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));
#define MPLP_PRINT_POS 0x4000
mplp.max_mq = 60;
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;
while ((c = getopt(argc, argv, "Agf:r:l:M:q:Q:uaRC:BDSd:L:b:P:o:e:h:Im:F:EG:6Os")) >= 0) {
switch (c) {
case 'f':
mplp.fai = fai_load(optarg);
if (mplp.fai == 0) return 1;
break;
case 'd': mplp.max_depth = atoi(optarg); break;
case 'r': mplp.reg = strdup(optarg); break;
case 'l': mplp.bed = bed_read(optarg); break;
case 'P': mplp.pl_list = strdup(optarg); break;
case 'g': mplp.flag |= MPLP_GLF; break;
case 'u': mplp.flag |= MPLP_NO_COMP | MPLP_GLF; break;
case 'a': mplp.flag |= MPLP_NO_ORPHAN | MPLP_REALN; break;
case 'B': mplp.flag &= ~MPLP_REALN; break;
case 'D': mplp.flag |= MPLP_FMT_DP; break;
case 'S': mplp.flag |= MPLP_FMT_SP; break;
case 'I': mplp.flag |= MPLP_NO_INDEL; break;
case 'E': mplp.flag |= MPLP_EXT_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 'M': mplp.max_mq = 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': mplp.openQ = atoi(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 = bcf_str2id_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...
bcf_str2id_add(mplp.rghash, strdup(buf));
fclose(fp_rg);
}
break;
}
}
if (use_orphan) mplp.flag &= ~MPLP_NO_ORPHAN;
if (argc == 1) {
fprintf(stderr, "\n");
fprintf(stderr, "Usage: samtools mpileup [options] in1.bam [in2.bam [...]]\n\n");
fprintf(stderr, "Input options:\n\n");
fprintf(stderr, " -6 assume the quality is in the Illumina-1.3+ encoding\n");
fprintf(stderr, " -A count anomalous read pairs\n");
fprintf(stderr, " -B disable BAQ computation\n");
fprintf(stderr, " -b FILE list of input BAM files [null]\n");
fprintf(stderr, " -C INT parameter for adjusting mapQ; 0 to disable [0]\n");
fprintf(stderr, " -d INT max per-BAM depth to avoid excessive memory usage [%d]\n", mplp.max_depth);
fprintf(stderr, " -E extended BAQ for higher sensitivity but lower specificity\n");
fprintf(stderr, " -f FILE faidx indexed reference sequence file [null]\n");
fprintf(stderr, " -G FILE exclude read groups listed in FILE [null]\n");
fprintf(stderr, " -l FILE list of positions (chr pos) or regions (BED) [null]\n");
fprintf(stderr, " -M INT cap mapping quality at INT [%d]\n", mplp.max_mq);
fprintf(stderr, " -r STR region in which pileup is generated [null]\n");
fprintf(stderr, " -R ignore RG tags\n");
fprintf(stderr, " -q INT skip alignments with mapQ smaller than INT [%d]\n", mplp.min_mq);
fprintf(stderr, " -Q INT skip bases with baseQ/BAQ smaller than INT [%d]\n", mplp.min_baseQ);
fprintf(stderr, "\nOutput options:\n\n");
fprintf(stderr, " -D output per-sample DP in BCF (require -g/-u)\n");
fprintf(stderr, " -g generate BCF output (genotype likelihoods)\n");
fprintf(stderr, " -O output base positions on reads (disabled by -g/-u)\n");
fprintf(stderr, " -s output mapping quality (disabled by -g/-u)\n");
fprintf(stderr, " -S output per-sample strand bias P-value in BCF (require -g/-u)\n");
fprintf(stderr, " -u generate uncompress BCF output\n");
fprintf(stderr, "\nSNP/INDEL genotype likelihoods options (effective with `-g' or `-u'):\n\n");
