static void init_data(args_t *args)
{
args->header = args->files->readers[0].header;
int i, nsamples = 0, *samples = NULL;
if ( args->sample_list && strcmp("-",args->sample_list) )
{
for (i=0; i<args->files->nreaders; i++)
{
int ret = bcf_hdr_set_samples(args->files->readers[i].header,args->sample_list,args->sample_is_file);
if ( ret<0 ) error("Error parsing the sample list\n");
else if ( ret>0 ) error("Sample name mismatch: sample #%d not found in the header\n", ret);
}
if ( args->sample_list[0]!='^' )
{
// the sample ordering may be different if not negated
int n;
char **smpls = hts_readlist(args->sample_list, args->sample_is_file, &n);
if ( !smpls ) error("Could not parse %s\n", args->sample_list);
if ( n!=bcf_hdr_nsamples(args->files->readers[0].header) )
error("The number of samples does not match, perhaps some are present multiple times?\n");
nsamples = bcf_hdr_nsamples(args->files->readers[0].header);
samples = (int*) malloc(sizeof(int)*nsamples);
for (i=0; i<n; i++)
{
samples[i] = bcf_hdr_id2int(args->files->readers[0].header, BCF_DT_SAMPLE,smpls[i]);
free(smpls[i]);
}
free(smpls);
}
}
args->convert = convert_init(args->header, samples, nsamples, args->format_str);
if ( args->allow_undef_tags ) convert_set_option(args->convert, allow_undef_tags, 1);
free(samples);
int max_unpack = convert_max_unpack(args->convert);
if ( args->filter_str )
{
args->filter = filter_init(args->header, args->filter_str);
max_unpack |= filter_max_unpack(args->filter);
}
args->files->max_unpack = max_unpack;
}
static void init_data(args_t *args)
{
bcf_srs_t *files = bcf_sr_init();
if ( args->regions_list )
{
if ( bcf_sr_set_regions(files, args->regions_list, args->regions_is_file)<0 )
error("Failed to read the regions: %s\n", args->regions_list);
}
if ( args->targets_list )
{
if ( bcf_sr_set_targets(files, args->targets_list, args->targets_is_file, 0)<0 )
error("Failed to read the targets: %s\n", args->targets_list);
}
if ( !bcf_sr_add_reader(files, args->fname) ) error("Failed to open %s: %s\n", args->fname,bcf_sr_strerror(files->errnum));
bcf_hdr_t *hdr = files->readers[0].header;
if ( !args->sample )
{
if ( bcf_hdr_nsamples(hdr)>1 ) error("Missing the option -s, --sample\n");
args->sample = hdr->samples[0];
}
else if ( bcf_hdr_id2int(hdr,BCF_DT_SAMPLE,args->sample)<0 ) error("No such sample: %s\n", args->sample);
int ret = bcf_hdr_set_samples(hdr, args->sample, 0);
if ( ret<0 ) error("Error setting the sample: %s\n", args->sample);
if ( !bcf_hdr_idinfo_exists(hdr,BCF_HL_FMT,bcf_hdr_id2int(hdr,BCF_DT_ID,"BAF")) )
error("The tag FORMAT/BAF is not present in the VCF: %s\n", args->fname);
int i;
args->xvals = (double*) calloc(args->nbins,sizeof(double));
for (i=0; i<args->nbins; i++) args->xvals[i] = 1.0*i/(args->nbins-1);
// collect BAF distributions for all chromosomes
int idist = -1, nbaf = 0, nprocessed = 0, ntotal = 0, prev_chr = -1;
float *baf = NULL;
while ( bcf_sr_next_line(files) )
{
ntotal++;
bcf1_t *line = bcf_sr_get_line(files,0);
if ( bcf_get_format_float(hdr,line,"BAF",&baf,&nbaf) != 1 ) continue;
if ( bcf_float_is_missing(baf[0]) ) continue;
nprocessed++;
if ( prev_chr==-1 || prev_chr!=line->rid )
{
// new chromosome
idist = args->ndist++;
args->dist = (dist_t*) realloc(args->dist, sizeof(dist_t)*args->ndist);
memset(&args->dist[idist],0,sizeof(dist_t));
