float estimate_gt() {
int i,j,count=0,g;
float den,af=af_start,old_af=100;
dosage=(float *)realloc(dosage,n*sizeof(float));
prior[0] = (1-af)*(1-af);
prior[1] = 2 * af * (1-af);
prior[2] = af * af;
for( i=0;i<n;i++)
if(bcf_float_is_missing(gl[i*3]) || bcf_float_is_missing(gl[i*3+1]) || bcf_float_is_missing(gl[i*3+2]) )
for( j=0;j<3;j++) gl[i*3+j]=1.;
//Expectation-Maximisation to estimate AF
while(1) {
af = 0.;
for(i=0;i<n;i++) {
dosage[i]=0.;
den = 0.;
for(j=0;j<3;j++) {
tmp_gl[j] = prior[j] *pow(10.,gl[i*3 + j]);
den += tmp_gl[j];
}
for(j=0;j<3;j++) tmp_gl[j]/=den;
for(j=1;j<3;j++) dosage[i] += j * tmp_gl[j];
af+=dosage[i];
}
af /= (2.*n);
prior[2] = af * af;
prior[1] = 2 * af * (1-af);
prior[0] = (1. - af)*(1. - af);
// assert(af>=0 && af<=1);
// fprintf(stderr,"count=%d af=%.5f old_af=%.5f %.5f\n",count,af,old_af,fabs(af-old_af));
if( (count>0 && fabs(af-old_af)<.000001) || count>100)
break;
else
old_af=af;
count++;
// fprintf(stderr,"%d\n",count);
}
//dosage -> GT
for(i=0;i<n;i++) {
g = roundf(dosage[i]);
if(g==0) {
gt[i*2] = bcf_gt_unphased(0);
gt[i*2+1] = bcf_gt_unphased(0);
}
else if(g==1) {
gt[i*2] = bcf_gt_unphased(0);
gt[i*2+1] = bcf_gt_unphased(1);
}
else {
gt[i*2] = bcf_gt_unphased(1);
gt[i*2+1] = bcf_gt_unphased(1);
}
}
return(af);
}
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 abcWriteBcf::print(funkyPars *pars){
if(doBcf==0)
return;
kstring_t buf;
if(fp==NULL){
buf.s=NULL;buf.l=buf.m=0;
fp=aio::openFileHts(outfiles,".bcf");
hdr = bcf_hdr_init("w");
rec = bcf_init1();
print_bcf_header(fp,hdr,args,buf,header);
}
lh3struct *lh3 = (lh3struct*) pars->extras[5];
freqStruct *freq = (freqStruct *) pars->extras[6];
genoCalls *geno = (genoCalls *) pars->extras[10];
for(int s=0;s<pars->numSites;s++){
if(pars->keepSites[s]==0)
continue;
rec->rid = bcf_hdr_name2id(hdr,header->target_name[pars->refId]);
rec->pos = pars->posi[s];//<- maybe one index?
// bcf_update_id(hdr, rec, "rs6054257");
char majmin[4]={intToRef[pars->major[s]],',',intToRef[pars->minor[s]],'\0'};
bcf_update_alleles_str(hdr, rec, majmin);
rec->qual = 29;
// .. FILTER
int32_t tmpi = bcf_hdr_id2int(hdr, BCF_DT_ID, "PASS");
bcf_update_filter(hdr, rec, &tmpi, 1);
// .. INFO
tmpi = pars->keepSites[s];
bcf_update_info_int32(hdr, rec, "NS", &tmpi, 1);
if(pars->counts){
int depth = 0;
for(int i=0; i<4*pars->nInd; i++)
depth += pars->counts[s][i];
tmpi = depth;
bcf_update_info_int32(hdr, rec, "DP", &tmpi, 1);
}
if(freq){
float tmpf = freq->freq_EM[s];
bcf_update_info_float(hdr, rec, "AF", &tmpf, 1);
}
// .. FORMAT
assert(geno);
if(geno){
int32_t *tmpia = (int*)malloc(bcf_hdr_nsamples(hdr)*2*sizeof(int32_t));
for(int i=0; i<pars->nInd;i++){
if(geno->dat[s][i]==0){
tmpia[2*i+0] = bcf_gt_unphased(0);
tmpia[2*i+1] = bcf_gt_unphased(0);
}else if(geno->dat[s][i]==1){
tmpia[2*i+0] = bcf_gt_unphased(0);
tmpia[2*i+1] = bcf_gt_unphased(1);
} else{
tmpia[2*i+0] = bcf_gt_unphased(1);
tmpia[2*i+1] = bcf_gt_unphased(1);
}
}
bcf_update_genotypes(hdr, rec, tmpia, bcf_hdr_nsamples(hdr)*2);
free(tmpia);
}
if(pars->counts){
int32_t *tmpfa = (int32_t*)malloc(sizeof(int32_t)*bcf_hdr_nsamples(hdr));
suint *ary=pars->counts[s];
for(int i=0;i<bcf_hdr_nsamples(hdr);i++)
tmpfa[i] = ary[0]+ary[1]+ary[2]+ary[3];
bcf_update_format_int32(hdr, rec, "DP", tmpfa,bcf_hdr_nsamples(hdr) );
free(tmpfa);
}
assert(lh3);
if(lh3){
float *tmpfa = (float*)malloc(3*bcf_hdr_nsamples(hdr)*sizeof(float ));
int32_t *tmpi = (int32_t*)malloc(3*bcf_hdr_nsamples(hdr)*sizeof(int32_t));
double *ary = lh3->lh3[s];
for(int i=0;i<bcf_hdr_nsamples(hdr);i++)
for(int j=0;j<3;j++){
tmpfa[i*3+j] = ary[i*3+j]/M_LN10;
tmpi[i*3+j] =(int) -log10(exp(ary[i*3+j]))*10.0;
// fprintf(stderr,"pl:%d raw:%f\n",tmpi[i*3+j],ary[i*3+j]);
}
bcf_update_format_float(hdr, rec, "GL", tmpfa,3*bcf_hdr_nsamples(hdr) );
bcf_update_format_int32(hdr, rec, "PL", tmpi,3*bcf_hdr_nsamples(hdr) );
free(tmpfa);
free(tmpi);
}
if ( bcf_write1(fp, hdr, rec)!=0 ){
fprintf(stderr,"Failed to write to \n");
exit(0);
}
// fprintf(stderr,"------\n");
bcf_clear1(rec);
}
}
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE. */
#include <stdio.h>
#include <stdlib.h>
#include <htslib/vcf.h>
#include <inttypes.h>
#include <getopt.h>
bcf_hdr_t *in_hdr, *out_hdr;
int32_t *gts = NULL, mgts = 0;
uint64_t nchanged = 0;
int new_gt = bcf_gt_unphased(0);
const char *about(void)
{
return "Set missing genotypes (\"./.\") to ref allele (\"0/0\" or \"0|0\").\n";
}
const char *usage(void)
{
return
"\n"
"About: Set missing genotypes\n"
"Usage: bcftools +missing2ref [General Options] -- [Plugin Options]\n"
"Options:\n"
" run \"bcftools plugin\" for a list of common options\n"
"\n"
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);
}
