=%d\n", sv1.tid1, sv1.tid2, sv1.pos1, sv1.pos2, sv1.ori1, sv1.ori2, sv1.allele);
}
}
res = get_qual_data(h, tid, pos, n, n_plp, plp, n_sv + 1, sv_h, geno_h, mp);
if (res < 0) {
fprintf(stderr, "Error collecting quality data from reads\n");
return -1;
}
kh_clear(sv_hash, sv_h);
}
}
print_header(h, optind, n, argv);
genotype_sv(h, n, geno_h, o.min_dp, ped_h, o.mi_prob);
free(n_plp);
free(plp);
bam_mplp_destroy(mplp);
mp_destroy(mp);
if (o.bed) bed_destroy(o.bed);
for (i = 0; i < n; ++i) {
bam_hdr_destroy(data[i]->hdr);
sam_close(data[i]->fp);
free(data[i]);
free(samples[i]);
}
free(data);
free(samples);
kh_destroy(sv_hash, sv_h);
kh_destroy(sv_geno, geno_h);
kh_destroy(colmap, smp_cols);
kh_destroy(ped, ped_h);
return 0;
}
int main_view(int argc, char *argv[])
{
int i, c, n_files = 0, out_bcf = 0, clevel = -1, multi_flag = 0, excl = 0, not_vcf = 0, in_mem = 0, u_set = 0;
long seekn = -1, n_rec = LONG_MAX, n_read = 0;
bgtm_t *bm = 0;
bcf1_t *b;
htsFile *out = 0;
char modew[8], *reg = 0, *site_flt = 0;
void *bed = 0;
int n_groups = 0;
char *gexpr[BGT_MAX_GROUPS], *aexpr = 0, *dbfn = 0, *fmt = 0;
bgt_file_t **files = 0;
fmf_t *vardb = 0;
while ((c = getopt(argc, argv, "ubs:r:l:CMGB:ef:g:a:i:n:SHt:d:")) >= 0) {
if (c == 'b') out_bcf = 1;
else if (c == 'r') reg = optarg;
else if (c == 'l') clevel = atoi(optarg);
else if (c == 'e') excl = 1;
else if (c == 'u') u_set = 1;
else if (c == 'B') bed = bed_read(optarg);
else if (c == 'C') multi_flag |= BGT_F_SET_AC;
else if (c == 'G') multi_flag |= BGT_F_NO_GT;
else if (c == 'S') multi_flag |= BGT_F_NO_GT | BGT_F_CNT_AL, not_vcf = 1;
else if (c == 'H') multi_flag |= BGT_F_NO_GT | BGT_F_CNT_HAP, not_vcf = 1;
else if (c == 'M') in_mem = 1;
else if (c == 'i') seekn = atol(optarg) - 1;
else if (c == 'n') n_rec = atol(optarg);
else if (c == 'f') site_flt = optarg;
else if (c == 't') fmt = optarg, not_vcf = 1;
else if (c == 'd') dbfn = optarg;
else if (c == 's' && n_groups < BGT_MAX_GROUPS) gexpr[n_groups++] = optarg;
else if (c == 'a') aexpr = optarg;
}
if (n_rec < 0) {
fprintf(stderr, "[E::%s] option -n must be at least 0.\n", __func__);
return 1;
}
if (clevel > 9) clevel = 9;
if (u_set) clevel = 0, out_bcf = 1;
if (n_groups > 1) multi_flag |= BGT_F_SET_AC;
if (argc - optind < 1) {
fprintf(stderr, "Usage: bgt %s [options] <bgt-prefix> [...]", argv[0]);
fputc('\n', stderr);
fprintf(stderr, "Options:\n");
fprintf(stderr, " Sample selection:\n");
fprintf(stderr, " -s EXPR samples list (,sample1,sample2 or a file or expr; see Notes below) [all]\n");
fprintf(stderr, " Site selection:\n");
fprintf(stderr, " -r STR region [all]\n");
fprintf(stderr, " -B FILE extract variants overlapping BED FILE []\n");
fprintf(stderr, " -e exclude variants overlapping BED FILE (effective with -B)\n");
fprintf(stderr, " -i INT process from the INT-th record (1-based) []\n");
fprintf(stderr, " -n INT process at most INT records []\n");
fprintf(stderr, " -d FILE variant annotations in FMF (to work with -a) []\n");
fprintf(stderr, " -M load variant annotations in RAM (only with -d)\n");
fprintf(stderr, " -a EXPR alleles list chr:1basedPos:refLen:seq (,allele1,allele2 or a file or expr) []\n");
fprintf(stderr, " -f STR frequency filters []\n");
fprintf(stderr, " VCF output:\n");
fprintf(stderr, " -b BCF output (effective without -S/-H)\n");
fprintf(stderr, " -l INT compression level for BCF [default]\n");
fprintf(stderr, " -u equivalent to -bl0 (overriding -b and -l)\n");
fprintf(stderr, " -G don't output sample genotypes\n");
