static void init_data(args_t *args)
{
args->hdr = args->files->readers[0].header;
args->hdr_out = bcf_hdr_dup(args->hdr);
init_plugin(args);
if ( args->filter_str )
args->filter = filter_init(args->hdr, args->filter_str);
bcf_hdr_append_version(args->hdr_out, args->argc, args->argv, "bcftools_plugin");
if ( !args->drop_header )
{
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_out);
}
}
static void init_data(args_t *args)
{
args->aux.srs = bcf_sr_init();
// Open files for input and output, initialize structures
if ( args->targets )
{
if ( bcf_sr_set_targets(args->aux.srs, args->targets, args->targets_is_file, args->aux.flag&CALL_CONSTR_ALLELES ? 3 : 0)<0 )
error("Failed to read the targets: %s\n", args->targets);
if ( args->aux.flag&CALL_CONSTR_ALLELES && args->flag&CF_INS_MISSED )
{
args->aux.srs->targets->missed_reg_handler = print_missed_line;
args->aux.srs->targets->missed_reg_data = args;
}
}
if ( args->regions )
{
if ( bcf_sr_set_regions(args->aux.srs, args->regions, args->regions_is_file)<0 )
error("Failed to read the targets: %s\n", args->regions);
}
int i;
if ( !bcf_sr_add_reader(args->aux.srs, args->bcf_fname) ) error("Failed to open: %s\n", args->bcf_fname);
if ( args->nsamples && args->nsamples != bcf_hdr_nsamples(args->aux.srs->readers[0].header) )
{
args->samples_map = (int *) malloc(sizeof(int)*args->nsamples);
args->aux.hdr = bcf_hdr_subset(args->aux.srs->readers[0].header, args->nsamples, args->samples, args->samples_map);
for (i=0; i<args->nsamples; i++)
if ( args->samples_map[i]<0 ) error("No such sample: %s\n", args->samples[i]);
if ( !bcf_hdr_nsamples(args->aux.hdr) ) error("No matching sample found\n");
}
else
{
args->aux.hdr = bcf_hdr_dup(args->aux.srs->readers[0].header);
for (i=0; i<args->nsamples; i++)
if ( bcf_hdr_id2int(args->aux.hdr,BCF_DT_SAMPLE,args->samples[i])<0 )
error("No such sample: %s\n", args->samples[i]);
}
// Reorder ploidy and family indexes to match mpileup's output and exclude samples which are not available
if ( args->aux.ploidy )
{
for (i=0; i<args->aux.nfams; i++)
{
int j;
for (j=0; j<3; j++)
{
int k = bcf_hdr_id2int(args->aux.hdr, BCF_DT_SAMPLE, args->samples[ args->aux.fams[i].sample[j] ]);
if ( k<0 ) error("No such sample: %s\n", args->samples[ args->aux.fams[i].sample[j] ]);
args->aux.fams[i].sample[j] = k;
}
}
uint8_t *ploidy = (uint8_t*) calloc(bcf_hdr_nsamples(args->aux.hdr), 1);
for (i=0; i<args->nsamples; i++) // i index in -s sample list
{
int j = bcf_hdr_id2int(args->aux.hdr, BCF_DT_SAMPLE, args->samples[i]); // j index in the output VCF / subset VCF
if ( j<0 )
{
fprintf(stderr,"Warning: no such sample: \"%s\"\n", args->samples[i]);
continue;
}
ploidy[j] = args->aux.ploidy[i];
}
args->nsamples = bcf_hdr_nsamples(args->aux.hdr);
for (i=0; i<args->nsamples; i++)
assert( ploidy[i]==0 || ploidy[i]==1 || ploidy[i]==2 );
free(args->aux.ploidy);
args->aux.ploidy = ploidy;
}
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));
if ( args->flag & CF_QCALL )
return;
if ( args->flag & CF_MCALL )
mcall_init(&args->aux);
if ( args->flag & CF_CCALL )
ccall_init(&args->aux);
if ( args->flag&CF_GVCF )
{
bcf_hdr_append(args->aux.hdr,"##INFO=<ID=END,Number=1,Type=Integer,Description=\"End position of the variant described in this record\">");
args->gvcf.rid = -1;
args->gvcf.line = bcf_init1();
args->gvcf.gt = (int32_t*) malloc(2*sizeof(int32_t)*bcf_hdr_nsamples(args->aux.hdr));
for (i=0; i<bcf_hdr_nsamples(args->aux.hdr); i++)
{
args->gvcf.gt[2*i+0] = bcf_gt_unphased(0);
args->gvcf.gt[2*i+1] = bcf_gt_unphased(0);
}
}
bcf_hdr_remove(args->aux.hdr, BCF_HL_INFO, "QS");
bcf_hdr_remove(args->aux.hdr, BCF_HL_INFO, "I16");
bcf_hdr_append_version(args->aux.hdr, args->argc, args->argv, "bcftools_call");
