bam_smpl_t *bam_smpl_init(void)
{
bam_smpl_t *bsmpl;
bsmpl = (bam_smpl_t*) calloc(1, sizeof(bam_smpl_t));
bsmpl->name2idx = khash_str2int_init();
return bsmpl;
}
// Add a new region into a list sorted by start,end. On input the coordinates
// are 1-based, stored 0-based, inclusive.
static void _regions_add(bcf_sr_regions_t *reg, const char *chr, int start, int end)
{
if ( start==-1 && end==-1 )
{
start = 0; end = MAX_CSI_COOR-1;
}
else
{
start--; end--; // store 0-based coordinates
}
if ( !reg->seq_hash )
reg->seq_hash = khash_str2int_init();
int iseq;
if ( khash_str2int_get(reg->seq_hash, chr, &iseq)<0 )
{
// the chromosome block does not exist
iseq = reg->nseqs++;
reg->seq_names = (char**) realloc(reg->seq_names,sizeof(char*)*reg->nseqs);
reg->regs = (region_t*) realloc(reg->regs,sizeof(region_t)*reg->nseqs);
memset(®->regs[reg->nseqs-1],0,sizeof(region_t));
reg->seq_names[iseq] = strdup(chr);
reg->regs[iseq].creg = -1;
khash_str2int_set(reg->seq_hash,reg->seq_names[iseq],iseq);
}
region_t *creg = ®->regs[iseq];
// the regions may not be sorted on input: binary search
int i, min = 0, max = creg->nregs - 1;
while ( min<=max )
{
i = (max+min)/2;
if ( start < creg->regs[i].start ) max = i - 1;
else if ( start > creg->regs[i].start ) min = i + 1;
else break;
}
if ( min>max || creg->regs[i].start!=start || creg->regs[i].end!=end )
{
// no such region, insert a new one just after max
hts_expand(region1_t,creg->nregs+1,creg->mregs,creg->regs);
if ( ++max < creg->nregs )
memmove(&creg->regs[max+1],&creg->regs[max],(creg->nregs - max)*sizeof(region1_t));
creg->regs[max].start = start;
creg->regs[max].end = end;
creg->nregs++;
}
}
static void list_columns(args_t *args)
{
void *has_sample = NULL;
if ( args->sample_list )
{
has_sample = khash_str2int_init();
int i, nsmpl;
char **smpl = hts_readlist(args->sample_list, args->sample_is_file, &nsmpl);
for (i=0; i<nsmpl; i++) khash_str2int_inc(has_sample, smpl[i]);
free(smpl);
}
int i;
bcf_sr_t *reader = &args->files->readers[0];
for (i=0; i<bcf_hdr_nsamples(reader->header); i++)
{
if ( has_sample && !khash_str2int_has_key(has_sample, reader->header->samples[i]) ) continue;
printf("%s\n", reader->header->samples[i]);
}
if ( has_sample )
khash_str2int_destroy_free(has_sample);
}
static void bsmpl_add_readgroup(bam_smpl_t *bsmpl, file_t *file, const char *rg_id, const char *smpl_name)
{
int ismpl = -1;
if ( smpl_name )
{
if ( khash_str2int_get(bsmpl->name2idx,smpl_name,&ismpl) < 0 )
{
// new sample
bsmpl->nsmpl++;
bsmpl->smpl = (char**) realloc(bsmpl->smpl,sizeof(char*)*bsmpl->nsmpl);
bsmpl->smpl[bsmpl->nsmpl-1] = strdup(smpl_name);
ismpl = khash_str2int_inc(bsmpl->name2idx,bsmpl->smpl[bsmpl->nsmpl-1]);
}
}
if ( !strcmp("*",rg_id) )
{
// all read groups in the bam treated as the same sample
file->default_idx = ismpl;
return;
}
if ( !file->rg2idx ) file->rg2idx = khash_str2int_init();
if ( khash_str2int_has_key(file->rg2idx,rg_id) ) return; // duplicate @RG:ID
khash_str2int_set(file->rg2idx, strdup(rg_id), ismpl);
}