fprintf(stderr, " -C write AC/AN to the INFO field (auto applied with -f or multipl -s)\n");
fprintf(stderr, " Non-VCF output:\n");
fprintf(stderr, " -S show samples with a set of alleles (with -a)\n");
fprintf(stderr, " -H count of haplotypes with a set of alleles (with -a)\n");
fprintf(stderr, " -t STR comma-delimited list of fields to output. Accepted variables:\n");
fprintf(stderr, " AC, AN, AC#, AN#, CHROM, POS, END, REF, ALT (# for a group number)\n");
fprintf(stderr, "Notes:\n");
fprintf(stderr, " For option -s/-a, EXPR can be one of:\n");
fprintf(stderr, " 1) comma-delimited list following a colon/comma. e.g. -s,NA12878,NA12044\n");
fprintf(stderr, " 2) space-delimited file with the first column giving a sample/allele name. e.g. -s list.txt\n");
fprintf(stderr, " 3) expression if .spl/-d file contains metadata. e.g.: -s\"gender=='M'&&population!='CEU'\"\n");
fprintf(stderr, " If multiple -s is specified, the AC/AN of the first group will be written to VCF INFO AC1/AN1,\n");
fprintf(stderr, " the second to AC2/AN2, etc.\n");
return 1;
}
if (dbfn && in_mem) vardb = fmf_read(dbfn), dbfn = 0;
if ((multi_flag&(BGT_F_CNT_AL|BGT_F_CNT_HAP)) && aexpr == 0) {
fprintf(stderr, "[E::%s] -a must be specified when -S/-H is in use.\n", __func__);
return 1;
}
n_files = argc - optind;
files = (bgt_file_t**)calloc(n_files, sizeof(bgt_file_t*));
for (i = 0; i < n_files; ++i) {
files[i] = bgt_open(argv[optind+i]);
if (files[i] == 0) {
fprintf(stderr, "[E::%s] failed to open BGT with prefix '%s'\n", __func__, argv[optind+i]);
return 1; // FIXME: memory leak
}
}
bm = bgtm_reader_init(n_files, files);
bgtm_set_flag(bm, multi_flag);
if (site_flt && bgtm_set_flt_site(bm, site_flt) != 0) {
fprintf(stderr, "[E::%s] failed to set frequency filters. Syntax error?\n", __func__);
return 1;
}
if (reg && bgtm_set_region(bm, reg) < 0) {
fprintf(stderr, "[E::%s] failed to set region. Region format error?\n", __func__);
return 1;
}
if (bed) bgtm_set_bed(bm, bed, excl);
if (fmt && bgtm_set_table(bm, fmt) < 0) {
fprintf(stderr, "[E::%s] failed to set tabular output.\n", __func__);
return 1;
}
if (seekn > 0) bgtm_set_start(bm, seekn);
if (aexpr) {
int n_al;
n_al = bgtm_set_alleles(bm, aexpr, vardb, dbfn);
if (n_al < 0) {
fprintf(stderr, "[E::%s] failed to set alleles.\n", __func__);
return 1;
} else if (n_al == 0)
fprintf(stderr, "[W::%s] no alleles selected.\n", __func__);
}
for (i = 0; i < n_groups; ++i) {
if (bgtm_add_group(bm, gexpr[i]) < 0) {
fprintf(stderr, "[E::%s] failed to add sample group '%s'.\n", __func__, gexpr[i]);
return 1;
}
}
bgtm_prepare(bm); // bgtm_prepare() generates the VCF header
if (!not_vcf) {
strcpy(modew, "w");
if (out_bcf) strcat(modew, "b");
sprintf(modew + strlen(modew), "%d", clevel);
out = hts_open("-", modew, 0);
vcf_hdr_write(out, bm->h_out);
}
b = bcf_init1();
while (bgtm_read(bm, b) >= 0 && n_read < n_rec) {
if (out) vcf_write1(out, bm->h_out, b);
if (fmt && bm->n_fields > 0) puts(bm->tbl_line.s);
++n_read;
}
bcf_destroy1(b);
if (not_vcf && bm->n_aal > 0) {
if (bm->flag & BGT_F_CNT_HAP) {
bgt_hapcnt_t *hc;
int n_hap;
char *s;
hc = bgtm_hapcnt(bm, &n_hap);
s = bgtm_hapcnt_print_destroy(bm, n_hap, hc);
fputs(s, stdout);
free(s);
}
if (bm->flag & BGT_F_CNT_AL) {
char *s;
if ((s = bgtm_alcnt_print(bm)) != 0)
fputs(s, stdout);
free(s);
}
}
if (out) hts_close(out);
bgtm_reader_destroy(bm);
if (bed) bed_destroy(bed);
for (i = 0; i < n_files; ++i) bgt_close(files[i]);
free(files);
if (vardb) fmf_destroy(vardb);
return 0;
}