args->dist[idist].chr = strdup(bcf_seqname(hdr,line));
args->dist[idist].yvals = (double*) calloc(args->nbins,sizeof(double));
args->dist[idist].xvals = args->xvals;
args->dist[idist].nvals = args->nbins;
prev_chr = line->rid;
}
int bin = baf[0]*(args->nbins-1);
args->dist[idist].yvals[bin]++; // the distribution
}
free(baf);
bcf_sr_destroy(files);
for (idist=0; idist<args->ndist; idist++)
{
#if 0
int j;
for (j=0; j<args->nbins; j++)
{
double x = args->dist[idist].xvals[j];
args->dist[idist].yvals[j] = exp(-(x-0.5)*(x-0.5)/1e-3);
}
#endif
init_dist(args, &args->dist[idist],args->verbose);
}
args->dat_fp = open_file(&args->dat_fname,"w","%s/dist.dat", args->output_dir);
fprintf(args->dat_fp, "# This file was produced by: bcftools polysomy(%s+htslib-%s), the command line was:\n", bcftools_version(),hts_version());
fprintf(args->dat_fp, "# \t bcftools %s ", args->argv[0]);
for (i=1; i<args->argc; i++)
fprintf(args->dat_fp, " %s",args->argv[i]);
fprintf(args->dat_fp,"\n#\n");
fprintf(args->dat_fp,"# DIST\t[2]Chrom\t[3]BAF\t[4]Normalized Count\n");
fprintf(args->dat_fp,"# FIT\t[2]Goodness of Fit\t[3]iFrom\t[4]iTo\t[5]The Fitted Function\n");
fprintf(args->dat_fp,"# CN\t[2]Chrom\t[3]Estimated Copy Number\t[4]Absolute fit deviation\n");
char *fname = NULL;
FILE *fp = open_file(&fname,"w","%s/dist.py", args->output_dir);
//-------- matplotlib script --------------
fprintf(fp,
"#!/usr/bin/env python\n"
"#\n"
"import matplotlib as mpl\n"
"mpl.use('Agg')\n"
"import matplotlib.pyplot as plt\n"
"import csv,sys,argparse\n"
"from math import exp\n"
"\n"
"outdir = '%s'\n"
"\n"
"def read_dat(dat,fit,cn):\n"
" csv.register_dialect('tab', delimiter='\t', quoting=csv.QUOTE_NONE)\n"
" with open(outdir+'/dist.dat', 'rb') as f:\n"
" reader = csv.reader(f, 'tab')\n"
" for row in reader:\n"
" if row[0][0]=='#': continue\n"
" type = row[0]\n"
" chr = row[1]\n"
" if type=='DIST':\n"
" if chr not in dat: dat[chr] = []\n"
" dat[chr].append(row)\n"
" elif type=='FIT':\n"
" if chr not in fit: fit[chr] = []\n"
" fit[chr].append(row)\n"
" elif type=='CN':\n"
" cn[chr] = row[2]\n"
"\n"
"def plot_dist(dat,fit,chr):\n"
" fig, ax = plt.subplots(1, 1, figsize=(7,5))\n"
" ax.plot([x[2] for x in dat[chr]],[x[3] for x in dat[chr]],'k-',label='Distribution')\n"
" if chr in fit:\n"
" for i in range(len(fit[chr])):\n"
" pfit = fit[chr][i]\n"
" exec('def xfit(x): return '+pfit[5])\n"
" istart = int(pfit[3])\n"
" iend = int(pfit[4])+1\n"
" vals = dat[chr][istart:iend]\n"
" args = {}\n"
" if i==0: args = {'label':'Target to Fit'}\n"
" ax.plot([x[2] for x in vals],[x[3] for x in vals],'r-',**args)\n"
" if i==0: args = {'label':'Best Fit'}\n"
" ax.plot([x[2] for x in vals],[xfit(float(x[2])) for x in vals],'g-',**args)\n"
" ax.set_title('BAF distribution, chr'+chr)\n"
" ax.set_xlabel('BAF')\n"
" ax.set_ylabel('Frequency')\n"
" ax.legend(loc='best',prop={'size':7},frameon=False)\n"
" plt.savefig(outdir+'/dist.chr'+chr+'.png')\n"
" plt.close()\n"
"\n"
"def plot_copy_number(cn):\n"
" fig, ax = plt.subplots(1, 1, figsize=(7,5))\n"
" xlabels = sorted(cn.keys())\n"
" xvals = range(len(xlabels))\n"
" yvals = [float(cn[x]) for x in xlabels]\n"
" ax.plot(xvals,yvals,'o',color='red')\n"
" for i in range(len(xvals)):\n"
" if yvals[i]==-1: ax.annotate('?', xy=(xvals[i],0.5),va='center',ha='center',color='red',fontweight='bold')\n"