fprintf(stderr, " -e INT Phred-scaled gap extension seq error probability [%d]\n", mplp.extQ);
fprintf(stderr, " -F FLOAT minimum fraction of gapped reads for candidates [%g]\n", mplp.min_frac);
fprintf(stderr, " -h INT coefficient for homopolymer errors [%d]\n", mplp.tandemQ);
fprintf(stderr, " -I do not perform indel calling\n");
fprintf(stderr, " -L INT max per-sample depth for INDEL calling [%d]\n", mplp.max_indel_depth);
fprintf(stderr, " -m INT minimum gapped reads for indel candidates [%d]\n", mplp.min_support);
fprintf(stderr, " -o INT Phred-scaled gap open sequencing error probability [%d]\n", mplp.openQ);
fprintf(stderr, " -P STR comma separated list of platforms for indels [all]\n");
fprintf(stderr, "\n");
fprintf(stderr, "Notes: Assuming diploid individuals.\n\n");
return 1;
}
if (file_list) {
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
mpileup(&mplp,nfiles,fn);
for (c=0; c<nfiles; c++) free(fn[c]);
free(fn);
} else mpileup(&mplp, argc - optind, argv + optind);
if (mplp.rghash) bcf_str2id_thorough_destroy(mplp.rghash);
free(mplp.reg); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);
if (mplp.bed) bed_destroy(mplp.bed);
return 0;
}
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;
}
int bam_mpileup(int argc, char *argv[])
{
int c;
const char *file_list = NULL;
char **fn = NULL;
int nfiles = 0, use_orphan = 0, noref = 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;
mplp.output_type = FT_VCF;
mplp.record_cmd_line = 1;
mplp.n_threads = 0;
mplp.bsmpl = bam_smpl_init();
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},
{"ignore-RG", no_argument, NULL, 5},
{"ignore-rg", no_argument, NULL, 5},
{"gvcf", required_argument, NULL, 'g'},
{"no-reference", no_argument, NULL, 7},
{"no-version", no_argument, NULL, 8},
{"threads",required_argument,NULL,9},
{"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'},
{"read-groups", required_argument, NULL, 'G'},
{"region", required_argument, NULL, 'r'},
{"regions", required_argument, NULL, 'r'},
{"regions-file", required_argument, NULL, 'R'},
{"targets", required_argument, NULL, 't'},
{"targets-file", required_argument, NULL, 'T'},
{"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'},
{"output-type", required_argument, NULL, 'O'},
{"samples", required_argument, NULL, 's'},
{"samples-file", required_argument, NULL, 'S'},
{"annotate", required_argument, NULL, 'a'},
{"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, "Ag:f:r:R:q:Q:C:Bd:L:b:P:po:e:h:Im:F:EG:6O:xa:s:S:t:T:",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 5 : bam_smpl_ignore_readgroups(mplp.bsmpl); break;
case 'g':
mplp.gvcf = gvcf_init(optarg);
if ( !mplp.gvcf ) error("Could not parse: --gvcf %s\n", optarg);
break;
case 'f':
mplp.fai = fai_load(optarg);
if (mplp.fai == NULL) return 1;
mplp.fai_fname = optarg;
break;
case 7 : noref = 1; break;
case 8 : mplp.record_cmd_line = 0; break;
case 9 : mplp.n_threads = strtol(optarg, 0, 0); break;
case 'd': mplp.max_depth = atoi(optarg); break;
case 'r': mplp.reg_fname = strdup(optarg); break;
case 'R': mplp.reg_fname = strdup(optarg); mplp.reg_is_file = 1; break;
case 't':
// 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.
