static void init_data(args_t *args)
{
args->files = bcf_sr_init();
args->files->require_index = 1;
if ( !bcf_sr_add_reader(args->files,args->fname) ) error("Failed to open %s: %s\n", args->fname, bcf_sr_strerror(args->files->errnum));
args->hdr = args->files->readers[0].header;
args->isample = -1;
if ( args->sample )
{
args->isample = bcf_hdr_id2int(args->hdr,BCF_DT_SAMPLE,args->sample);
if ( args->isample<0 ) error("No such sample: %s\n", args->sample);
}
if ( args->haplotype && args->isample<0 )
{
if ( bcf_hdr_nsamples(args->hdr) > 1 ) error("The --sample option is expected with --haplotype\n");
args->isample = 0;
}
if ( args->mask_fname )
{
args->mask = regidx_init(args->mask_fname,NULL,NULL,0,NULL);
if ( !args->mask ) error("Failed to initialize mask regions\n");
}
// In case we want to store the chains
if ( args->chain_fname )
{
args->fp_chain = fopen(args->chain_fname,"w");
if ( ! args->fp_chain ) error("Failed to create %s: %s\n", args->chain_fname, strerror(errno));
args->chain_id = 0;
}
rbuf_init(&args->vcf_rbuf, 100);
args->vcf_buf = (bcf1_t**) calloc(args->vcf_rbuf.m, sizeof(bcf1_t*));
if ( args->output_fname ) {
args->fp_out = fopen(args->output_fname,"w");
if ( ! args->fp_out ) error("Failed to create %s: %s\n", args->output_fname, strerror(errno));
}
else args->fp_out = stdout;
}
int bam_mpileup(int argc, char *argv[])
{
int c;
const char *file_list = NULL;
char **fn = NULL;
int nfiles = 0, use_orphan = 0, noref = 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;
mplp.output_type = FT_VCF;
mplp.record_cmd_line = 1;
mplp.n_threads = 0;
mplp.bsmpl = bam_smpl_init();
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},
{"ignore-RG", no_argument, NULL, 5},
{"ignore-rg", no_argument, NULL, 5},
{"gvcf", required_argument, NULL, 'g'},
{"no-reference", no_argument, NULL, 7},
{"no-version", no_argument, NULL, 8},
{"threads",required_argument,NULL,9},
{"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'},
{"read-groups", required_argument, NULL, 'G'},
{"region", required_argument, NULL, 'r'},
{"regions", required_argument, NULL, 'r'},
{"regions-file", required_argument, NULL, 'R'},
{"targets", required_argument, NULL, 't'},
{"targets-file", required_argument, NULL, 'T'},
{"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'},
{"output-type", required_argument, NULL, 'O'},
{"samples", required_argument, NULL, 's'},
{"samples-file", required_argument, NULL, 'S'},
{"annotate", required_argument, NULL, 'a'},
{"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, "Ag:f:r:R:q:Q:C:Bd:L:b:P:po:e:h:Im:F:EG:6O:xa:s:S:t:T:",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 5 : bam_smpl_ignore_readgroups(mplp.bsmpl); break;
case 'g':
mplp.gvcf = gvcf_init(optarg);
if ( !mplp.gvcf ) error("Could not parse: --gvcf %s\n", optarg);
break;
case 'f':
mplp.fai = fai_load(optarg);
if (mplp.fai == NULL) return 1;
mplp.fai_fname = optarg;
break;
case 7 : noref = 1; break;
case 8 : mplp.record_cmd_line = 0; break;
case 9 : mplp.n_threads = strtol(optarg, 0, 0); break;
case 'd': mplp.max_depth = atoi(optarg); break;
case 'r': mplp.reg_fname = strdup(optarg); break;
case 'R': mplp.reg_fname = strdup(optarg); mplp.reg_is_file = 1; break;
case 't':
// 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.