" ax.tick_params(axis='both', which='major', labelsize=9)\n"
" ax.set_xticks(xvals)\n"
" ax.set_xticklabels(xlabels,rotation=45)\n"
" ax.set_xlim(-1,len(xlabels))\n"
" ax.set_ylim(0,5.0)\n"
" ax.set_yticks([1.0,2.0,3.0,4.0])\n"
" ax.set_xlabel('Chromosome')\n"
" ax.set_ylabel('Copy Number')\n"
" plt.savefig(outdir+'/copy-number.png')\n"
" plt.close()\n"
"\n"
"class myParser(argparse.ArgumentParser):\n"
" def error(self, message):\n"
" self.print_help()\n"
" sys.stderr.write('error: %%s\\n' %% message)\n"
" sys.exit(2)\n"
"\n"
"def main():\n"
" parser = myParser()\n"
" parser.add_argument('-a', '--all', action='store_true', help='Create all plots')\n"
" parser.add_argument('-c', '--copy-number', action='store_true', help='Create copy-number plot')\n"
" parser.add_argument('-d', '--distrib', metavar='CHR', help='Plot BAF distribution of a single chromosome')\n"
" args = parser.parse_args()\n"
" dat = {}; fit = {}; cn = {}\n"
" read_dat(dat,fit,cn)\n"
" if args.distrib!=None:\n"
" plot_dist(dat,fit,args.distrib)\n"
" if args.all:\n"
" for chr in dat: plot_dist(dat,fit,chr)\n"
" plot_copy_number(cn)\n"
" elif args.copy_number:\n"
" plot_copy_number(cn)\n"
" else:\n"
" for chr in dat: plot_dist(dat,fit,chr)\n"
"\n"
"if __name__ == '__main__':\n"
" main()\n",
args->output_dir);
//---------------------------------------
chmod(fname, S_IWUSR|S_IRUSR|S_IRGRP|S_IROTH|S_IXUSR|S_IXGRP|S_IXOTH);
free(fname);
fclose(fp);
}
//{{{ void print_query_result_offset(uint32_t *mask,
void print_query_result_offset(uint32_t *mask,
uint32_t mask_len,
uint32_t *vids,
struct gqt_query *q,
uint32_t **counts,
uint32_t *id_lens,
uint32_t *U_R,
uint32_t U_R_len,
char **id_query_list,
char **gt_query_list,
uint32_t num_qs,
uint32_t num_fields,
char *off_file_name,
char *source_file,
char *full_cmd)
{
struct off_file *off_f = open_off_file(off_file_name);
struct bcf_file bcf_f = init_bcf_file(source_file);
char *sample_names = NULL;
uint32_t i,j,k,line_idx,bytes, bit_i = 0;
int r;
for (i = 0; i < U_R_len; ++i) {
if (i == 0 )
r = asprintf(&sample_names,
"%s",
bcf_f.hdr->samples[U_R[i]]);
else
r = asprintf(&sample_names,
"%s,%s",
sample_names,
bcf_f.hdr->samples[U_R[i]]);
if (r == -1)
err(EX_OSERR, "asprintf error");
}
if (bcf_hdr_set_samples(bcf_f.hdr, sample_names, 0) != 0)
errx(EX_DATAERR, "Error setting samples: %s\n", source_file);
char *info_s;
for (i = 0; i < num_qs; i++) {
if ( q[i].variant_op == p_count ) {
r = asprintf(&info_s, "##INFO=<ID=GQT_%u,Number=1,Type=Integer,"
"Description=\"GQT count result from "
"phenotype:'%s' genotype:'%s'\">",
i, id_query_list[i], gt_query_list[i]);
if (r == -1) err(EX_OSERR, "asprintf error");
if (bcf_hdr_append(bcf_f.hdr, info_s) != 0)
errx(EX_DATAERR, "Error updating header: %s\n", source_file);
} else if ( q[i].variant_op == p_pct ) {
r = asprintf(&info_s, "##INFO=<ID=GQT_%u,Number=1,Type=Float,"
"Description=\"GQT percent result from "
"phenotype:'%s' genotype:'%s'\">",
i, id_query_list[i], gt_query_list[i]);
if (r == -1) err(EX_OSERR, "asprintf error");
if (bcf_hdr_append(bcf_f.hdr, info_s) != 0)
errx(EX_DATAERR, "Error updating header: %s\n", source_file);
} else if ( q[i].variant_op == p_maf ) {
r = asprintf(&info_s, "##INFO=<ID=GQT_%u,Number=1,Type=Float,"
"Description=\"GQT maf result from "
