static void init_header(bcf_hdr_t *hdr, const char *ori, int ori_type, const char *new_hdr_line)
{
if ( ori )
bcf_hdr_remove(hdr,ori_type,ori);
bcf_hdr_append(hdr, new_hdr_line);
}
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);
}
int ingest1(const char *input,const char *output,char *ref,bool exit_on_mismatch=true) {
cerr << "Input: " << input << "\tOutput: "<<output<<endl;
kstream_t *ks;
kstring_t str = {0,0,0};
gzFile fp = gzopen(input, "r");
VarBuffer vbuf(1000);
int prev_rid = -1;
if(fp==NULL) {
fprintf(stderr,"problem opening %s\n",input);
exit(1);
}
char *out_fname = (char *)malloc(strlen(output)+5);
strcpy(out_fname,output);
strcat(out_fname,".tmp");
if(fileexists(out_fname)) {
fprintf(stderr,"%s file already exists. will not overwrite\n",out_fname);
exit(1);
}
printf("depth: %s\n",out_fname);
gzFile depth_fp = gzopen(out_fname, "wb1");
strcpy(out_fname,output);
strcat(out_fname,".bcf");
if(fileexists(out_fname)) {
fprintf(stderr,"%s file already exists. will not overwrite\n",out_fname);
exit(1);
}
printf("variants: %s\n",out_fname);
htsFile *variant_fp=hts_open(out_fname,"wb1");
if(variant_fp==NULL) {
fprintf(stderr,"problem opening %s\n",input);
exit(1);
}
ks = ks_init(fp);
htsFile *hfp=hts_open(input, "r");
bcf_hdr_t *hdr_in = bcf_hdr_read(hfp);
hts_close(hfp);
//this is a hack to fix gvcfs where AD is incorrectly defined in the header. (vcf4.2 does not technically allow Number=R)
bcf_hdr_remove(hdr_in,BCF_HL_FMT,"AD");
assert( bcf_hdr_append(hdr_in,"##FORMAT=<ID=AD,Number=R,Type=Integer,Description=\"Allelic depths for the ref and alt alleles in the order listed. For indels this value only includes reads which confidently support each allele (posterior prob 0.999 or higher that read contains indicated allele vs all other intersecting indel alleles)\">") == 0);
//this is a hack to fix broken gvcfs where GQ is incorrectly labelled as float (v4.3 spec says it should be integer)
bcf_hdr_remove(hdr_in,BCF_HL_FMT,"GQ");
assert( bcf_hdr_append(hdr_in,"##FORMAT=<ID=GQ,Number=1,Type=Integer,Description=\"Genotype Quality\">") == 0);
// bcf_hdr_t *hdr_out=hdr_in;
bcf_hdr_t *hdr_out = bcf_hdr_dup(hdr_in);
remove_hdr_lines(hdr_out,BCF_HL_INFO);
remove_hdr_lines(hdr_out,BCF_HL_FLT);
bcf_hdr_sync(hdr_out);
//here we add FORMAT/PF. which is the pass filter flag for alts.
assert( bcf_hdr_append(hdr_out,"##FORMAT=<ID=PF,Number=A,Type=Integer,Description=\"variant was PASS filter in original sample gvcf\">") == 0);
args_t *norm_args = init_vcfnorm(hdr_out,ref);
norm_args->check_ref |= CHECK_REF_WARN;
bcf1_t *bcf_rec = bcf_init();
bcf_hdr_write(variant_fp, hdr_out);
kstring_t work1 = {0,0,0};
int buf[5];
ks_tokaux_t aux;
int ndec=0;
int ref_len,alt_len;
while( ks_getuntil(ks, '\n', &str, 0) >=0) {
// fprintf(stderr,"%s\n",str.s);
if(str.s[0]!='#') {
char *ptr = kstrtok(str.s,"\t",&aux);//chrom
ptr = kstrtok(NULL,NULL,&aux);//pos
work1.l=0;
kputsn(str.s,ptr-str.s-1, &work1);
buf[0] = bcf_hdr_name2id(hdr_in, work1.s);
assert( buf[0]>=0);
buf[1]=atoi(ptr)-1;
ptr = kstrtok(NULL,NULL,&aux);//ID
ptr = kstrtok(NULL,NULL,&aux);//REF
ref_len=0;
while(ptr[ref_len]!='\t') ref_len++;
ptr = kstrtok(NULL,NULL,&aux);//ALT
bool is_variant=false;
alt_len=0;
while(ptr[alt_len]!='\t') alt_len++;
if(ptr[0]!='.')