if ( optarg[0]=='^' ) optarg++;
else mplp.bed_logic = 1;
mplp.bed = regidx_init(NULL,regidx_parse_reg,NULL,0,NULL);
mplp.bed_itr = regitr_init(mplp.bed);
if ( regidx_insert_list(mplp.bed,optarg,',') !=0 )
{
fprintf(stderr,"Could not parse the targets: %s\n", optarg);
exit(EXIT_FAILURE);
}
break;
case 'T':
if ( optarg[0]=='^' ) optarg++;
else mplp.bed_logic = 1;
mplp.bed = regidx_init(optarg,NULL,NULL,0,NULL);
if (!mplp.bed) { fprintf(stderr, "bcftools mpileup: 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 'B': mplp.flag &= ~MPLP_REALN; 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 's': if ( bam_smpl_add_samples(mplp.bsmpl,optarg,0)<0 ) error("Could not read samples: %s\n",optarg); break;
case 'S': if ( bam_smpl_add_samples(mplp.bsmpl,optarg,1)<0 ) error("Could not read samples: %s\n",optarg); break;
case 'O':
switch (optarg[0]) {
case 'b': mplp.output_type = FT_BCF_GZ; break;
case 'u': mplp.output_type = FT_BCF; break;
case 'z': mplp.output_type = FT_VCF_GZ; break;
case 'v': mplp.output_type = FT_VCF; break;
default: error("[error] The option \"-O\" changed meaning when mpileup moved to bcftools. Did you mean: \"bcftools mpileup --output-type\" or \"samtools mpileup --output-BP\"?\n");
}
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': bam_smpl_add_readgroups(mplp.bsmpl, optarg, 1); break;
case 'a':
if (optarg[0]=='?') {
list_annotations(stderr);
return 1;
}
mplp.fmt_flag |= parse_format_flag(optarg);
break;
default:
fprintf(stderr,"Invalid option: '%c'\n", c);
return 1;
}
}
if ( mplp.gvcf && !(mplp.fmt_flag&B2B_FMT_DP) )
{
fprintf(stderr,"[warning] The -t DP option is required with --gvcf, switching on.\n");
mplp.fmt_flag |= B2B_FMT_DP;
}
if ( mplp.flag&(MPLP_BCF|MPLP_VCF|MPLP_NO_COMP) )
{
if ( mplp.flag&MPLP_VCF )
{
if ( mplp.flag&MPLP_NO_COMP ) mplp.output_type = FT_VCF;
else mplp.output_type = FT_VCF_GZ;
}
else if ( mplp.flag&MPLP_BCF )
{
if ( mplp.flag&MPLP_NO_COMP ) mplp.output_type = FT_BCF;
else mplp.output_type = FT_BCF_GZ;
}
}
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;
}
if (!mplp.fai && !noref) {
fprintf(stderr,"Error: mpileup requires the --fasta-ref option by default; use --no-reference to run without a fasta reference\n");
return 1;
}
int ret,i;
if (file_list)
{
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
mplp.files = fn;
mplp.nfiles = nfiles;
}
else
{
mplp.nfiles = argc - optind;
mplp.files = (char**) malloc(mplp.nfiles*sizeof(char*));
for (i=0; i<mplp.nfiles; i++) mplp.files[i] = strdup(argv[optind+i]);
}
ret = mpileup(&mplp);
for (i=0; i<mplp.nfiles; i++) free(mplp.files[i]);
free(mplp.files);
free(mplp.reg_fname); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);
if (mplp.bed) regidx_destroy(mplp.bed);
if (mplp.bed_itr) regitr_destroy(mplp.bed_itr);
if (mplp.reg) regidx_destroy(mplp.reg);
bam_smpl_destroy(mplp.bsmpl);
return ret;
}
int run(int argc, char **argv)
{
char *trio_samples = NULL, *trio_file = NULL, *rules_fname = NULL, *rules_string = NULL;
memset(&args,0,sizeof(args_t));
args.mode = 0;
args.output_fname = "-";
static struct option loptions[] =
{
{"trio",1,0,'t'},
{"trio-file",1,0,'T'},
{"delete",0,0,'d'},
{"list",1,0,'l'},
{"count",0,0,'c'},
{"rules",1,0,'r'},
{"rules-file",1,0,'R'},
{"output",required_argument,NULL,'o'},
{"output-type",required_argument,NULL,'O'},
{0,0,0,0}
};
int c;
while ((c = getopt_long(argc, argv, "?ht:T:l:cdr:R:o:O:",loptions,NULL)) >= 0)
{
switch (c)
{
case 'o': args.output_fname = optarg; break;
case 'O':
switch (optarg[0]) {
case 'b': args.output_type = FT_BCF_GZ; break;