bcf_hdr_write(args->out_fh, args->aux.hdr);
if ( args->flag&CF_INS_MISSED ) init_missed_line(args);
}
void isec_vcf(args_t *args)
{
bcf_srs_t *files = args->files;
kstring_t str = {0,0,0};
htsFile *out_fh = NULL;
// When only one VCF is output, print VCF to pysam_stdout or -o file
int out_std = 0;
if ( args->nwrite==1 && !args->prefix ) out_std = 1;
if ( args->targets_list && files->nreaders==1 ) out_std = 1;
if ( out_std )
{
out_fh = hts_open(args->output_fname? args->output_fname : "-",hts_bcf_wmode(args->output_type));
if ( out_fh == NULL ) error("Can't write to %s: %s\n", args->output_fname? args->output_fname : "standard output", strerror(errno));
if ( args->n_threads ) hts_set_threads(out_fh, args->n_threads);
if (args->record_cmd_line) bcf_hdr_append_version(files->readers[args->iwrite].header,args->argc,args->argv,"bcftools_isec");
bcf_hdr_write(out_fh, files->readers[args->iwrite].header);
}
if ( !args->nwrite && !out_std && !args->prefix )
fprintf(pysam_stderr,"Note: -w option not given, printing list of sites...\n");
int n;
while ( (n=bcf_sr_next_line(files)) )
{
bcf_sr_t *reader = NULL;
bcf1_t *line = NULL;
int i, ret = 0;
for (i=0; i<files->nreaders; i++)
{
if ( !bcf_sr_has_line(files,i) ) continue;
if ( args->nflt && args->flt[i] )
{
bcf1_t *rec = bcf_sr_get_line(files, i);
int pass = filter_test(args->flt[i], rec, NULL);
if ( args->flt_logic[i] & FLT_EXCLUDE ) pass = pass ? 0 : 1;
if ( !pass )
{
files->has_line[i] = 0;
n--;
continue;
}
}
if ( !line )
{
line = files->readers[i].buffer[0];
reader = &files->readers[i];
}
ret |= 1<<i; // this may overflow for many files, but will be used only with two (OP_VENN)
}
switch (args->isec_op)
{
case OP_COMPLEMENT: if ( n!=1 || !bcf_sr_has_line(files,0) ) continue; break;
case OP_EQUAL: if ( n != args->isec_n ) continue; break;
case OP_PLUS: if ( n < args->isec_n ) continue; break;
case OP_MINUS: if ( n > args->isec_n ) continue; break;
case OP_EXACT:
for (i=0; i<files->nreaders; i++)
if ( files->has_line[i] != args->isec_exact[i] ) break;
if ( i<files->nreaders ) continue;
break;
}
if ( out_std )
{
if ( bcf_sr_has_line(files,args->iwrite) )
bcf_write1(out_fh, files->readers[args->iwrite].header, files->readers[args->iwrite].buffer[0]);
continue;
}
else if ( args->fh_sites )
{
str.l = 0;
kputs(reader->header->id[BCF_DT_CTG][line->rid].key, &str); kputc('\t', &str);
kputw(line->pos+1, &str); kputc('\t', &str);
if (line->n_allele > 0) kputs(line->d.allele[0], &str);
else kputc('.', &str);
kputc('\t', &str);
if (line->n_allele > 1) kputs(line->d.allele[1], &str);
else kputc('.', &str);
for (i=2; i<line->n_allele; i++)
{
kputc(',', &str);
kputs(line->d.allele[i], &str);
}
kputc('\t', &str);
for (i=0; i<files->nreaders; i++)
kputc(bcf_sr_has_line(files,i)?'1':'0', &str);
kputc('\n', &str);
fwrite(str.s,sizeof(char),str.l,args->fh_sites);
}
if ( args->prefix )
{
if ( args->isec_op==OP_VENN && ret==3 )
{
if ( !args->nwrite || args->write[0] )
bcf_write1(args->fh_out[2], bcf_sr_get_header(files,0), bcf_sr_get_line(files,0));
if ( !args->nwrite || args->write[1] )
bcf_write1(args->fh_out[3], bcf_sr_get_header(files,1), bcf_sr_get_line(files,1));
}
else
{
for (i=0; i<files->nreaders; i++)
{
if ( !bcf_sr_has_line(files,i) ) continue;
if ( args->write && !args->write[i] ) continue;
bcf_write1(args->fh_out[i], files->readers[i].header, files->readers[i].buffer[0]);
}
}
}
}
if ( str.s ) free(str.s);
if ( out_fh ) hts_close(out_fh);
}
static void init_data(args_t *args)
{
bcf1_t *line = NULL;
// With phased concat, the chunks overlap and come in the right order. To
// avoid opening all files at once, store start positions to recognise need
// for the next one. This way we can keep only two open chunks at once.