if ( optarg[0]=='^' ) optarg++;
else mplp.bed_logic = 1;
mplp.bed = regidx_init(NULL,regidx_parse_reg,NULL,0,NULL);
mplp.bed_itr = regitr_init(mplp.bed);
if ( regidx_insert_list(mplp.bed,optarg,',') !=0 )
{
fprintf(stderr,"Could not parse the targets: %s\n", optarg);
exit(EXIT_FAILURE);
}
break;
case 'T':
if ( optarg[0]=='^' ) optarg++;
else mplp.bed_logic = 1;
mplp.bed = regidx_init(optarg,NULL,NULL,0,NULL);
if (!mplp.bed) { fprintf(stderr, "bcftools mpileup: 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 'B': mplp.flag &= ~MPLP_REALN; 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 's': if ( bam_smpl_add_samples(mplp.bsmpl,optarg,0)<0 ) error("Could not read samples: %s\n",optarg); break;
case 'S': if ( bam_smpl_add_samples(mplp.bsmpl,optarg,1)<0 ) error("Could not read samples: %s\n",optarg); break;
case 'O':
switch (optarg[0]) {
case 'b': mplp.output_type = FT_BCF_GZ; break;
case 'u': mplp.output_type = FT_BCF; break;
case 'z': mplp.output_type = FT_VCF_GZ; break;
case 'v': mplp.output_type = FT_VCF; break;
default: error("[error] The option \"-O\" changed meaning when mpileup moved to bcftools. Did you mean: \"bcftools mpileup --output-type\" or \"samtools mpileup --output-BP\"?\n");
}
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': bam_smpl_add_readgroups(mplp.bsmpl, optarg, 1); break;
case 'a':
if (optarg[0]=='?') {
list_annotations(stderr);
return 1;
}
mplp.fmt_flag |= parse_format_flag(optarg);
break;
default:
fprintf(stderr,"Invalid option: '%c'\n", c);
return 1;
}
}
if ( mplp.gvcf && !(mplp.fmt_flag&B2B_FMT_DP) )
{
fprintf(stderr,"[warning] The -t DP option is required with --gvcf, switching on.\n");
mplp.fmt_flag |= B2B_FMT_DP;
}
if ( mplp.flag&(MPLP_BCF|MPLP_VCF|MPLP_NO_COMP) )
{
if ( mplp.flag&MPLP_VCF )
{
if ( mplp.flag&MPLP_NO_COMP ) mplp.output_type = FT_VCF;
else mplp.output_type = FT_VCF_GZ;
}
else if ( mplp.flag&MPLP_BCF )
{
if ( mplp.flag&MPLP_NO_COMP ) mplp.output_type = FT_BCF;
else mplp.output_type = FT_BCF_GZ;
}
}
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;
}
if (!mplp.fai && !noref) {
fprintf(stderr,"Error: mpileup requires the --fasta-ref option by default; use --no-reference to run without a fasta reference\n");
return 1;
}
int ret,i;
if (file_list)
{
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
mplp.files = fn;
mplp.nfiles = nfiles;
}
else
{
mplp.nfiles = argc - optind;
mplp.files = (char**) malloc(mplp.nfiles*sizeof(char*));
for (i=0; i<mplp.nfiles; i++) mplp.files[i] = strdup(argv[optind+i]);
}
ret = mpileup(&mplp);
for (i=0; i<mplp.nfiles; i++) free(mplp.files[i]);
free(mplp.files);
free(mplp.reg_fname); free(mplp.pl_list);
if (mplp.fai) fai_destroy(mplp.fai);
if (mplp.bed) regidx_destroy(mplp.bed);
if (mplp.bed_itr) regitr_destroy(mplp.bed_itr);
if (mplp.reg) regidx_destroy(mplp.reg);
bam_smpl_destroy(mplp.bsmpl);
return ret;
}
int
main_uniq(int argc, char *argv[])
{
int c, i;
char *bam_file = NULL;
char *vcf_in = NULL; /* - == stdout */
char *vcf_out = NULL; /* - == stdout */
mplp_conf_t mplp_conf;
uniq_conf_t uniq_conf;
void (*plp_proc_func)(const plp_col_t*, void*);
int rc = 0;
var_t **vars = NULL;
int num_vars = 0;
char *vcf_header = NULL;
static int use_det_lim = 0;
static int use_orphan = 0;
static int output_all = 0;
static int is_somatic = 0;
/* default uniq options */
memset(&uniq_conf, 0, sizeof(uniq_conf_t));
uniq_conf.uni_freq = DEFAULT_UNI_FREQ;
uniq_conf.use_det_lim = 0;
uniq_conf.uniq_filter.mtc_type = MTC_FDR;
uniq_conf.uniq_filter.alpha = 0.001;
/* default pileup options */
memset(&mplp_conf, 0, sizeof(mplp_conf_t));
mplp_conf.max_mq = DEFAULT_MAX_MQ;
mplp_conf.min_mq = 1;
mplp_conf.min_plp_bq = DEFAULT_MIN_PLP_BQ;
mplp_conf.max_depth = DEFAULT_MAX_PLP_DEPTH;
mplp_conf.flag = MPLP_NO_ORPHAN;
/* keep in sync with long_opts_str and usage
*
* getopt is a pain in the whole when it comes to syncing of long
* and short args and usage. check out gopt, libcfu...