case 'u': args.output_type = FT_BCF; break;
case 'z': args.output_type = FT_VCF_GZ; break;
case 'v': args.output_type = FT_VCF; break;
default: error("The output type \"%s\" not recognised\n", optarg);
};
break;
case 'R': rules_fname = optarg; break;
case 'r': rules_string = optarg; break;
case 'd': args.mode |= MODE_DELETE; break;
case 'c': args.mode |= MODE_COUNT; break;
case 'l':
if ( !strcmp("+",optarg) ) args.mode |= MODE_LIST_GOOD;
else if ( !strcmp("x",optarg) ) args.mode |= MODE_LIST_BAD;
else error("The argument not recognised: --list %s\n", optarg);
break;
case 't': trio_samples = optarg; break;
case 'T': trio_file = optarg; break;
case 'h':
case '?':
default: error("%s",usage()); break;
}
}
if ( rules_fname )
args.rules = regidx_init(rules_fname, parse_rules, NULL, sizeof(rule_t), &args);
else
args.rules = init_rules(&args, rules_string);
if ( !args.rules ) return -1;
args.itr = regitr_init(args.rules);
args.itr_ori = regitr_init(args.rules);
char *fname = NULL;
if ( optind>=argc || argv[optind][0]=='-' )
{
if ( !isatty(fileno((FILE *)stdin)) ) fname = "-"; // reading from stdin
else error("%s",usage());
}
else
fname = argv[optind];
if ( !trio_samples && !trio_file ) error("Expected the -t/T option\n");
if ( !args.mode ) error("Expected one of the -c, -d or -l options\n");
if ( args.mode&MODE_DELETE && !(args.mode&(MODE_LIST_GOOD|MODE_LIST_BAD)) ) args.mode |= MODE_LIST_GOOD|MODE_LIST_BAD;
args.sr = bcf_sr_init();
if ( !bcf_sr_add_reader(args.sr, fname) ) error("Failed to open %s: %s\n", fname,bcf_sr_strerror(args.sr->errnum));
args.hdr = bcf_sr_get_header(args.sr, 0);
args.out_fh = hts_open(args.output_fname,hts_bcf_wmode(args.output_type));
if ( args.out_fh == NULL ) error("Can't write to \"%s\": %s\n", args.output_fname, strerror(errno));
bcf_hdr_write(args.out_fh, args.hdr);
int i, n = 0;
char **list;
if ( trio_samples )
{
args.ntrios = 1;
args.trios = (trio_t*) calloc(1,sizeof(trio_t));
list = hts_readlist(trio_samples, 0, &n);
if ( n!=3 ) error("Expected three sample names with -t\n");
args.trios[0].imother = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[0]);
args.trios[0].ifather = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[1]);
args.trios[0].ichild = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[2]);
for (i=0; i<n; i++) free(list[i]);
free(list);
}
if ( trio_file )
{
list = hts_readlist(trio_file, 1, &n);
args.ntrios = n;
args.trios = (trio_t*) calloc(n,sizeof(trio_t));
for (i=0; i<n; i++)
{
char *ss = list[i], *se;
se = strchr(ss, ',');
if ( !se ) error("Could not parse %s: %s\n",trio_file, ss);
*se = 0;
args.trios[i].imother = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, ss);
if ( args.trios[i].imother<0 ) error("No such sample: \"%s\"\n", ss);
ss = ++se;
se = strchr(ss, ',');
if ( !se ) error("Could not parse %s\n",trio_file);
*se = 0;
args.trios[i].ifather = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, ss);
if ( args.trios[i].ifather<0 ) error("No such sample: \"%s\"\n", ss);
ss = ++se;
if ( *ss=='\0' ) error("Could not parse %s\n",trio_file);
args.trios[i].ichild = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, ss);
if ( args.trios[i].ichild<0 ) error("No such sample: \"%s\"\n", ss);
free(list[i]);
}
free(list);
}
while ( bcf_sr_next_line(args.sr) )
{
bcf1_t *line = bcf_sr_get_line(args.sr,0);
line = process(line);
if ( line )
{
if ( line->errcode ) error("TODO: Unchecked error (%d), exiting\n",line->errcode);
bcf_write1(args.out_fh, args.hdr, line);
}
}
fprintf(stderr,"# [1]nOK\t[2]nBad\t[3]nSkipped\t[4]Trio\n");