if ( args->phased_concat )
{
args->start_pos = (int*) malloc(sizeof(int)*args->nfnames);
line = bcf_init();
}
kstring_t str = {0,0,0};
int i, prev_chrid = -1;
for (i=0; i<args->nfnames; i++)
{
htsFile *fp = hts_open(args->fnames[i], "r"); if ( !fp ) error("Failed to open: %s\n", args->fnames[i]);
bcf_hdr_t *hdr = bcf_hdr_read(fp); if ( !hdr ) error("Failed to parse header: %s\n", args->fnames[i]);
args->out_hdr = bcf_hdr_merge(args->out_hdr,hdr);
if ( bcf_hdr_nsamples(hdr) != bcf_hdr_nsamples(args->out_hdr) )
error("Different number of samples in %s. Perhaps \"bcftools merge\" is what you are looking for?\n", args->fnames[i]);
int j;
for (j=0; j<bcf_hdr_nsamples(hdr); j++)
if ( strcmp(args->out_hdr->samples[j],hdr->samples[j]) )
error("Different sample names in %s. Perhaps \"bcftools merge\" is what you are looking for?\n", args->fnames[i]);
if ( args->phased_concat )
{
int ret = bcf_read(fp, hdr, line);
if ( ret!=0 ) args->start_pos[i] = -2; // empty file
else
{
int chrid = bcf_hdr_id2int(args->out_hdr,BCF_DT_CTG,bcf_seqname(hdr,line));
args->start_pos[i] = chrid==prev_chrid ? line->pos : -1;
prev_chrid = chrid;
}
}
bcf_hdr_destroy(hdr);
hts_close(fp);
}
free(str.s);
if ( line ) bcf_destroy(line);
args->seen_seq = (int*) calloc(args->out_hdr->n[BCF_DT_CTG],sizeof(int));
if ( args->phased_concat )
{
bcf_hdr_append(args->out_hdr,"##FORMAT=<ID=PQ,Number=1,Type=Integer,Description=\"Phasing Quality (bigger is better)\">");
bcf_hdr_append(args->out_hdr,"##FORMAT=<ID=PS,Number=1,Type=Integer,Description=\"Phase Set\">");
}
if (args->record_cmd_line) bcf_hdr_append_version(args->out_hdr, args->argc, args->argv, "bcftools_concat");
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));
if ( args->n_threads ) hts_set_threads(args->out_fh, args->n_threads);
bcf_hdr_write(args->out_fh, args->out_hdr);
if ( args->allow_overlaps )
{
args->files = bcf_sr_init();
args->files->require_index = 1;
if ( args->regions_list )
{
if ( bcf_sr_set_regions(args->files, args->regions_list, args->regions_is_file)<0 )
error("Failed to read the regions: %s\n", args->regions_list);
}
if ( args->remove_dups )
{
if ( !strcmp(args->remove_dups,"snps") ) args->files->collapse |= COLLAPSE_SNPS;
else if ( !strcmp(args->remove_dups,"indels") ) args->files->collapse |= COLLAPSE_INDELS;
else if ( !strcmp(args->remove_dups,"both") ) args->files->collapse |= COLLAPSE_SNPS | COLLAPSE_INDELS;
else if ( !strcmp(args->remove_dups,"any") ) args->files->collapse |= COLLAPSE_ANY;
else if ( !strcmp(args->remove_dups,"all") ) args->files->collapse |= COLLAPSE_ANY;
else if ( !strcmp(args->remove_dups,"none") ) args->files->collapse = COLLAPSE_NONE;
else error("The -D string \"%s\" not recognised.\n", args->remove_dups);
}
for (i=0; i<args->nfnames; i++)
if ( !bcf_sr_add_reader(args->files,args->fnames[i]) ) error("Failed to open %s: %s\n", args->fnames[i],bcf_sr_strerror(args->files->errnum));
}
else if ( args->phased_concat )
{
// Remove empty files from the list
int nok = 0;