*/
while (1) {
static struct option long_opts[] = {
/* see usage sync */
{"help", no_argument, NULL, 'h'},
{"verbose", no_argument, &verbose, 1},
{"debug", no_argument, &debug, 1},
{"use-det-lim", no_argument, &use_det_lim, 1},
{"use-orphan", no_argument, &use_orphan, 1},
{"output-all", no_argument, &output_all, 1},
{"is-somatic", no_argument, &is_somatic, 1},
{"vcf-in", required_argument, NULL, 'v'},
{"vcf-out", required_argument, NULL, 'o'},
{"uni-freq", required_argument, NULL, 'f'},
{"uniq-thresh", required_argument, NULL, 't'},
{"uniq-mtc", required_argument, NULL, 'm'},
{"uniq-alpha", required_argument, NULL, 'a'},
{"uniq-ntests", required_argument, NULL, 'n'},
{0, 0, 0, 0} /* sentinel */
};
/* keep in sync with long_opts and usage */
static const char *long_opts_str = "hv:o:f:t:m:a:n:";
/* getopt_long stores the option index here. */
int long_opts_index = 0;
c = getopt_long(argc-1, argv+1, /* skipping 'lofreq', just leaving 'command', i.e. call */
long_opts_str, long_opts, & long_opts_index);
if (c == -1) {
break;
}
switch (c) {
/* keep in sync with long_opts etc */
case 'h':
usage(& uniq_conf);
return 0;
case 'v':
if (0 != strcmp(optarg, "-")) {
if (! file_exists(optarg)) {
LOG_FATAL("Input file '%s' does not exist. Exiting...\n", optarg);
return 1;
}
}
vcf_in = strdup(optarg);
break;
case 'o':
if (0 != strcmp(optarg, "-")) {
if (file_exists(optarg)) {
LOG_FATAL("Cowardly refusing to overwrite file '%s'. Exiting...\n", optarg);
return 1;
}
}
vcf_out = strdup(optarg);
break;
case 'f':
uniq_conf.uni_freq = strtof(optarg, (char **)NULL); /* atof */
if (uniq_conf.uni_freq<=0) {
LOG_WARN("%s\n", "Ignoring uni-freq option");
}
if (uniq_conf.uni_freq>1.0) {
LOG_FATAL("%s\n", "Value for uni-freq has to be <1.0");
return 1;
}
break;
case 't':
if (! isdigit(optarg[0])) {
LOG_FATAL("Non-numeric argument provided: %s\n", optarg);
return -1;
}
uniq_conf.uniq_filter.thresh = atoi(optarg);
uniq_conf.uniq_filter.mtc_type = MTC_NONE;
break;
case 'm':
uniq_conf.uniq_filter.mtc_type = mtc_str_to_type(optarg);
if (-1 == uniq_conf.uniq_filter.mtc_type) {
LOG_FATAL("Unknown multiple testing correction type '%s' for snv quality filtering\n", optarg);
return -1;
}
break;
case 'a':
if (! isdigit(optarg[0])) {
LOG_FATAL("Non-numeric argument provided: %s\n", optarg);
return -1;
}
uniq_conf.uniq_filter.alpha = strtof(optarg, NULL);
break;
case 'n':
if (! isdigit(optarg[0])) {
LOG_FATAL("Non-numeric argument provided: %s\n", optarg);
return -1;
}
uniq_conf.uniq_filter.ntests = atol(optarg);
break;
case '?':
LOG_FATAL("%s\n", "unrecognized arguments found. Exiting...\n");
return 1;
default:
break;
}
}
if (use_orphan) {