for (i=0; i<args.ntrios; i++)
{
trio_t *trio = &args.trios[i];
fprintf(stderr,"%d\t%d\t%d\t%s,%s,%s\n",
trio->nok,trio->nbad,args.nrec-(trio->nok+trio->nbad),
bcf_hdr_int2id(args.hdr, BCF_DT_SAMPLE, trio->imother),
bcf_hdr_int2id(args.hdr, BCF_DT_SAMPLE, trio->ifather),
bcf_hdr_int2id(args.hdr, BCF_DT_SAMPLE, trio->ichild)
);
}
free(args.gt_arr);
free(args.trios);
regitr_destroy(args.itr);
regitr_destroy(args.itr_ori);
regidx_destroy(args.rules);
bcf_sr_destroy(args.sr);
if ( hts_close(args.out_fh)!=0 ) error("Error: close failed\n");
return 0;
}
static int mpileup(mplp_conf_t *conf)
{
if (conf->nfiles == 0) {
fprintf(stderr,"[%s] no input file/data given\n", __func__);
exit(EXIT_FAILURE);
}
mplp_ref_t mp_ref = MPLP_REF_INIT;
conf->gplp = (mplp_pileup_t *) calloc(1,sizeof(mplp_pileup_t));
conf->mplp_data = (mplp_aux_t**) calloc(conf->nfiles, sizeof(mplp_aux_t*));
conf->plp = (const bam_pileup1_t**) calloc(conf->nfiles, sizeof(bam_pileup1_t*));
conf->n_plp = (int*) calloc(conf->nfiles, sizeof(int));
// Allow to run mpileup on multiple regions in one go. This comes at cost: the bai index
// must be kept in the memory for the whole time which can be a problem with many bams.
// Therefore if none or only one region is requested, we initialize the bam iterator as
// before and free the index. Only when multiple regions are queried, we keep the index.
int nregs = 0;
if ( conf->reg_fname )
{
if ( conf->reg_is_file )
{
conf->reg = regidx_init(conf->reg_fname,NULL,NULL,0,NULL);
if ( !conf->reg ) {
fprintf(stderr,"Could not parse the regions: %s\n", conf->reg_fname);
exit(EXIT_FAILURE);
}
}
else
{
conf->reg = regidx_init(NULL,regidx_parse_reg,NULL,sizeof(char*),NULL);
if ( regidx_insert_list(conf->reg,conf->reg_fname,',') !=0 ) {
fprintf(stderr,"Could not parse the regions: %s\n", conf->reg_fname);
exit(EXIT_FAILURE);
}
}
nregs = regidx_nregs(conf->reg);
conf->reg_itr = regitr_init(conf->reg);
regitr_loop(conf->reg_itr); // region iterator now positioned at the first region
}
// read the header of each file in the list and initialize data
// beware: mpileup has always assumed that tid's are consistent in the headers, add sanity check at least!
bam_hdr_t *hdr = NULL; // header of first file in input list
int i;
for (i = 0; i < conf->nfiles; ++i) {
bam_hdr_t *h_tmp;
conf->mplp_data[i] = (mplp_aux_t*) calloc(1, sizeof(mplp_aux_t));
conf->mplp_data[i]->fp = sam_open(conf->files[i], "rb");
if ( !conf->mplp_data[i]->fp )
{
fprintf(stderr, "[%s] failed to open %s: %s\n", __func__, conf->files[i], strerror(errno));
exit(EXIT_FAILURE);
}
if (hts_set_opt(conf->mplp_data[i]->fp, CRAM_OPT_DECODE_MD, 0)) {
fprintf(stderr, "Failed to set CRAM_OPT_DECODE_MD value\n");
exit(EXIT_FAILURE);
}
if (conf->fai_fname && hts_set_fai_filename(conf->mplp_data[i]->fp, conf->fai_fname) != 0) {
fprintf(stderr, "[%s] failed to process %s: %s\n",
__func__, conf->fai_fname, strerror(errno));
exit(EXIT_FAILURE);
}
conf->mplp_data[i]->conf = conf;
conf->mplp_data[i]->ref = &mp_ref;
h_tmp = sam_hdr_read(conf->mplp_data[i]->fp);
if ( !h_tmp ) {
fprintf(stderr,"[%s] fail to read the header of %s\n", __func__, conf->files[i]);
exit(EXIT_FAILURE);
}
conf->mplp_data[i]->h = i ? hdr : h_tmp; // for j==0, "h" has not been set yet
conf->mplp_data[i]->bam_id = bam_smpl_add_bam(conf->bsmpl,h_tmp->text,conf->files[i]);
if ( conf->mplp_data[i]->bam_id<0 )
{
// no usable readgroups in this bam, it can be skipped
sam_close(conf->mplp_data[i]->fp);
free(conf->mplp_data[i]);
bam_hdr_destroy(h_tmp);
free(conf->files[i]);