static void bcf_sr_sort_set(bcf_srs_t *readers, sr_sort_t *srt, const char *chr, int min_pos)
{
if ( !srt->grp_str2int )
{
// first time here, initialize
if ( !srt->pair )
{
if ( readers->collapse==COLLAPSE_NONE ) readers->collapse = BCF_SR_PAIR_EXACT;
bcf_sr_set_opt(readers, BCF_SR_PAIR_LOGIC, readers->collapse);
}
bcf_sr_init_scores(srt);
srt->grp_str2int = khash_str2int_init();
srt->var_str2int = khash_str2int_init();
}
int k;
khash_t(str2int) *hash;
hash = srt->grp_str2int;
for (k=0; k < kh_end(hash); k++)
if ( kh_exist(hash,k) ) free((char*)kh_key(hash,k));
hash = srt->var_str2int;
for (k=0; k < kh_end(hash); k++)
if ( kh_exist(hash,k) ) free((char*)kh_key(hash,k));
kh_clear(str2int, srt->grp_str2int);
kh_clear(str2int, srt->var_str2int);
srt->ngrp = srt->nvar = srt->nvset = 0;
grp_t grp;
memset(&grp,0,sizeof(grp_t));
// group VCFs into groups, each with a unique combination of variants in the duplicate lines
int ireader,ivar,irec,igrp,ivset,iact;
for (ireader=0; ireader<readers->nreaders; ireader++) srt->vcf_buf[ireader].nrec = 0;
for (iact=0; iact<srt->nactive; iact++)
{
ireader = srt->active[iact];
bcf_sr_t *reader = &readers->readers[ireader];
int rid = bcf_hdr_name2id(reader->header, chr);
grp.nvar = 0;
hts_expand(int,reader->nbuffer,srt->moff,srt->off);
srt->noff = 0;
srt->str.l = 0;
for (irec=1; irec<=reader->nbuffer; irec++)
{
bcf1_t *line = reader->buffer[irec];
if ( line->rid!=rid || line->pos!=min_pos ) break;
if ( srt->str.l ) kputc(';',&srt->str);
srt->off[srt->noff++] = srt->str.l;
size_t beg = srt->str.l;
for (ivar=1; ivar<line->n_allele; ivar++)
{
if ( ivar>1 ) kputc(',',&srt->str);
kputs(line->d.allele[0],&srt->str);
kputc('>',&srt->str);
kputs(line->d.allele[ivar],&srt->str);
}
if ( line->n_allele==1 )
{
kputs(line->d.allele[0],&srt->str);
kputsn(">.",2,&srt->str);
}
// Create new variant or attach to existing one. But careful, there can be duplicate
// records with the same POS,REF,ALT (e.g. in dbSNP-b142)
char *var_str = beg + srt->str.s;
int ret, var_idx = 0, var_end = srt->str.l;
while ( 1 )
{
ret = khash_str2int_get(srt->var_str2int, var_str, &ivar);
if ( ret==-1 ) break;
var_t *var = &srt->var[ivar];
if ( var->vcf[var->nvcf-1] != ireader ) break;
srt->str.l = var_end;
kputw(var_idx, &srt->str);
var_str = beg + srt->str.s;
var_idx++;
}
if ( ret==-1 )
{
ivar = srt->nvar++;
hts_expand0(var_t,srt->nvar,srt->mvar,srt->var);
srt->var[ivar].nvcf = 0;
khash_str2int_set(srt->var_str2int, strdup(var_str), ivar);
free(srt->var[ivar].str); // possible left-over from the previous position
}
var_t *var = &srt->var[ivar];
var->nalt = line->n_allele - 1;
var->type = bcf_get_variant_types(line);
srt->str.s[var_end] = 0;
if ( ret==-1 )
var->str = strdup(var_str);
int mvcf = var->mvcf;
var->nvcf++;
hts_expand0(int*, var->nvcf, var->mvcf, var->vcf);
if ( mvcf != var->mvcf ) var->rec = (bcf1_t **) realloc(var->rec,sizeof(bcf1_t*)*var->mvcf);
var->vcf[var->nvcf-1] = ireader;
var->rec[var->nvcf-1] = line;
grp.nvar++;
hts_expand(var_t,grp.nvar,grp.mvar,grp.var);
grp.var[grp.nvar-1] = ivar;
}