"phenotype:'%s' genotype:'%s'\">",
i, id_query_list[i], gt_query_list[i]);
if (bcf_hdr_append(bcf_f.hdr, info_s) != 0)
errx(EX_DATAERR, "Error updating header: %s\n", source_file);
}
}
r = asprintf(&info_s, "##%s_queryVersion=%s", PROGRAM_NAME, VERSION);
if (r == -1) err(EX_OSERR, "asprintf error");
if (bcf_hdr_append(bcf_f.hdr, info_s) != 0)
errx(EX_DATAERR, "Error updating header: %s\n", source_file);
r = asprintf(&info_s, "##%s_queryCommand=%s", PROGRAM_NAME, full_cmd);
if (r == -1) err(EX_OSERR, "asprintf error");
if (bcf_hdr_append(bcf_f.hdr, info_s) != 0)
errx(EX_DATAERR, "Error updating header: %s\n", source_file);
htsFile *out_f = hts_open("-","w");
if ( !out_f )
err(EX_DATAERR, "Could open output file");
bcf_hdr_write(out_f, bcf_f.hdr);
bcf_f.line = bcf_init1();
for (i=0; i < mask_len; ++i) {
bytes = mask[i];
if (bytes == 0)
continue; /* skip a bunch of ops if you can */
for (j=0; j < 32; j++) {
if (bytes & 1 << (31 - j)) {
line_idx = i*32+j;
r = goto_bcf_line(&bcf_f, off_f, line_idx);
if (r == -1)
err(EX_NOINPUT,
"Error seeking file '%s'", bcf_f.file_name);
r = get_bcf_line(&bcf_f);
if (r == -1)
err(EX_NOINPUT,
"Error reading file '%s'", bcf_f.file_name);
for (k=0; k < num_qs; k++) {
r = asprintf(&info_s, "GQT_%u", k);
if (r == -1)
err(EX_OSERR, "asprintf error");
if ( q[k].variant_op == p_count ) {
int32_t v = counts[k][line_idx];
if (bcf_update_info_int32(bcf_f.hdr,
bcf_f.line,
info_s,
&v,
1) != 0)
errx(EX_DATAERR,
"Error adding to info field: %s\n",
bcf_f.file_name);
} else if (q[k].variant_op == p_pct) {
float v = ((float)counts[k][line_idx])/
((float) id_lens[k]);
if (bcf_update_info_float(bcf_f.hdr,
bcf_f.line,
info_s,
&v,
1) != 0)
errx(EX_DATAERR,
"Error adding to info field: %s\n",
bcf_f.file_name);
} else if (q[k].variant_op == p_maf) {
float v = ((float)counts[k][line_idx])/
(((float) id_lens[k])*2.0);
if (bcf_update_info_float(bcf_f.hdr,
bcf_f.line,
info_s,
&v,
1) != 0)
errx(EX_DATAERR,
"Error adding to info field: %s\n",
bcf_f.file_name);
}
}
bcf_write(out_f, bcf_f.hdr, bcf_f.line);
}
bit_i++;
if (bit_i == num_fields)
break;
}
if (bit_i == num_fields)
break;
}
hts_close(out_f);
destroy_off_file(off_f);
}
static void init_data(args_t *args)
{
args->prev_rid = args->skip_rid = -1;
args->hdr = args->files->readers[0].header;
if ( !args->sample )
{
if ( bcf_hdr_nsamples(args->hdr)>1 ) error("Missing the option -s, --sample\n");
args->sample = strdup(args->hdr->samples[0]);
}
if ( !bcf_hdr_nsamples(args->hdr) ) error("No samples in the VCF?\n");
// Set samples
kstring_t str = {0,0,0};
if ( args->estimate_AF && strcmp("-",args->estimate_AF) )
{
int i, n;
char **smpls = hts_readlist(args->estimate_AF, 1, &n);
// Make sure the query sample is included
for (i=0; i<n; i++)
if ( !strcmp(args->sample,smpls[i]) ) break;
// Add the query sample if not present
if ( i!=n ) kputs(args->sample, &str);
for (i=0; i<n; i++)
{
if ( str.l ) kputc(',', &str);
kputs(smpls[i], &str);
free(smpls[i]);
}
free(smpls);
}
else if ( !args->estimate_AF )
kputs(args->sample, &str);
if ( str.l )
{
int ret = bcf_hdr_set_samples(args->hdr, str.s, 0);
if ( ret<0 ) error("Error parsing the list of samples: %s\n", str.s);
else if ( ret>0 ) error("The %d-th sample not found in the VCF\n", ret);
}
if ( args->af_tag )