if ( i+1<conf->nfiles ) memmove(&conf->files[i],&conf->files[i+1],sizeof(*conf->files)*(conf->nfiles-i-1));
conf->nfiles--;
i--;
continue;
}
if (conf->reg) {
hts_idx_t *idx = sam_index_load(conf->mplp_data[i]->fp, conf->files[i]);
if (idx == NULL) {
fprintf(stderr, "[%s] fail to load index for %s\n", __func__, conf->files[i]);
exit(EXIT_FAILURE);
}
conf->buf.l = 0;
ksprintf(&conf->buf,"%s:%u-%u",conf->reg_itr->seq,conf->reg_itr->beg+1,conf->reg_itr->end+1);
conf->mplp_data[i]->iter = sam_itr_querys(idx, conf->mplp_data[i]->h, conf->buf.s);
if ( !conf->mplp_data[i]->iter )
{
conf->mplp_data[i]->iter = sam_itr_querys(idx, conf->mplp_data[i]->h, conf->reg_itr->seq);
if ( conf->mplp_data[i]->iter ) {
fprintf(stderr,"[E::%s] fail to parse region '%s'\n", __func__, conf->buf.s);
exit(EXIT_FAILURE);
}
fprintf(stderr,"[E::%s] the sequence \"%s\" not found: %s\n",__func__,conf->reg_itr->seq,conf->files[i]);
exit(EXIT_FAILURE);
}
if ( nregs==1 ) // no need to keep the index in memory
hts_idx_destroy(idx);
else
conf->mplp_data[i]->idx = idx;
}
if ( !hdr ) hdr = h_tmp; /* save the header of first file in list */
else {
// FIXME: check consistency between h and h_tmp
bam_hdr_destroy(h_tmp);
// we store only the first file's header; it's (alleged to be)
// compatible with the i-th file's target_name lookup needs
conf->mplp_data[i]->h = hdr;
}
}
// allocate data storage proportionate to number of samples being studied sm->n
bam_smpl_get_samples(conf->bsmpl, &conf->gplp->n);
conf->gplp->n_plp = (int*) calloc(conf->gplp->n, sizeof(int));
conf->gplp->m_plp = (int*) calloc(conf->gplp->n, sizeof(int));
conf->gplp->plp = (bam_pileup1_t**) calloc(conf->gplp->n, sizeof(bam_pileup1_t*));
fprintf(stderr, "[%s] %d samples in %d input files\n", __func__, conf->gplp->n, conf->nfiles);
// write the VCF header
conf->bcf_fp = hts_open(conf->output_fname?conf->output_fname:"-", hts_bcf_wmode(conf->output_type));
if (conf->bcf_fp == NULL) {
fprintf(stderr, "[%s] failed to write to %s: %s\n", __func__, conf->output_fname? conf->output_fname : "standard output", strerror(errno));
exit(EXIT_FAILURE);
}
if ( conf->n_threads ) hts_set_threads(conf->bcf_fp, conf->n_threads);
// BCF header creation
conf->bcf_hdr = bcf_hdr_init("w");
conf->buf.l = 0;
if (conf->record_cmd_line)
{
ksprintf(&conf->buf, "##bcftoolsVersion=%s+htslib-%s\n",bcftools_version(),hts_version());
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
conf->buf.l = 0;
ksprintf(&conf->buf, "##bcftoolsCommand=mpileup");
for (i=1; i<conf->argc; i++) ksprintf(&conf->buf, " %s", conf->argv[i]);
kputc('\n', &conf->buf);
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
}
if (conf->fai_fname)
{
conf->buf.l = 0;
ksprintf(&conf->buf, "##reference=file://%s\n", conf->fai_fname);
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
}
// Translate BAM @SQ tags to BCF ##contig tags
// todo: use/write new BAM header manipulation routines, fill also UR, M5
for (i=0; i<hdr->n_targets; i++)
{
conf->buf.l = 0;
ksprintf(&conf->buf, "##contig=<ID=%s,length=%d>", hdr->target_name[i], hdr->target_len[i]);
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
}
conf->buf.l = 0;