char *grp_key = grp_create_key(srt);
int ret = khash_str2int_get(srt->grp_str2int, grp_key, &igrp);
if ( ret==-1 )
{
igrp = srt->ngrp++;
hts_expand0(grp_t, srt->ngrp, srt->mgrp, srt->grp);
free(srt->grp[igrp].var);
srt->grp[igrp] = grp;
srt->grp[igrp].key = grp_key;
khash_str2int_set(srt->grp_str2int, grp_key, igrp);
memset(&grp,0,sizeof(grp_t));
}
else
free(grp_key);
srt->grp[igrp].nvcf++;
}
free(grp.var);
// initialize bitmask - which groups is the variant present in
for (ivar=0; ivar<srt->nvar; ivar++)
{
srt->var[ivar].mask = kbs_resize(srt->var[ivar].mask, srt->ngrp);
kbs_clear(srt->var[ivar].mask);
}
for (igrp=0; igrp<srt->ngrp; igrp++)
{
for (ivar=0; ivar<srt->grp[igrp].nvar; ivar++)
{
int i = srt->grp[igrp].var[ivar];
kbs_insert(srt->var[i].mask, igrp);
}
}
// create the initial list of variant sets
for (ivar=0; ivar<srt->nvar; ivar++)
{
ivset = srt->nvset++;
hts_expand0(varset_t, srt->nvset, srt->mvset, srt->vset);
varset_t *vset = &srt->vset[ivset];
vset->nvar = 1;
hts_expand0(var_t, vset->nvar, vset->mvar, vset->var);
vset->var[vset->nvar-1] = ivar;
var_t *var = &srt->var[ivar];
vset->cnt = var->nvcf;
vset->mask = kbs_resize(vset->mask, srt->ngrp);
kbs_clear(vset->mask);
kbs_bitwise_or(vset->mask, var->mask);
int type = 0;
if ( var->type==VCF_REF ) type |= SR_REF;
else
{
if ( var->type & VCF_SNP ) type |= SR_SNP;
if ( var->type & VCF_MNP ) type |= SR_SNP;
if ( var->type & VCF_INDEL ) type |= SR_INDEL;
if ( var->type & VCF_OTHER ) type |= SR_OTHER;
}
var->type = type;
}
#if DEBUG_VSETS
debug_vsets(srt);
#endif
// initialize the pairing matrix
hts_expand(int, srt->ngrp*srt->nvset, srt->mpmat, srt->pmat);
hts_expand(int, srt->nvset, srt->mcnt, srt->cnt);
memset(srt->pmat, 0, sizeof(*srt->pmat)*srt->ngrp*srt->nvset);
for (ivset=0; ivset<srt->nvset; ivset++)
{
varset_t *vset = &srt->vset[ivset];
for (igrp=0; igrp<srt->ngrp; igrp++) srt->pmat[ivset*srt->ngrp+igrp] = 0;
srt->cnt[ivset] = vset->cnt;
}
// pair the lines
while ( srt->nvset )
{
#if DEBUG_VSETS
fprintf(stderr,"\n");
debug_vsets(srt);
#endif
int imax = 0;
for (ivset=1; ivset<srt->nvset; ivset++)
if ( srt->cnt[imax] < srt->cnt[ivset] ) imax = ivset;
int ipair = -1;
uint32_t max_score = 0;
for (ivset=0; ivset<srt->nvset; ivset++)
{
if ( kbs_logical_and(srt->vset[imax].mask,srt->vset[ivset].mask) ) continue; // cannot be merged
uint32_t score = pairing_score(srt, imax, ivset);
// fprintf(stderr,"score: %d %d, logic=%d \t..\t %u\n", imax,ivset,srt->pair,score);
if ( max_score < score ) { max_score = score; ipair = ivset; }
}
// merge rows creating a new variant set this way
if ( ipair!=-1 && ipair!=imax )
{
imax = merge_vsets(srt, imax, ipair);
continue;
}
push_vset(srt, imax);
}
srt->chr = chr;
srt->pos = min_pos;
}
bcf_sr_regions_t *bcf_sr_regions_init(const char *regions, int is_file, int ichr, int ifrom, int ito)
{
bcf_sr_regions_t *reg;
if ( !is_file ) return _regions_init_string(regions);
reg = (bcf_sr_regions_t *) calloc(1, sizeof(bcf_sr_regions_t));
reg->start = reg->end = -1;
reg->prev_start = reg->prev_seq = -1;