while (1)
{
while ( nok<args->nfnames && args->start_pos[nok]!=-2 ) nok++;
if ( nok==args->nfnames ) break;
i = nok;
while ( i<args->nfnames && args->start_pos[i]==-2 ) i++;
if ( i==args->nfnames ) break;
int tmp = args->start_pos[nok]; args->start_pos[nok] = args->start_pos[i]; args->start_pos[i] = tmp;
char *str = args->fnames[nok]; args->fnames[nok] = args->fnames[i]; args->fnames[i] = str;
}
for (i=nok; i<args->nfnames; i++) free(args->fnames[i]);
args->nfnames = nok;
for (i=1; i<args->nfnames; i++)
if ( args->start_pos[i-1]!=-1 && args->start_pos[i]!=-1 && args->start_pos[i]<args->start_pos[i-1] )
error("The files not in ascending order: %d in %s, %d in %s\n", args->start_pos[i-1]+1,args->fnames[i-1],args->start_pos[i]+1,args->fnames[i]);
args->prev_chr = -1;
args->swap_phase = (int*) calloc(bcf_hdr_nsamples(args->out_hdr),sizeof(int));
args->nmatch = (int*) calloc(bcf_hdr_nsamples(args->out_hdr),sizeof(int));
args->nmism = (int*) calloc(bcf_hdr_nsamples(args->out_hdr),sizeof(int));
args->phase_qual = (int32_t*) malloc(bcf_hdr_nsamples(args->out_hdr)*sizeof(int32_t));
args->phase_set = (int32_t*) malloc(bcf_hdr_nsamples(args->out_hdr)*sizeof(int32_t));
args->files = bcf_sr_init();
args->files->require_index = 1;
args->ifname = 0;
}
}
static void init_data(args_t *args)
{
args->sr = bcf_sr_init();
if ( args->region )
{
args->sr->require_index = 1;
if ( bcf_sr_set_regions(args->sr, args->region, args->region_is_file)<0 ) error("Failed to read the regions: %s\n",args->region);
}
if ( args->target && bcf_sr_set_targets(args->sr, args->target, args->target_is_file, 0)<0 ) error("Failed to read the targets: %s\n",args->target);
if ( !bcf_sr_add_reader(args->sr,args->fname) ) error("Error: %s\n", bcf_sr_strerror(args->sr->errnum));
args->hdr_in = bcf_sr_get_header(args->sr,0);
args->hdr_out = bcf_hdr_dup(args->hdr_in);
if ( args->filter_str )
args->filter = filter_init(args->hdr_in, args->filter_str);
mkdir_p("%s/",args->output_dir);
int i, nsmpl = bcf_hdr_nsamples(args->hdr_in);
if ( !nsmpl ) error("No samples to split: %s\n", args->fname);
args->fh = (htsFile**)calloc(nsmpl,sizeof(*args->fh));
args->bnames = set_file_base_names(args);
kstring_t str = {0,0,0};
for (i=0; i<nsmpl; i++)
{
if ( !args->bnames[i] ) continue;
str.l = 0;
kputs(args->output_dir, &str);
if ( str.s[str.l-1] != '/' ) kputc('/', &str);
int k, l = str.l;
kputs(args->bnames[i], &str);
for (k=l; k<str.l; k++) if ( isspace(str.s[k]) ) str.s[k] = '_';
if ( args->output_type & FT_BCF ) kputs(".bcf", &str);
else if ( args->output_type & FT_GZ ) kputs(".vcf.gz", &str);
else kputs(".vcf", &str);
args->fh[i] = hts_open(str.s, hts_bcf_wmode(args->output_type));
if ( args->fh[i] == NULL ) error("Can't write to \"%s\": %s\n", str.s, strerror(errno));
bcf_hdr_nsamples(args->hdr_out) = 1;
args->hdr_out->samples[0] = args->bnames[i];
bcf_hdr_write(args->fh[i], args->hdr_out);
}
free(str.s);
// parse tags
int is_info = 0, is_fmt = 0;
char *beg = args->keep_tags;