mplp_conf.flag &= ~MPLP_NO_ORPHAN;
}
if (debug) {
dump_mplp_conf(& mplp_conf, stderr);
}
uniq_conf.output_all = output_all;
uniq_conf.use_det_lim = use_det_lim;
#if DEBUG
LOG_DEBUG("uniq_conf.uniq_filter.thresh = %d\n", uniq_conf.uniq_filter.thresh);
LOG_DEBUG("uniq_conf.uniq_filter.mtc_type = %d\n", uniq_conf.uniq_filter.mtc_type);
LOG_DEBUG("uniq_conf.uniq_filter.alpha = %f\n", uniq_conf.uniq_filter.alpha);
LOG_DEBUG("uniq_conf.uniq_filter.ntests = %d\n", uniq_conf.uniq_filter.ntests);
#endif
if (uniq_conf.uniq_filter.thresh && uniq_conf.uniq_filter.mtc_type != MTC_NONE) {
LOG_FATAL("%s\n", "Can't use fixed Unique quality threshold *and* multiple testing correction.");
return 1;
}
if (argc == 2) {
fprintf(stderr, "\n");
usage(& uniq_conf);
return 1;
}
if (1 != argc - optind - 1) {
fprintf(stderr, "Need exactly one BAM file as last argument\n");
return 1;
}
bam_file = (argv + optind + 1)[0];
if (! file_exists(bam_file)) {
LOG_FATAL("BAM file %s does not exist. Exiting...\n", bam_file);
return -1;
}
if (! vcf_in) {
#if 0
vcf_in = malloc(2 * sizeof(char));
strcpy(vcf_in, "-");
#else
LOG_FATAL("%s\n", "No input vcf specified. Exiting...");
return -1;
#endif
}
if (! vcf_out) {
vcf_out = malloc(2 * sizeof(char));
strcpy(vcf_out, "-");
}
if (vcf_file_open(& uniq_conf.vcf_in, vcf_in,
HAS_GZIP_EXT(vcf_in), 'r')) {
LOG_ERROR("Couldn't open %s\n", vcf_in);
return 1;
}
if (vcf_file_open(& uniq_conf.vcf_out, vcf_out,
HAS_GZIP_EXT(vcf_out), 'w')) {
LOG_ERROR("Couldn't open %s\n", vcf_out);
return 1;
}
if (0 != vcf_parse_header(&vcf_header, & uniq_conf.vcf_in)) {
LOG_WARN("%s\n", "vcf_parse_header() failed. trying to rewind to start...");
if (vcf_file_seek(& uniq_conf.vcf_in, 0, SEEK_SET)) {
LOG_FATAL("%s\n", "Couldn't rewind file to parse variants"
" after header parsing failed");
return 1;
}
} else {
vcf_header_add(&vcf_header, "##INFO=<ID=UNIQ,Number=0,Type=Flag,Description=\"Unique, i.e. not detectable in paired sample\">\n");
vcf_header_add(&vcf_header, "##INFO=<ID=UQ,Number=1,Type=Integer,Description=\"Phred-scaled uniq score at this position\">\n");
if (is_somatic) {
vcf_header_add(&vcf_header, "##INFO=<ID=SOMATIC,Number=0,Type=Flag,Description=\"Somatic event\">\n");
}
if (! uniq_conf.use_det_lim) {
char full_filter_str[FILTER_STRSIZE];
if (uniq_conf.uniq_filter.thresh > 0) {
snprintf(uniq_conf.uniq_filter.id, FILTER_ID_STRSIZE, "min_uq_%d", uniq_conf.uniq_filter.thresh);
snprintf(full_filter_str, FILTER_STRSIZE,
"##FILTER=<ID=%s,Description=\"Minimum Uniq Phred %d\">\n",
uniq_conf.uniq_filter.id, uniq_conf.uniq_filter.thresh);
vcf_header_add(&vcf_header, full_filter_str);
} else if (uniq_conf.uniq_filter.mtc_type != MTC_NONE) {