reg->file = hts_open(regions, "rb");
if ( !reg->file )
{
fprintf(stderr,"[%s:%d %s] Could not open file: %s\n", __FILE__,__LINE__,__FUNCTION__,regions);
free(reg);
return NULL;
}
reg->tbx = tbx_index_load(regions);
if ( !reg->tbx )
{
int len = strlen(regions);
int is_bed = strcasecmp(".bed",regions+len-4) ? 0 : 1;
if ( !is_bed && !strcasecmp(".bed.gz",regions+len-7) ) is_bed = 1;
if ( reg->file->format.format==vcf ) ito = 1;
// read the whole file, tabix index is not present
while ( hts_getline(reg->file, KS_SEP_LINE, ®->line) > 0 )
{
char *chr, *chr_end;
int from, to, ret;
ret = _regions_parse_line(reg->line.s, ichr,ifrom,abs(ito), &chr,&chr_end,&from,&to);
if ( ret < 0 )
{
if ( ito<0 )
ret = _regions_parse_line(reg->line.s, ichr,ifrom,ifrom, &chr,&chr_end,&from,&to);
if ( ret<0 )
{
fprintf(stderr,"[%s:%d] Could not parse the file %s, using the columns %d,%d[,%d]\n", __FILE__,__LINE__,regions,ichr+1,ifrom+1,ito+1);
hts_close(reg->file); reg->file = NULL; free(reg);
return NULL;
}
}
if ( !ret ) continue;
if ( is_bed ) from++;
*chr_end = 0;
_regions_add(reg, chr, from, to);
*chr_end = '\t';
}
hts_close(reg->file); reg->file = NULL;
if ( !reg->nseqs ) { free(reg); return NULL; }
return reg;
}
reg->seq_names = (char**) tbx_seqnames(reg->tbx, ®->nseqs);
if ( !reg->seq_hash )
reg->seq_hash = khash_str2int_init();
int i;
for (i=0; i<reg->nseqs; i++)
{
khash_str2int_set(reg->seq_hash,reg->seq_names[i],i);
}
reg->fname = strdup(regions);
reg->is_bin = 1;
return reg;
}
int bcf_sr_set_samples(bcf_srs_t *files, const char *fname, int is_file)
{
int i, j, nsmpl, free_smpl = 0;
char **smpl = NULL;
void *exclude = (fname[0]=='^') ? khash_str2int_init() : NULL;
if ( exclude || strcmp("-",fname) ) // "-" stands for all samples
{
smpl = hts_readlist(fname, is_file, &nsmpl);
if ( !smpl )
{
fprintf(stderr,"Could not read the file: \"%s\"\n", fname);
return 0;
}
if ( exclude )
{
for (i=0; i<nsmpl; i++)
khash_str2int_inc(exclude, smpl[i]);
}
free_smpl = 1;
}
if ( !smpl )
{
smpl = files->readers[0].header->samples; // intersection of all samples
nsmpl = bcf_hdr_nsamples(files->readers[0].header);
}
files->samples = NULL;
files->n_smpl = 0;
for (i=0; i<nsmpl; i++)
{
if ( exclude && khash_str2int_has_key(exclude,smpl[i]) ) continue;
int n_isec = 0;
for (j=0; j<files->nreaders; j++)
{
if ( bcf_hdr_id2int(files->readers[j].header, BCF_DT_SAMPLE, smpl[i])<0 ) break;
n_isec++;
}
if ( n_isec!=files->nreaders )
{
fprintf(stderr,"Warning: The sample \"%s\" was not found in %s, skipping\n", smpl[i], files->readers[n_isec].fname);
continue;
}
files->samples = (char**) realloc(files->samples, (files->n_smpl+1)*sizeof(const char*));
files->samples[files->n_smpl++] = strdup(smpl[i]);
}
if ( exclude ) khash_str2int_destroy(exclude);
if ( free_smpl )
{
for (i=0; i<nsmpl; i++) free(smpl[i]);
free(smpl);
}
if ( !files->n_smpl )
{
if ( files->nreaders>1 )
fprintf(stderr,"No samples in common.\n");
return 0;
}
for (i=0; i<files->nreaders; i++)
{
bcf_sr_t *reader = &files->readers[i];
reader->samples = (int*) malloc(sizeof(int)*files->n_smpl);