while ( beg && *beg )
{
if ( !strncasecmp("INFO/",beg,5) ) { is_info = 1; is_fmt = 0; beg += 5; }
else if ( !strcasecmp("INFO",beg) ) { args->keep_info = 1; break; }
else if ( !strncasecmp("INFO,",beg,5) ) { args->keep_info = 1; beg += 5; continue; }
else if ( !strncasecmp("FMT/",beg,4) ) { is_info = 0; is_fmt = 1; beg += 4; }
else if ( !strncasecmp("FORMAT/",beg,7) ) { is_info = 0; is_fmt = 1; beg += 7; }
else if ( !strcasecmp("FMT",beg) ) { args->keep_fmt = 1; break; }
else if ( !strcasecmp("FORMAT",beg) ) { args->keep_fmt = 1; break; }
else if ( !strncasecmp("FMT,",beg,4) ) { args->keep_fmt = 1; beg += 4; continue; }
else if ( !strncasecmp("FORMAT,",beg,7) ) { args->keep_fmt = 1; beg += 7; continue; }
char *end = beg;
while ( *end && *end!=',' ) end++;
char tmp = *end; *end = 0;
int id = bcf_hdr_id2int(args->hdr_in, BCF_DT_ID, beg);
beg = tmp ? end + 1 : end;
if ( is_info && bcf_hdr_idinfo_exists(args->hdr_in,BCF_HL_INFO,id) )
{
if ( id >= args->ninfo_tags ) args->ninfo_tags = id + 1;
hts_expand0(uint8_t, args->ninfo_tags, args->minfo_tags, args->info_tags);
args->info_tags[id] = 1;
}
if ( is_fmt && bcf_hdr_idinfo_exists(args->hdr_in,BCF_HL_FMT,id) )
{
if ( id >= args->nfmt_tags ) args->nfmt_tags = id + 1;
hts_expand0(uint8_t, args->nfmt_tags, args->mfmt_tags, args->fmt_tags);
args->fmt_tags[id] = 1;
}
}
if ( !args->keep_info && !args->keep_fmt && !args->ninfo_tags && !args->nfmt_tags )
{
args->keep_info = args->keep_fmt = 1;
}
}
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;
}
void isec_vcf(args_t *args)
{
bcf_srs_t *files = args->files;
kstring_t str = {0,0,0};
htsFile *out_fh = NULL;
// When only one VCF is output, print VCF to stdout
int out_std = 0;
if ( args->nwrite==1 ) out_std = 1;
if ( args->targets_list && files->nreaders==1 ) out_std = 1;
if ( out_std )
{
out_fh = hts_open("-",hts_bcf_wmode(args->output_type));
bcf_hdr_append_version(files->readers[args->iwrite].header,args->argc,args->argv,"bcftools_isec");
bcf_hdr_write(out_fh, files->readers[args->iwrite].header);
}
if ( !args->nwrite && !out_std && !args->prefix )
fprintf(stderr,"Note: -w option not given, printing list of sites...\n");
int n;
while ( (n=bcf_sr_next_line(files)) )
{
bcf_sr_t *reader = NULL;
bcf1_t *line = NULL;
int i, ret = 0;
for (i=0; i<files->nreaders; i++)
{
if ( !bcf_sr_has_line(files,i) ) continue;
if ( !line )
{
line = files->readers[i].buffer[0];
reader = &files->readers[i];
}
ret |= 1<<i; // this may overflow for many files, but will be used only with two (OP_VENN)
}
switch (args->isec_op)
{
case OP_COMPLEMENT: if ( n!=1 || !bcf_sr_has_line(files,0) ) continue; break;
case OP_EQUAL: if ( n != args->isec_n ) continue; break;
case OP_PLUS: if ( n < args->isec_n ) continue; break;
case OP_MINUS: if ( n > args->isec_n ) continue;
}
if ( out_std )
{
if ( bcf_sr_has_line(files,args->iwrite) )
bcf_write1(out_fh, files->readers[args->iwrite].header, files->readers[args->iwrite].buffer[0]);
continue;
}
else if ( args->fh_sites )
{
str.l = 0;