bcf_hdr_append(conf->bcf_hdr,"##ALT=<ID=*,Description=\"Represents allele(s) other than observed.\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=INDEL,Number=0,Type=Flag,Description=\"Indicates that the variant is an INDEL.\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=IDV,Number=1,Type=Integer,Description=\"Maximum number of reads supporting an indel\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=IMF,Number=1,Type=Float,Description=\"Maximum fraction of reads supporting an indel\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=DP,Number=1,Type=Integer,Description=\"Raw read depth\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=VDB,Number=1,Type=Float,Description=\"Variant Distance Bias for filtering splice-site artefacts in RNA-seq data (bigger is better)\",Version=\"3\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=RPB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Read Position Bias (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality Bias (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=BQB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Base Quality Bias (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQSB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality vs Strand Bias (bigger is better)\">");
#if CDF_MWU_TESTS
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=RPB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Read Position Bias [CDF] (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality Bias [CDF] (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=BQB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Base Quality Bias [CDF] (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQSB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality vs Strand Bias [CDF] (bigger is better)\">");
#endif
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=SGB,Number=1,Type=Float,Description=\"Segregation based metric.\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQ0F,Number=1,Type=Float,Description=\"Fraction of MQ0 reads (smaller is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=I16,Number=16,Type=Float,Description=\"Auxiliary tag used for calling, see description of bcf_callret1_t in bam2bcf.h\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=QS,Number=R,Type=Float,Description=\"Auxiliary tag used for calling\">");
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=PL,Number=G,Type=Integer,Description=\"List of Phred-scaled genotype likelihoods\">");
if ( conf->fmt_flag&B2B_FMT_DP )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DP,Number=1,Type=Integer,Description=\"Number of high-quality bases\">");
if ( conf->fmt_flag&B2B_FMT_DV )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DV,Number=1,Type=Integer,Description=\"Number of high-quality non-reference bases\">");
if ( conf->fmt_flag&B2B_FMT_DPR )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DPR,Number=R,Type=Integer,Description=\"Number of high-quality bases observed for each allele\">");
if ( conf->fmt_flag&B2B_INFO_DPR )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=DPR,Number=R,Type=Integer,Description=\"Number of high-quality bases observed for each allele\">");
if ( conf->fmt_flag&B2B_FMT_DP4 )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DP4,Number=4,Type=Integer,Description=\"Number of high-quality ref-fwd, ref-reverse, alt-fwd and alt-reverse bases\">");
if ( conf->fmt_flag&B2B_FMT_SP )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=SP,Number=1,Type=Integer,Description=\"Phred-scaled strand bias P-value\">");
if ( conf->fmt_flag&B2B_FMT_AD )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=AD,Number=R,Type=Integer,Description=\"Allelic depths\">");
if ( conf->fmt_flag&B2B_FMT_ADF )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=ADF,Number=R,Type=Integer,Description=\"Allelic depths on the forward strand\">");