reader->n_smpl = files->n_smpl;
for (j=0; j<files->n_smpl; j++)
reader->samples[j] = bcf_hdr_id2int(reader->header, BCF_DT_SAMPLE, files->samples[j]);
}
return 1;
}
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;
}
static void init_data(args_t *args)
{
int i;
args->hdr = args->files->readers[0].header;
if (args->calc_ac && args->update_info)
{
bcf_hdr_append(args->hdr,"##INFO=<ID=AC,Number=A,Type=Integer,Description=\"Allele count in genotypes\">");
bcf_hdr_append(args->hdr,"##INFO=<ID=AN,Number=1,Type=Integer,Description=\"Total number of alleles in called genotypes\">");
}
bcf_hdr_append_version(args->hdr, args->argc, args->argv, "bcftools_view");
// setup sample data
if (args->sample_names)
{
void *hdr_samples = khash_str2int_init();
for (i=0; i<bcf_hdr_nsamples(args->hdr); i++)
khash_str2int_inc(hdr_samples, bcf_hdr_int2id(args->hdr,BCF_DT_SAMPLE,i));
void *exclude = (args->sample_names[0]=='^') ? khash_str2int_init() : NULL;
int nsmpl;
char **smpl = NULL;
args->samples = NULL; args->n_samples = 0;
smpl = hts_readlist(exclude ? &args->sample_names[1] : args->sample_names, args->sample_is_file, &nsmpl);
if ( !smpl )
{
error("Could not read the list: \"%s\"\n", exclude ? &args->sample_names[1] : args->sample_names);
}
if ( exclude )
{
for (i=0; i<nsmpl; i++) {
if (!khash_str2int_has_key(hdr_samples,smpl[i])) {
if (args->force_samples) {
fprintf(stderr, "Warn: exclude called for sample that does not exist in header: \"%s\"... skipping\n", smpl[i]);
} else {
error("Error: exclude called for sample that does not exist in header: \"%s\". Use \"--force-samples\" to ignore this error.\n", smpl[i]);
}
}
khash_str2int_inc(exclude, smpl[i]);
}
for (i=0; i<bcf_hdr_nsamples(args->hdr); i++)
{
if ( exclude && khash_str2int_has_key(exclude,bcf_hdr_int2id(args->hdr,BCF_DT_SAMPLE,i)) ) continue;
args->samples = (char**) realloc(args->samples, (args->n_samples+1)*sizeof(const char*));
args->samples[args->n_samples++] = strdup(bcf_hdr_int2id(args->hdr,BCF_DT_SAMPLE,i));
}
khash_str2int_destroy(exclude);
}
else
{
for (i=0; i<nsmpl; i++) {
if (!khash_str2int_has_key(hdr_samples,smpl[i])) {
if (args->force_samples) {
fprintf(stderr, "Warn: subset called for sample that does not exist in header: \"%s\"... skipping\n", smpl[i]);
continue;
} else {
error("Error: subset called for sample that does not exist in header: \"%s\". Use \"--force-samples\" to ignore this error.\n", smpl[i]);
}
}
args->samples = (char**) realloc(args->samples, (args->n_samples+1)*sizeof(const char*));
args->samples[args->n_samples++] = strdup(smpl[i]);
}
}
for (i=0; i<nsmpl; i++) free(smpl[i]);
free(smpl);
khash_str2int_destroy(hdr_samples);
if (args->n_samples == 0) {
fprintf(stderr, "Warn: subsetting has removed all samples\n");
args->sites_only = 1;
}
}
if (args->n_samples)
args->imap = (int*)malloc(args->n_samples * sizeof(int));
// determine variant types to include/exclude
if (args->include_types || args->exclude_types) {
if (args->include_types && args->exclude_types) {