char buf[64];
mtc_str(buf, uniq_conf.uniq_filter.mtc_type);
snprintf(uniq_conf.uniq_filter.id, FILTER_ID_STRSIZE, "uq_%s", buf);
snprintf(full_filter_str, FILTER_STRSIZE,
"##FILTER=<ID=%s,Description=\"Uniq Multiple Testing Correction: %s corr. pvalue < %f\">\n",
uniq_conf.uniq_filter.id, buf, uniq_conf.uniq_filter.alpha);
vcf_header_add(& vcf_header, full_filter_str);
}
}
vcf_write_header(& uniq_conf.vcf_out, vcf_header);
free(vcf_header);
}
num_vars = vcf_parse_vars(&vars, & uniq_conf.vcf_in, 1);
if (0 == num_vars) {
LOG_WARN("%s\n", "Didn't find any variants in input");
goto clean_and_exit;
}
if (! uniq_conf.uniq_filter.ntests) {
uniq_conf.uniq_filter.ntests = num_vars;
}
plp_proc_func = &uniq_snv;
for (i=0; i<num_vars; i++) {
char reg_buf[BUF_SIZE];
if (i%100==0) {
LOG_VERBOSE("Processing variant %d of %d\n", i+1, num_vars);
}
uniq_conf.var = vars[i];
snprintf(reg_buf, BUF_SIZE, "%s:%ld-%ld",
vars[i]->chrom, vars[i]->pos+1, vars[i]->pos+1);
mplp_conf.reg = strdup(reg_buf);
LOG_DEBUG("pileup for var no %d at %s %d\n",
i+1, uniq_conf.var->chrom, uniq_conf.var->pos+1);
#ifdef DISABLE_INDELS
if (vcf_var_has_info_key(NULL, uniq_conf.var, "INDEL")) {
LOG_WARN("Skipping indel var at %s %d\n",
uniq_conf.var->chrom, uniq_conf.var->pos+1);
free(mplp_conf.reg);
mplp_conf.reg = NULL;
continue;
}
#endif
/* no need to check for filter because done by parse_vars */
rc = mpileup(&mplp_conf, plp_proc_func, (void*)&uniq_conf,
1, (const char **) argv + optind + 1);
if (uniq_conf.uniq_filter.thresh) {
apply_uniq_threshold(uniq_conf.var, & uniq_conf.uniq_filter);
}
free(mplp_conf.reg);
mplp_conf.reg = NULL;
}
uniq_conf.var = NULL;/* just be sure to not use it accidentally again */
/* print whatever we've got. there's no UQ to test or we
* are supposed to print all
*/
if (uniq_conf.use_det_lim) {
for (i=0; i<num_vars; i++) {
var_t *var = vars[i];
vcf_write_var(& uniq_conf.vcf_out, var);
}
/* all done */
goto clean_and_exit;
}
if (uniq_conf.uniq_filter.mtc_type != MTC_NONE) {
if (apply_uniq_filter_mtc(& uniq_conf.uniq_filter, vars, num_vars)) {
LOG_FATAL("%s\n", "Multiple testing correction on uniq pvalues failed");
return -1;
}
}
for (i=0; i<num_vars; i++) {
var_t *var = vars[i];
if (VCF_VAR_PASSES(var) || uniq_conf.output_all) {
vcf_write_var(& uniq_conf.vcf_out, var);
}
}
clean_and_exit:
vcf_file_close(& uniq_conf.vcf_in);
vcf_file_close(& uniq_conf.vcf_out);
for (i=0; i<num_vars; i++) {
vcf_free_var(& vars[i]);
}
free(vars);
free(vcf_in);
free(vcf_out);
if (0==rc) {
LOG_VERBOSE("%s\n", "Successful exit.");
}
/* LOG_FIXME("%s\n", "allow user setting of -S and -J. Currently just using default") */
return rc;
}