if ( !bcf_hdr_idinfo_exists(args->hdr,BCF_HL_INFO,bcf_hdr_id2int(args->hdr,BCF_DT_ID,args->af_tag)) )
error("No such INFO tag in the VCF: %s\n", args->af_tag);
args->nsmpl = bcf_hdr_nsamples(args->hdr);
args->ismpl = bcf_hdr_id2int(args->hdr, BCF_DT_SAMPLE, args->sample);
free(str.s);
int i;
for (i=0; i<256; i++) args->pl2p[i] = pow(10., -i/10.);
// Init transition matrix and HMM
double tprob[4];
MAT(tprob,2,STATE_HW,STATE_HW) = 1 - args->t2AZ;
MAT(tprob,2,STATE_HW,STATE_AZ) = args->t2HW;
MAT(tprob,2,STATE_AZ,STATE_HW) = args->t2AZ;
MAT(tprob,2,STATE_AZ,STATE_AZ) = 1 - args->t2HW;
if ( args->genmap_fname )
{
args->hmm = hmm_init(2, tprob, 0);
hmm_set_tprob_func(args->hmm, set_tprob_genmap, args);
}
else if ( args->rec_rate > 0 )
{
args->hmm = hmm_init(2, tprob, 0);
hmm_set_tprob_func(args->hmm, set_tprob_recrate, args);
}
else
args->hmm = hmm_init(2, tprob, 10000);
// print header
printf("# This file was produced by: bcftools roh(%s+htslib-%s)\n", bcftools_version(),hts_version());
printf("# The command line was:\tbcftools %s", args->argv[0]);
for (i=1; i<args->argc; i++)
printf(" %s",args->argv[i]);
printf("\n#\n");
printf("# [1]Chromosome\t[2]Position\t[3]State (0:HW, 1:AZ)\t[4]Quality\n");
}
int init(int argc, char **argv, bcf_hdr_t *in, bcf_hdr_t *out)
{
char *trio_samples = NULL, *unrelated_samples = NULL;
memset(&args,0,sizeof(args_t));
args.prev_rid = -1;
args.hdr = in;
args.pij = 2e-8;
args.pgt_err = 1e-9;
static struct option loptions[] =
{
{"prefix",1,0,'p'},
{"trio",1,0,'t'},
{"unrelated",1,0,'u'},
{0,0,0,0}
};
int c;
while ((c = getopt_long(argc, argv, "?ht:u:p:",loptions,NULL)) >= 0)
{
switch (c)
{
case 'p': args.prefix = optarg; break;
case 't': trio_samples = optarg; break;
case 'u': unrelated_samples = optarg; break;
case 'h':
case '?':
default: error("%s", usage()); break;
}
}
if ( optind != argc ) error(usage());
if ( trio_samples && unrelated_samples ) error("Expected only one of the -t/-u options\n");
if ( !trio_samples && !unrelated_samples ) error("Expected one of the -t/-u options\n");
if ( !args.prefix ) error("Expected the -p option\n");
int ret = bcf_hdr_set_samples(args.hdr, trio_samples ? trio_samples : unrelated_samples, 0);
if ( ret<0 ) error("Could not parse samples: %s\n", trio_samples ? trio_samples : unrelated_samples);
else if ( ret>0 ) error("%d-th sample not found: %s\n", ret,trio_samples ? trio_samples : unrelated_samples);
if ( trio_samples )
{
int i,n = 0;
char **list = hts_readlist(trio_samples, 0, &n);
if ( n!=3 ) error("Expected three sample names with -t\n");
args.imother = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[0]);
args.ifather = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[1]);
args.ichild = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[2]);
for (i=0; i<n; i++) free(list[i]);
free(list);
args.set_observed_prob = set_observed_prob_trio;
args.mode = C_TRIO;
init_hmm_trio(&args);
}
else
{
int i,n = 0;
char **list = hts_readlist(unrelated_samples, 0, &n);
if ( n!=2 ) error("Expected two sample names with -u\n");
args.isample = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[0]);
args.jsample = bcf_hdr_id2int(args.hdr, BCF_DT_SAMPLE, list[1]);
for (i=0; i<n; i++) free(list[i]);
free(list);
args.set_observed_prob = set_observed_prob_unrelated;
args.mode = C_UNRL;
init_hmm_unrelated(&args);
}
return 1;
}