fprintf(stderr, "Error: only supply one of --include-types, --exclude-types options\n");
exit(1);
}
char **type_list = 0;
int m = 0, n = 0;
const char *q, *p;
for (q = p = args->include_types ? args->include_types : args->exclude_types;; ++p) {
if (*p == ',' || *p == 0) {
if (m == n) {
m = m? m<<1 : 16;
type_list = (char**)realloc(type_list, m * sizeof(char*));
}
type_list[n] = (char*)calloc(p - q + 1, 1);
strncpy(type_list[n++], q, p - q);
q = p + 1;
if (*p == 0) break;
}
}
type_list = (char**)realloc(type_list, n * sizeof(char*));
if (args->include_types) {
args->include = 0;
for (i = 0; i < n; ++i) {
if (strcmp(type_list[i], "snps") == 0) args->include |= VCF_SNP;
else if (strcmp(type_list[i], "indels") == 0) args->include |= VCF_INDEL;
else if (strcmp(type_list[i], "mnps") == 0) args->include |= VCF_MNP;
else if (strcmp(type_list[i], "other") == 0) args->include |= VCF_OTHER;
else {
fprintf(stderr, "[E::%s] unknown type\n", type_list[i]);
fprintf(stderr, "Accepted types are snps, indels, mnps, other\n");
exit(1);
}
}
}
if (args->exclude_types) {
args->exclude = 0;
for (i = 0; i < n; ++i) {
if (strcmp(type_list[i], "snps") == 0) args->exclude |= VCF_SNP;
else if (strcmp(type_list[i], "indels") == 0) args->exclude |= VCF_INDEL;
else if (strcmp(type_list[i], "mnps") == 0) args->exclude |= VCF_MNP;
else if (strcmp(type_list[i], "other") == 0) args->exclude |= VCF_OTHER;
else {
fprintf(stderr, "[E::%s] unknown type\n", type_list[i]);
fprintf(stderr, "Accepted types are snps, indels, mnps, other\n");
exit(1);
}
}
}
for (i = 0; i < n; ++i)
free(type_list[i]);
free(type_list);
}
// setup output
char modew[8];
strcpy(modew, "w");
if (args->clevel >= 0 && args->clevel <= 9) sprintf(modew + 1, "%d", args->clevel);
if (args->output_type==FT_BCF) strcat(modew, "bu"); // uncompressed BCF
else if (args->output_type & FT_BCF) strcat(modew, "b"); // compressed BCF
else if (args->output_type & FT_GZ) strcat(modew,"z"); // compressed VCF
args->out = hts_open(args->fn_out ? args->fn_out : "-", modew);
if ( !args->out ) error("%s: %s\n", args->fn_out,strerror(errno));
// headers: hdr=full header, hsub=subset header, hnull=sites only header
if (args->sites_only){
args->hnull = bcf_hdr_subset(args->hdr, 0, 0, 0);
bcf_hdr_remove(args->hnull, BCF_HL_FMT, NULL);
}
if (args->n_samples > 0)
{
args->hsub = bcf_hdr_subset(args->hdr, args->n_samples, args->samples, args->imap);
if ( !args->hsub ) error("Error occurred while subsetting samples\n");
if ( args->n_samples != bcf_hdr_nsamples(args->hsub) )
{
int i;
for (i=0; i<args->n_samples; i++)
if ( args->imap[i]<0 ) error("Error: No such sample: \"%s\"\n", args->samples[i]);
}
}
if ( args->filter_str )
args->filter = filter_init(args->hdr, args->filter_str);
}
/*
The logic of this function is a bit complicated because we want to work
also with broken bams containing read groups that are not listed in the
header. The desired behavior is as follows:
- when -G is given, read groups which are not listed in the header must
be given explicitly using the "?" symbol in -G.