kputs(reader->header->id[BCF_DT_CTG][line->rid].key, &str); kputc('\t', &str);
kputw(line->pos+1, &str); kputc('\t', &str);
if (line->n_allele > 0) kputs(line->d.allele[0], &str);
else kputc('.', &str);
kputc('\t', &str);
if (line->n_allele > 1) kputs(line->d.allele[1], &str);
else kputc('.', &str);
for (i=2; i<line->n_allele; i++)
{
kputc(',', &str);
kputs(line->d.allele[i], &str);
}
kputc('\t', &str);
for (i=0; i<files->nreaders; i++)
kputc(bcf_sr_has_line(files,i)?'1':'0', &str);
kputc('\n', &str);
fwrite(str.s,sizeof(char),str.l,args->fh_sites);
}
if ( args->prefix )
{
if ( args->isec_op==OP_VENN )
bcf_write1(args->fh_out[ret-1], reader->header, line);
else
{
for (i=0; i<files->nreaders; i++)
{
if ( !bcf_sr_has_line(files,i) ) continue;
if ( args->write && !args->write[i] ) continue;
bcf_write1(args->fh_out[i], files->readers[i].header, files->readers[i].buffer[0]);
}
}
}
}
if ( str.s ) free(str.s);
if ( out_fh ) hts_close(out_fh);
}
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 void init_data(args_t *args)
{
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));
if ( args->n_threads ) hts_set_threads(args->out_fh, args->n_threads);
args->hdr = args->files->readers[0].header;
args->flt_pass = bcf_hdr_id2int(args->hdr,BCF_DT_ID,"PASS"); assert( !args->flt_pass ); // sanity check: required by BCF spec
// -i or -e: append FILTER line
if ( args->soft_filter && args->filter_logic )
{
kstring_t flt_name = {0,0,0};
if ( strcmp(args->soft_filter,"+") )
kputs(args->soft_filter, &flt_name);
else
{
// Make up a filter name
int i = 0, id = -1;
do
{
ksprintf(&flt_name,"Filter%d", ++i);
id = bcf_hdr_id2int(args->hdr,BCF_DT_ID,flt_name.s);
}
while ( bcf_hdr_idinfo_exists(args->hdr,BCF_HL_FLT,id) );
}
// escape quotes
kstring_t tmp = {0,0,0};
char *t = args->filter_str;
while ( *t )
{
if ( *t=='"' ) kputc('\\',&tmp);
kputc(*t,&tmp);
t++;
}
int ret = bcf_hdr_printf(args->hdr, "##FILTER=<ID=%s,Description=\"Set if %s: %s\">", flt_name.s,args->filter_logic & FLT_INCLUDE ? "not true" : "true", tmp.s);
if ( ret!=0 )
error("Failed to append header line: ##FILTER=<ID=%s,Description=\"Set if %s: %s\">\n", flt_name.s,args->filter_logic & FLT_INCLUDE ? "not true" : "true", tmp.s);
args->flt_fail = bcf_hdr_id2int(args->hdr,BCF_DT_ID,flt_name.s); assert( args->flt_fail>=0 );
free(flt_name.s);
free(tmp.s);
}
if ( args->snp_gap || args->indel_gap )
{
if ( !args->filter_logic && args->soft_filter && strcmp(args->soft_filter,"+") )
{
kstring_t tmp = {0,0,0};
if ( args->snp_gap ) kputs("\"SnpGap\"", &tmp);
if ( args->indel_gap )
{
if ( tmp.s ) kputs(" and ", &tmp);
kputs("\"IndelGap\"", &tmp);
}
fprintf(stderr,"Warning: using %s filter name instead of \"%s\"\n", tmp.s,args->soft_filter);
free(tmp.s);
}
rbuf_init(&args->rbuf, 64);
args->rbuf_lines = (bcf1_t**) calloc(args->rbuf.m, sizeof(bcf1_t*));
if ( args->snp_gap )
{
bcf_hdr_printf(args->hdr, "##FILTER=<ID=SnpGap,Description=\"SNP within %d bp of an indel\">", args->snp_gap);