if ( conf->fmt_flag&B2B_FMT_ADR )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=ADR,Number=R,Type=Integer,Description=\"Allelic depths on the reverse strand\">");
if ( conf->fmt_flag&B2B_INFO_AD )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=AD,Number=R,Type=Integer,Description=\"Total allelic depths\">");
if ( conf->fmt_flag&B2B_INFO_ADF )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=ADF,Number=R,Type=Integer,Description=\"Total allelic depths on the forward strand\">");
if ( conf->fmt_flag&B2B_INFO_ADR )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=ADR,Number=R,Type=Integer,Description=\"Total allelic depths on the reverse strand\">");
if ( conf->gvcf )
gvcf_update_header(conf->gvcf, conf->bcf_hdr);
int nsmpl;
const char **smpl = bam_smpl_get_samples(conf->bsmpl, &nsmpl);
for (i=0; i<nsmpl; i++)
bcf_hdr_add_sample(conf->bcf_hdr, smpl[i]);
bcf_hdr_write(conf->bcf_fp, conf->bcf_hdr);
conf->bca = bcf_call_init(-1., conf->min_baseQ);
conf->bcr = (bcf_callret1_t*) calloc(nsmpl, sizeof(bcf_callret1_t));
conf->bca->openQ = conf->openQ, conf->bca->extQ = conf->extQ, conf->bca->tandemQ = conf->tandemQ;
conf->bca->min_frac = conf->min_frac;
conf->bca->min_support = conf->min_support;
conf->bca->per_sample_flt = conf->flag & MPLP_PER_SAMPLE;
conf->bc.bcf_hdr = conf->bcf_hdr;
conf->bc.n = nsmpl;
conf->bc.PL = (int32_t*) malloc(15 * nsmpl * sizeof(*conf->bc.PL));
if (conf->fmt_flag)
{
assert( sizeof(float)==sizeof(int32_t) );
conf->bc.DP4 = (int32_t*) malloc(nsmpl * sizeof(int32_t) * 4);
conf->bc.fmt_arr = (uint8_t*) malloc(nsmpl * sizeof(float)); // all fmt_flag fields, float and int32
if ( conf->fmt_flag&(B2B_INFO_DPR|B2B_FMT_DPR|B2B_INFO_AD|B2B_INFO_ADF|B2B_INFO_ADR|B2B_FMT_AD|B2B_FMT_ADF|B2B_FMT_ADR) )
{
// first B2B_MAX_ALLELES fields for total numbers, the rest per-sample
conf->bc.ADR = (int32_t*) malloc((nsmpl+1)*B2B_MAX_ALLELES*sizeof(int32_t));
conf->bc.ADF = (int32_t*) malloc((nsmpl+1)*B2B_MAX_ALLELES*sizeof(int32_t));
for (i=0; i<nsmpl; i++)
{
conf->bcr[i].ADR = conf->bc.ADR + (i+1)*B2B_MAX_ALLELES;
conf->bcr[i].ADF = conf->bc.ADF + (i+1)*B2B_MAX_ALLELES;
}
}
}
// init mpileup
conf->iter = bam_mplp_init(conf->nfiles, mplp_func, (void**)conf->mplp_data);
if ( conf->flag & MPLP_SMART_OVERLAPS ) bam_mplp_init_overlaps(conf->iter);
if ( (double)conf->max_depth * conf->nfiles > 1<<20)
fprintf(stderr, "Warning: Potential memory hog, up to %.0fM reads in the pileup!\n", (double)conf->max_depth*conf->nfiles);
if ( (double)conf->max_depth * conf->nfiles / nsmpl < 250 )
fprintf(stderr, "Note: The maximum per-sample depth with -d %d is %.1fx\n", conf->max_depth,(double)conf->max_depth * conf->nfiles / nsmpl);
bam_mplp_set_maxcnt(conf->iter, conf->max_depth);
conf->max_indel_depth = conf->max_indel_depth * nsmpl;
conf->bcf_rec = bcf_init1();
bam_mplp_constructor(conf->iter, pileup_constructor);
// Run mpileup for multiple regions
if ( nregs )
{
int ireg = 0;
do
{
// first region is already positioned
if ( ireg++ > 0 )
{
conf->buf.l = 0;
ksprintf(&conf->buf,"%s:%u-%u",conf->reg_itr->seq,conf->reg_itr->beg,conf->reg_itr->end);
for (i=0; i<conf->nfiles; i++)
{
hts_itr_destroy(conf->mplp_data[i]->iter);
conf->mplp_data[i]->iter = sam_itr_querys(conf->mplp_data[i]->idx, conf->mplp_data[i]->h, conf->buf.s);
if ( !conf->mplp_data[i]->iter )
{
conf->mplp_data[i]->iter = sam_itr_querys(conf->mplp_data[i]->idx, conf->mplp_data[i]->h, conf->reg_itr->seq);
if ( conf->mplp_data[i]->iter ) {
fprintf(stderr,"[E::%s] fail to parse region '%s'\n", __func__, conf->buf.s);
exit(EXIT_FAILURE);
}
fprintf(stderr,"[E::%s] the sequence \"%s\" not found: %s\n",__func__,conf->reg_itr->seq,conf->files[i]);
