static void copy_check_alignment(const char *infname, const char *informat,
const char *outfname, const char *outmode, const char *outref)
{
samFile *in = sam_open(infname, "r");
samFile *out = sam_open(outfname, outmode);
bam1_t *aln = bam_init1();
bam_hdr_t *header = NULL;
int res;
if (!in) {
fail("couldn't open %s", infname);
goto err;
}
if (!out) {
fail("couldn't open %s with mode %s", outfname, outmode);
goto err;
}
if (!aln) {
fail("bam_init1() failed");
goto err;
}
if (outref) {
if (hts_set_opt(out, CRAM_OPT_REFERENCE, outref) < 0) {
fail("setting reference %s for %s", outref, outfname);
goto err;
}
}
header = sam_hdr_read(in);
if (!header) {
fail("reading header from %s", infname);
goto err;
}
if (sam_hdr_write(out, header) < 0) fail("writing headers to %s", outfname);
while ((res = sam_read1(in, header, aln)) >= 0) {
int mod4 = ((intptr_t) bam_get_cigar(aln)) % 4;
if (mod4 != 0)
fail("%s CIGAR not 4-byte aligned; offset is 4k+%d for \"%s\"",
informat, mod4, bam_get_qname(aln));
if (sam_write1(out, header, aln) < 0) fail("writing to %s", outfname);
}
if (res < -1) {
fail("failed to read alignment from %s", infname);
}
err:
bam_destroy1(aln);
bam_hdr_destroy(header);
if (in) sam_close(in);
if (out) sam_close(out);
}
int bam_idxstats(int argc, char *argv[])
{
hts_idx_t* idx;
bam_hdr_t* header;
samFile* fp;
if (argc < 2) {
fprintf(pysamerr, "Usage: samtools idxstats <in.bam>\n");
return 1;
}
fp = sam_open(argv[1], "r");
if (fp == NULL) { fprintf(pysamerr, "[%s] fail to open BAM.\n", __func__); return 1; }
header = sam_hdr_read(fp);
idx = sam_index_load(fp, argv[1]);
if (idx == NULL) { fprintf(pysamerr, "[%s] fail to load the index.\n", __func__); return 1; }
int i;
for (i = 0; i < header->n_targets; ++i) {
// Print out contig name and length
printf("%s\t%d", header->target_name[i], header->target_len[i]);
// Now fetch info about it from the meta bin
uint64_t u, v;
hts_idx_get_stat(idx, i, &u, &v);
printf("\t%" PRIu64 "\t%" PRIu64 "\n", u, v);
}
// Dump information about unmapped reads
printf("*\t0\t0\t%" PRIu64 "\n", hts_idx_get_n_no_coor(idx));
bam_hdr_destroy(header);
hts_idx_destroy(idx);
sam_close(fp);
return 0;
}
// remember to clean up with bam_destroy1(b);
bam1_t* alignment_to_bam(const string& sam_header,
const Alignment& alignment,
const string& refseq,
const int32_t refpos,
const string& cigar,
const string& mateseq,
const int32_t matepos,
const int32_t tlen) {
assert(!sam_header.empty());
string sam_file = "data:" + sam_header + alignment_to_sam(alignment, refseq, refpos, cigar, mateseq, matepos, tlen);
const char* sam = sam_file.c_str();
samFile *in = sam_open(sam, "r");
bam_hdr_t *header = sam_hdr_read(in);
bam1_t *aln = bam_init1();
if (sam_read1(in, header, aln) >= 0) {
bam_hdr_destroy(header);
sam_close(in); // clean up
return aln;
} else {
cerr << "[vg::alignment] Failure to parse SAM record" << endl
<< sam << endl;
exit(1);
}
}
BamFilePrivate(const std::string& fn)
: filename_(fn)
, firstAlignmentOffset_(-1)
{
// ensure we've updated htslib verbosity with requested verbosity here
hts_verbose = ( PacBio::BAM::HtslibVerbosity == -1 ? 0 : PacBio::BAM::HtslibVerbosity);
// attempt open
auto f = RawOpen();
#if !defined (PBBAM_NO_CHECK_EOF) || PBBAM_AUTOVALIDATE
// sanity check on file
const int eofCheck = bgzf_check_EOF(f->fp.bgzf);
if (eofCheck <= 0 ) {
// 1: EOF present & correct
// 2: not seekable (e.g. reading from stdin)
// 0: EOF absent
// -1: some other error
std::stringstream e;
if (eofCheck == 0)
e << fn << " : is missing EOF block" << std::endl;
else
e << fn << " : unknown error while checking EOF block" << std::endl;
throw std::runtime_error(e.str());
}
#endif
// attempt fetch header
std::unique_ptr<bam_hdr_t, internal::HtslibHeaderDeleter> hdr(sam_hdr_read(f.get()));
header_ = internal::BamHeaderMemory::FromRawData(hdr.get());
// cache first alignment offset
firstAlignmentOffset_ = bgzf_tell(f->fp.bgzf);
}
bufReader initBufReader2(const char*fname,int doCheck,char *fai_fname){
bufReader ret;
ret.fn = strdup(fname);
int newlen=strlen(fname);//<-just to avoid valgrind -O3 uninitialized warning
ret.fp = openBAM(ret.fn,doCheck);
if (fai_fname && hts_set_fai_filename(ret.fp, fai_fname) != 0) {
fprintf(stderr, "[%s] failed to process %s\n",
__func__, fai_fname);
exit(EXIT_FAILURE);
}
ret.isEOF =0;
ret.itr=NULL;
ret.idx=NULL;
ret.hdr = sam_hdr_read(ret.fp);
if(strlen(ret.hdr->text)==0){
fprintf(stderr,"\t-> No header information could be found for BAM/CRAM file: \'%s\' will exit\n",fname);
exit(1);
}
checkIfSorted(ret.hdr->text);
if(ret.hdr==NULL) {
fprintf(stderr, "[main_samview] fail to read the header from \"%s\".\n", ret.fn);
exit(0);
}
return ret;
}
int bam_access_openhts(char *hts_file, char *ref_file){
assert(hts_file != NULL);
//Assign memory for the file name etc holding struct
fholder = malloc(sizeof(file_holder));
check_mem(fholder);
//Beginning and end of tmp struct for bam access
fholder->beg = 0; fholder->end = 0x7fffffff; // The max 32 bit integer.