Otherwise:
- if the bam has no header, all reads in the file are assigned to a
single sample named after the file
- if there is at least one sample defined in the header, reads with no
read group id or with a read group id not listed in the header are
assigned to the first sample encountered in the header
*/
int bam_smpl_add_bam(bam_smpl_t *bsmpl, char *bam_hdr, const char *fname)
{
bsmpl->nfiles++;
bsmpl->files = (file_t*) realloc(bsmpl->files,bsmpl->nfiles*sizeof(file_t));
file_t *file = &bsmpl->files[bsmpl->nfiles-1];
memset(file,0,sizeof(file_t));
file->fname = strdup(fname);
file->default_idx = -1;
if ( bsmpl->ignore_rg || !bam_hdr )
{
// The option --ignore-RG is set or there is no BAM header: use the file name as the sample name
bsmpl_add_readgroup(bsmpl,file,"*",file->fname);
return bsmpl->nfiles-1;
}
void *bam_smpls = khash_str2int_init();
int first_smpl = -1, nskipped = 0;
const char *p = bam_hdr, *q, *r;
while (p != NULL && (q = strstr(p, "@RG")) != 0)
{
char *eol = strchr(q + 3, '\n');
if (q > bam_hdr && *(q - 1) != '\n') { // @RG must be at start of line
p = eol;
continue;
}
p = q + 3;
if ((q = strstr(p, "\tID:")) != 0) q += 4;
if ((r = strstr(p, "\tSM:")) != 0) r += 4;
if (r && q)
{
char *u, *v;
int ioq, ior;
for (u = (char*)q; *u && *u != '\t' && *u != '\n'; ++u);
for (v = (char*)r; *v && *v != '\t' && *v != '\n'; ++v);
ioq = *u; ior = *v; *u = *v = '\0';
// q now points to a null terminated read group id
// r points to a null terminated sample name
if ( !strcmp("*",q) || !strcmp("?",q) )
error("Error: the read group IDs \"*\" and \"?\" have a special meaning in the mpileup code. Please fix the code or the bam: %s\n", fname);
int accept_rg = 1;
if ( bsmpl->sample_list )
{
// restrict samples based on the -s/-S options
char *name = khash_str2str_get(bsmpl->sample_list,r);
if ( bsmpl->sample_logic==0 )
accept_rg = name ? 0 : 1;
else if ( !name )
accept_rg = 0;
else
r = name;
}
if ( accept_rg && bsmpl->rg_list )
{
// restrict readgroups based on the -G option, possibly renaming the sample
accept_rg = bsmpl_keep_readgroup(bsmpl,file,q,&r);
}
if ( accept_rg )
bsmpl_add_readgroup(bsmpl,file,q,r);
else
{
bsmpl_add_readgroup(bsmpl,file,q,NULL); // ignore this RG but note that it was seen in the header
nskipped++;
}
if ( first_smpl<0 )
khash_str2int_get(bsmpl->name2idx,r,&first_smpl);
if ( !khash_str2int_has_key(bam_smpls,r) )
khash_str2int_inc(bam_smpls,strdup(r));
*u = ioq; *v = ior;
}
else
break;
p = eol;
}
int nsmpls = khash_str2int_size(bam_smpls);
khash_str2int_destroy_free(bam_smpls);
const char *smpl_name = NULL;
int accept_null_rg = 1;
if ( bsmpl->rg_list && !bsmpl_keep_readgroup(bsmpl,file,"?",&smpl_name) ) accept_null_rg = 0;
if ( bsmpl->sample_list && first_smpl==-1 ) accept_null_rg = 0;
if ( !accept_null_rg && first_smpl==-1 )
{
// no suitable read group is available in this bam: ignore the whole file.
free(file->fname);
if ( file->rg2idx ) khash_str2int_destroy_free(file->rg2idx);
bsmpl->nfiles--;
return -1;
}
if ( !accept_null_rg ) return bsmpl->nfiles-1;
if ( nsmpls==1 && !nskipped )
{
file->default_idx = first_smpl;
return bsmpl->nfiles-1;
}
if ( !smpl_name ) smpl_name = first_smpl==-1 ? file->fname : bsmpl->smpl[first_smpl];
bsmpl_add_readgroup(bsmpl,file,"?",smpl_name);
return bsmpl->nfiles-1;
}