exit(EXIT_FAILURE);
}
bam_mplp_reset(conf->iter);
}
}
mpileup_reg(conf,conf->reg_itr->beg,conf->reg_itr->end);
}
while ( regitr_loop(conf->reg_itr) );
}
else
mpileup_reg(conf,0,0);
flush_bcf_records(conf, conf->bcf_fp, conf->bcf_hdr, NULL);
// clean up
free(conf->bc.tmp.s);
bcf_destroy1(conf->bcf_rec);
if (conf->bcf_fp)
{
hts_close(conf->bcf_fp);
bcf_hdr_destroy(conf->bcf_hdr);
bcf_call_destroy(conf->bca);
free(conf->bc.PL);
free(conf->bc.DP4);
free(conf->bc.ADR);
free(conf->bc.ADF);
free(conf->bc.fmt_arr);
free(conf->bcr);
}
if ( conf->gvcf ) gvcf_destroy(conf->gvcf);
free(conf->buf.s);
for (i = 0; i < conf->gplp->n; ++i) free(conf->gplp->plp[i]);
free(conf->gplp->plp); free(conf->gplp->n_plp); free(conf->gplp->m_plp); free(conf->gplp);
bam_mplp_destroy(conf->iter);
bam_hdr_destroy(hdr);
for (i = 0; i < conf->nfiles; ++i) {
if ( nregs>1 ) hts_idx_destroy(conf->mplp_data[i]->idx);
sam_close(conf->mplp_data[i]->fp);
if ( conf->mplp_data[i]->iter) hts_itr_destroy(conf->mplp_data[i]->iter);
free(conf->mplp_data[i]);
}
if ( conf->reg_itr ) regitr_destroy(conf->reg_itr);
free(conf->mplp_data); free(conf->plp); free(conf->n_plp);
free(mp_ref.ref[0]);
free(mp_ref.ref[1]);
return 0;
}
static int query_regions(args_t *args, char *fname, char **regs, int nregs)
{
int i;
htsFile *fp = hts_open(fname,"r");
if ( !fp ) error("Could not read %s\n", fname);
enum htsExactFormat format = hts_get_format(fp)->format;
regidx_t *reg_idx = NULL;
if ( args->targets_fname )
{
reg_idx = regidx_init(args->targets_fname, NULL, NULL, 0, NULL);
if ( !reg_idx ) error("Could not read %s\n", args->targets_fname);
}
if ( format == bcf )
{
htsFile *out = hts_open("-","w");
if ( !out ) error("Could not open stdout\n", fname);
hts_idx_t *idx = bcf_index_load(fname);
if ( !idx ) error("Could not load .csi index of %s\n", fname);
bcf_hdr_t *hdr = bcf_hdr_read(fp);
if ( !hdr ) error("Could not read the header: %s\n", fname);
if ( args->print_header )
bcf_hdr_write(out,hdr);
if ( !args->header_only )
{
bcf1_t *rec = bcf_init();
for (i=0; i<nregs; i++)
{
hts_itr_t *itr = bcf_itr_querys(idx,hdr,regs[i]);
while ( bcf_itr_next(fp, itr, rec) >=0 )
{
if ( reg_idx && !regidx_overlap(reg_idx, bcf_seqname(hdr,rec),rec->pos,rec->pos+rec->rlen-1, NULL) ) continue;
bcf_write(out,hdr,rec);
}
tbx_itr_destroy(itr);
}
bcf_destroy(rec);
}
if ( hts_close(out) ) error("hts_close returned non-zero status for stdout\n");
bcf_hdr_destroy(hdr);
hts_idx_destroy(idx);
}
else if ( format==vcf || format==sam || format==unknown_format )
{
tbx_t *tbx = tbx_index_load(fname);
if ( !tbx ) error("Could not load .tbi/.csi index of %s\n", fname);
kstring_t str = {0,0,0};
if ( args->print_header )
{
while ( hts_getline(fp, KS_SEP_LINE, &str) >= 0 )
{
if ( !str.l || str.s[0]!=tbx->conf.meta_char ) break;
puts(str.s);
}
}
if ( !args->header_only )
{
int nseq;
const char **seq = NULL;
if ( reg_idx ) seq = tbx_seqnames(tbx, &nseq);
for (i=0; i<nregs; i++)
{
hts_itr_t *itr = tbx_itr_querys(tbx, regs[i]);
if ( !itr ) continue;
while (tbx_itr_next(fp, tbx, itr, &str) >= 0)
{
if ( reg_idx && !regidx_overlap(reg_idx,seq[itr->curr_tid],itr->curr_beg,itr->curr_end, NULL) ) continue;
puts(str.s);
}
tbx_itr_destroy(itr);
}
free(seq);
}
free(str.s);
tbx_destroy(tbx);
}
else if ( format==bam )
error("Please use \"samtools view\" for querying BAM files.\n");
if ( reg_idx ) regidx_destroy(reg_idx);
if ( hts_close(fp) ) error("hts_close returned non-zero status: %s\n", fname);
for (i=0; i<nregs; i++) free(regs[i]);
free(regs);
return 0;
}