//Open a file for read from compressed bam.
fholder->in = hts_open(hts_file, "r");
check(fholder->in != 0,"HTS file %s failed to open.",hts_file);
fholder->idx = sam_index_load(fholder->in,hts_file);
check(fholder->idx != 0,"HTS index file %s failed to open.",hts_file);
if(ref_file){
hts_set_fai_filename(fholder->in, ref_file);
}else{
if(fholder->in->format.format == cram) log_warn("No reference file provided for a cram input file, if the reference described in the cram header can't be located this script may fail.");
}
//Check for generic header read method.
fholder->head = sam_hdr_read(fholder->in);
return 0;
error:
if(fholder->in) hts_close(fholder->in);
if(fholder) free(fholder);
return -1;
}
BAMOrderedReader::BAMOrderedReader(std::string bam_file, std::vector<GenomeInterval>& intervals)
:bam_file(bam_file), intervals(intervals), sam(0), hdr(0), idx(0), itr(0)
{
const char* fname = bam_file.c_str();
int len = strlen(fname);
if ( strcasecmp(".bam",fname+len-4) )
{
fprintf(stderr, "[%s:%d %s] Not a BAM file: %s\n", __FILE__, __LINE__, __FUNCTION__, bam_file.c_str());
exit(1);
}
sam = sam_open(bam_file.c_str(), "r");
hdr = sam_hdr_read(sam);
s = bam_init1();
idx = bam_index_load(bam_file.c_str());
if (idx==0)
{
fprintf(stderr, "[%s:%d %s] fail to load index for %s\n", __FILE__, __LINE__, __FUNCTION__, bam_file.c_str());
abort();
}
else
{
index_loaded = true;
}
str = {0,0,0};
intervals_present = intervals.size()!=0;
interval_index = 0;
random_access_enabled = intervals_present && index_loaded;
};
static int aux_fields1(void)
{
static const char sam[] = "data:"
"@SQ\tSN:one\tLN:1000\n"
"@SQ\tSN:two\tLN:500\n"
"r1\t0\tone\t500\t20\t8M\t*\t0\t0\tATGCATGC\tqqqqqqqq\tXA:A:k\tXi:i:37\tXf:f:" xstr(PI) "\tXd:d:" xstr(E) "\tXZ:Z:" HELLO "\tXH:H:" BEEF "\tXB:B:c,-2,0,+2\tZZ:i:1000000\n";
// Canonical form of the alignment record above, as output by sam_format1()
static const char r1[] = "r1\t0\tone\t500\t20\t8M\t*\t0\t0\tATGCATGC\tqqqqqqqq\tXA:A:k\tXi:i:37\tXf:f:3.14159\tXd:d:2.71828\tXZ:Z:" HELLO "\tXH:H:" BEEF "\tXB:B:c,-2,0,2\tZZ:i:1000000";
samFile *in = sam_open(sam, "r");
bam_hdr_t *header = sam_hdr_read(in);
bam1_t *aln = bam_init1();
uint8_t *p;
uint32_t n;
kstring_t ks = { 0, 0, NULL };
if (sam_read1(in, header, aln) >= 0) {
if ((p = check_bam_aux_get(aln, "XA", 'A')) && bam_aux2A(p) != 'k')
fail("XA field is '%c', expected 'k'", bam_aux2A(p));
if ((p = check_bam_aux_get(aln, "Xi", 'C')) && bam_aux2i(p) != 37)
fail("Xi field is %d, expected 37", bam_aux2i(p));
if ((p = check_bam_aux_get(aln, "Xf", 'f')) && fabs(bam_aux2f(p) - PI) > 1E-6)
fail("Xf field is %.12f, expected pi", bam_aux2f(p));
if ((p = check_bam_aux_get(aln, "Xd", 'd')) && fabs(bam_aux2f(p) - E) > 1E-6)
fail("Xf field is %.12f, expected e", bam_aux2f(p));
if ((p = check_bam_aux_get(aln, "XZ", 'Z')) && strcmp(bam_aux2Z(p), HELLO) != 0)
fail("XZ field is \"%s\", expected \"%s\"", bam_aux2Z(p), HELLO);
if ((p = check_bam_aux_get(aln, "XH", 'H')) && strcmp(bam_aux2Z(p), BEEF) != 0)
fail("XH field is \"%s\", expected \"%s\"", bam_aux2Z(p), BEEF);
// TODO Invent and use bam_aux2B()
if ((p = check_bam_aux_get(aln, "XB", 'B')) && ! (memcmp(p, "Bc", 2) == 0 && (memcpy(&n, p+2, 4), n) == 3 && memcmp(p+6, "\xfe\x00\x02", 3) == 0))
fail("XB field is %c,..., expected c,-2,0,+2", p[1]);