is_variant=true;
char * QUAL_ptr = kstrtok(NULL, NULL, &aux);
assert (QUAL_ptr != NULL);
for(int i=0;i<2;i++) ptr = kstrtok(NULL,NULL,&aux);// gets us to INFO
//find END if it is there
char *end_ptr=strstr(ptr,"END=") ;
if(end_ptr!=NULL)
buf[2]=atoi(end_ptr+4)-1;
else
buf[2]=buf[1]+alt_len-1;
ptr = kstrtok(NULL,NULL,&aux);//FORMAT
//find index of DP (if present)
//if not present, dont output anything (indels ignored)
char *DP_ptr = find_format(ptr,"DP");
int GQX = 0;
int QUAL = 0;
// AH: change code to use the minimum of GQ and QUAL fields if
// GQX is not defined. See here:
// https://support.basespace.illumina.com/knowledgebase/articles/144844-vcf-file
// "GQXGenotype quality. GQX is the minimum of the GQ value
// and the QUAL column. In general, these are similar values;
// taking the minimum makes GQX the more conservative measure of
// genotype quality."
if(DP_ptr!=NULL) {
buf[3]=atoi(DP_ptr);
char *GQX_ptr = find_format(ptr,"GQX");
if (GQX_ptr == NULL)
{
GQX_ptr = find_format(ptr,"GQ");
GQX = atoi(GQX_ptr);
if (QUAL_ptr[0] != '.')
{
QUAL = atoi(QUAL_ptr);
if (QUAL < GQX)
GQX = QUAL;
}
}
else
{
GQX = atoi(GQX_ptr);
}
//trying to reduce entropy on GQ to get better compression performance.
//1. rounds down to nearest 10.
//2. sets gq to min(gq,100).
buf[4]=GQX/10;
buf[4]*=10;
if(buf[4]>100) buf[4]=100;
// printf("%d\t%d\t%d\t%d\t%d\n",buf[0],buf[1],buf[2],buf[3],buf[4]);
if(gzwrite(depth_fp,buf,5*sizeof(int))!=(5*sizeof(int)))
die("ERROR: problem writing "+(string)out_fname+".tmp");
}
if(is_variant) {//wass this a variant? if so write it out to the bcf
norm_args->ntotal++;
vcf_parse(&str,hdr_in,bcf_rec);
// cerr<<bcf_rec->rid<<":"<<bcf_rec->pos<<endl;
if(prev_rid!=bcf_rec->rid)
vbuf.flush(variant_fp,hdr_out);
else
vbuf.flush(bcf_rec->pos,variant_fp,hdr_out);
prev_rid=bcf_rec->rid;
int32_t pass = bcf_has_filter(hdr_in, bcf_rec, ".");
bcf_update_format_int32(hdr_out,bcf_rec,"PF",&pass,1);
bcf_update_filter(hdr_out,bcf_rec,NULL,0);
if(bcf_rec->n_allele>2) {//split multi-allelics (using vcfnorm.c from bcftools1.3
norm_args->nsplit++;
split_multiallelic_to_biallelics(norm_args,bcf_rec );
for(int i=0;i<norm_args->ntmp_lines;i++){
remove_info(norm_args->tmp_lines[i]);
if(realign(norm_args,norm_args->tmp_lines[i]) != ERR_REF_MISMATCH)
ndec+=decompose(norm_args->tmp_lines[i],hdr_out,vbuf);
else
if(exit_on_mismatch)
die("vcf did not match the reference");
else
norm_args->nskipped++;
}
}
else {
remove_info(bcf_rec);
if( realign(norm_args,bcf_rec) != ERR_REF_MISMATCH)
ndec+=decompose(bcf_rec,hdr_out,vbuf);
else
if(exit_on_mismatch)
die("vcf did not match the reference");
else
norm_args->nskipped++;
}
vbuf.flush(bcf_rec->pos,variant_fp,hdr_out);
}
}
}
vbuf.flush(variant_fp,hdr_out);
bcf_hdr_destroy(hdr_in);
bcf_hdr_destroy(hdr_out);
bcf_destroy1(bcf_rec);
ks_destroy(ks);
gzclose(fp);
gzclose(depth_fp);
free(str.s);
free(work1.s);
hts_close(variant_fp);
destroy_data(norm_args);
fprintf(stderr,"Variant lines total/split/realigned/skipped:\t%d/%d/%d/%d\n", norm_args->ntotal,norm_args->nsplit,norm_args->nchanged,norm_args->nskipped);
fprintf(stderr,"Decomposed %d MNPs\n", ndec);
fprintf(stderr,"Indexing %s\n",out_fname);
bcf_index_build(out_fname, BCF_LIDX_SHIFT);
free(out_fname);
return 0;
}
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);
}
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);
}