args->SnpGap_id = bcf_hdr_id2int(args->hdr, BCF_DT_ID, "SnpGap");
assert( args->SnpGap_id>=0 );
}
if ( args->indel_gap )
{
bcf_hdr_printf(args->hdr, "##FILTER=<ID=IndelGap,Description=\"Indel within %d bp of an indel\">", args->indel_gap);
args->IndelGap_id = bcf_hdr_id2int(args->hdr, BCF_DT_ID, "IndelGap");
assert( args->IndelGap_id>=0 );
}
}
if (args->record_cmd_line) bcf_hdr_append_version(args->hdr, args->argc, args->argv, "bcftools_filter");
if ( args->filter_str )
args->filter = filter_init(args->hdr, args->filter_str);
}
int main(int argc, char **argv)
{
if ( argc == 1 )
error("Usage : bed_annos -c config.json -O z -o output.vcf.gz input.vcf.gz");
int i;
for ( i = 1; i < argc; ) {
const char *a = argv[i++];
const char **var = 0;
if ( strcmp(a, "-c") == 0 )
var = &json_fname;
else if ( strcmp(a, "-O") == 0 )
var = &output_fname_type;
else if ( strcmp(a, "-o") == 0 )
var = &output_fname;
if ( var != 0 ) {
if ( i == argc )
error("Missing an argument after %s", a);
*var = argv[i++];
continue;
}
if ( input_fname == 0 ) {
input_fname = a;
continue;
}
error("Unknown argument : %s.", a);
}
struct vcfanno_config *con = vcfanno_config_init();
if ( vcfanno_load_config(con, json_fname) != 0 )
error("Failed to load configure file. %s : %s", json_fname, strerror(errno));
vcfanno_config_debug(con);
if ( con->beds.n_beds == 0)
error("No bed database specified.");
if ( input_fname == 0 && (!isatty(fileno(stdin))) )
input_fname = "-";
if ( input_fname == 0 )
error("No input file.");
int out_type = FT_VCF;
if ( output_fname_type != 0 ) {
switch (output_fname_type[0]) {
case 'b':
out_type = FT_BCF_GZ; break;
case 'u':
out_type = FT_BCF; break;
case 'z':
out_type = FT_VCF_GZ; break;
case 'v':
out_type = FT_VCF; break;
default :
error("The output type \"%d\" not recognised\n", out_type);
};
}
htsFile *fp = hts_open(input_fname, "r");
if ( fp == NULL )
error("Failed to open %s : %s.", input_fname, strerror(errno));
htsFormat type = *hts_get_format(fp);
if ( type.format != vcf && type.format != bcf )
error("Unsupported input format. %s", input_fname);
bcf_hdr_t *hdr = bcf_hdr_read(fp);
if ( hdr == NULL )
error("Failed to parse header.");
bcf_hdr_t *hdr_out = bcf_hdr_dup(hdr);
htsFile *fout = output_fname == 0 ? hts_open("-", hts_bcf_wmode(out_type)) : hts_open(output_fname, hts_bcf_wmode(out_type));
struct beds_options opts = { .beds_is_inited = 0,};
beds_options_init(&opts);
opts.hdr_out = hdr_out;
for ( i = 0; i < con->beds.n_beds; ++i ) {
beds_database_add(&opts, con->beds.files[i].fname, con->beds.files[i].columns);
}
bcf_hdr_write(fout, hdr_out);
bcf1_t *line = bcf_init();
while ( bcf_read(fp, hdr, line) == 0 ) {
anno_beds_core(&opts, line);
bcf_write(fout, hdr_out, line);
}
bcf_destroy(line);
bcf_hdr_destroy(hdr);
bcf_hdr_destroy(hdr_out);
beds_options_destroy(&opts);
hts_close(fp);
hts_close(fout);
return 0;
}
static void init_data(args_t *args)
{
args->aux.srs = bcf_sr_init();
// Open files for input and output, initialize structures
if ( args->targets )
{
if ( bcf_sr_set_targets(args->aux.srs, args->targets, args->targets_is_file, args->aux.flag&CALL_CONSTR_ALLELES ? 3 : 0)<0 )
error("Failed to read the targets: %s\n", args->targets);
if ( args->aux.flag&CALL_CONSTR_ALLELES && args->flag&CF_INS_MISSED )
{
args->aux.srs->targets->missed_reg_handler = print_missed_line;
args->aux.srs->targets->missed_reg_data = args;
}
}
if ( args->regions )
{
if ( bcf_sr_set_regions(args->aux.srs, args->regions, args->regions_is_file)<0 )
error("Failed to read the targets: %s\n", args->regions);
}
if ( !bcf_sr_add_reader(args->aux.srs, args->bcf_fname) ) error("Failed to open %s: %s\n", args->bcf_fname,bcf_sr_strerror(args->aux.srs->errnum));
args->aux.hdr = bcf_sr_get_header(args->aux.srs,0);
int i;
if ( args->samples_fname )
{
set_samples(args, args->samples_fname, args->samples_is_file);
if ( args->aux.flag&CALL_CONSTR_TRIO )
{
if ( 3*args->aux.nfams!=args->nsamples ) error("Expected only trios in %s, sorry!\n", args->samples_fname);
fprintf(stderr,"Detected %d samples in %d trio families\n", args->nsamples,args->aux.nfams);
}
args->nsex = ploidy_nsex(args->ploidy);
args->sex2ploidy = (int*) calloc(args->nsex,sizeof(int));
args->sex2ploidy_prev = (int*) calloc(args->nsex,sizeof(int));
args->aux.ploidy = (uint8_t*) malloc(args->nsamples);
for (i=0; i<args->nsamples; i++) args->aux.ploidy[i] = 2;
for (i=0; i<args->nsex; i++) args->sex2ploidy_prev[i] = 2;
}
if ( args->samples_map )
{
args->aux.hdr = bcf_hdr_subset(bcf_sr_get_header(args->aux.srs,0), args->nsamples, args->samples, args->samples_map);
if ( !args->aux.hdr ) error("Error occurred while subsetting samples\n");
for (i=0; i<args->nsamples; i++)
if ( args->samples_map[i]<0 ) error("No such sample: %s\n", args->samples[i]);
if ( !bcf_hdr_nsamples(args->aux.hdr) ) error("No matching sample found\n");
}
else
{
args->aux.hdr = bcf_hdr_dup(bcf_sr_get_header(args->aux.srs,0));
for (i=0; i<args->nsamples; i++)
if ( bcf_hdr_id2int(args->aux.hdr,BCF_DT_SAMPLE,args->samples[i])<0 )
error("No such sample: %s\n", args->samples[i]);
}
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));
if ( args->flag & CF_QCALL )
return;
if ( args->flag & CF_MCALL )
mcall_init(&args->aux);
if ( args->flag & CF_CCALL )
ccall_init(&args->aux);
if ( args->flag&CF_GVCF )
{
bcf_hdr_append(args->aux.hdr,"##INFO=<ID=END,Number=1,Type=Integer,Description=\"End position of the variant described in this record\">");
args->gvcf.rid = -1;
args->gvcf.line = bcf_init1();
args->gvcf.gt = (int32_t*) malloc(2*sizeof(int32_t)*bcf_hdr_nsamples(args->aux.hdr));
for (i=0; i<bcf_hdr_nsamples(args->aux.hdr); i++)
{
args->gvcf.gt[2*i+0] = bcf_gt_unphased(0);
args->gvcf.gt[2*i+1] = bcf_gt_unphased(0);
}
}
bcf_hdr_remove(args->aux.hdr, BCF_HL_INFO, "QS");
bcf_hdr_remove(args->aux.hdr, BCF_HL_INFO, "I16");
bcf_hdr_append_version(args->aux.hdr, args->argc, args->argv, "bcftools_call");
bcf_hdr_write(args->out_fh, args->aux.hdr);
if ( args->flag&CF_INS_MISSED ) init_missed_line(args);
}