if ((p = check_bam_aux_get(aln, "ZZ", 'I')) && bam_aux2i(p) != 1000000)
fail("ZZ field is %d, expected 1000000", bam_aux2i(p));
if (sam_format1(header, aln, &ks) < 0)
fail("can't format record");
if (strcmp(ks.s, r1) != 0)
fail("record formatted incorrectly: \"%s\"", ks.s);
free(ks.s);
}
else fail("can't read record");
bam_destroy1(aln);
bam_hdr_destroy(header);
sam_close(in);
return 1;
}
bool init() {
// Open files
file_iter = sam_open(file_name.c_str(), "rb", 0);
replace_iter = sam_open(replace_name.c_str(), "r", 0);
out_file = sam_open(out_name.c_str(), "wb", 0);
if (file_iter == NULL || replace_iter == NULL || out_file == NULL) {
return false;
}
// Iterate through BAM
file_header = sam_hdr_read(file_iter);
replace_header = sam_hdr_read(replace_iter);
build_translation();
return true;
}
bam_hdr_t* hts_file_header(string& filename, string& header) {
samFile *in = hts_open(filename.c_str(), "r");
if (in == NULL) {
cerr << "[vg::alignment] could not open " << filename << endl;
exit(1);
}
bam_hdr_t *hdr = sam_hdr_read(in);
header = hdr->text;
bam_hdr_destroy(hdr);
hts_close(in);
return hdr;
}
samfile_t *samopen(const char *fn, const char *mode, const void *aux)
{
// hts_open() is really sam_open(), except for #define games
samFile *hts_fp = hts_open(fn, mode);
if (hts_fp == NULL) return NULL;
samfile_t *fp = malloc(sizeof (samfile_t));
if (!fp) {
sam_close(hts_fp);
return NULL;
}
fp->file = hts_fp;
fp->x.bam = hts_fp->fp.bgzf;
if (strchr(mode, 'r')) {
if (aux) {
if (hts_set_fai_filename(fp->file, aux) != 0) {
sam_close(hts_fp);
free(fp);
return NULL;
}
}
fp->header = sam_hdr_read(fp->file); // samclose() will free this
if (fp->header == NULL) {
sam_close(hts_fp);
free(fp);
return NULL;
}
fp->is_write = 0;
if (fp->header->n_targets == 0 && bam_verbose >= 1)
fprintf(samtools_stderr, "[samopen] no @SQ lines in the header.\n");
}
else {
enum htsExactFormat fmt = hts_get_format(fp->file)->format;
fp->header = (bam_hdr_t *)aux; // For writing, we won't free it
fp->is_write = 1;
if (!(fmt == text_format || fmt == sam) || strchr(mode, 'h')) {
if (sam_hdr_write(fp->file, fp->header) < 0) {
if (bam_verbose >= 1)
fprintf(samtools_stderr, "[samopen] Couldn't write header\n");
sam_close(hts_fp);
free(fp);
return NULL;
}
}
}
return fp;
}
void bam_parser(opt_t *opt) {
samFile *in = sam_open(opt->in_name, "r");
if(in == NULL) die("bam_parser: fail to open file '%s'", opt->in_name);
// if output file exists but not force to overwrite
if(access(opt->out_name, F_OK)!=-1 && opt->f==false) die("bam_parser: %s exists, use opetion -f to overwrite", opt->out_name);
bam_hdr_t *header = sam_hdr_read(in);
bam1_t *aln = bam_init1();
int8_t *p;
int32_t n;
int ret;
while((ret=sam_read1(in, header, aln)) >= 0)
printf("name=%s\nflag=%d\nseq=%s\nqual=%s\nlane_id=%d\n", get_read_name(aln), aln->core.flag, get_sequence(aln), get_qualities(aln), get_lane_id(aln));
bam_destroy1(aln);
sam_close(in);
}
static void flanks_sam_open()
{
if(!futil_path_has_extension(sam_path, ".bam") &&
!futil_path_has_extension(sam_path, ".sam"))
{
cmd_print_usage("Mapped flanks is not .sam or .bam file: %s", sam_path);
}
bool isbam = futil_path_has_extension(sam_path, ".bam");
samfh = sam_open(sam_path, isbam ? "rb" : "rs");
if(samfh == NULL) die("Cannot open SAM/BAM %s", sam_path);
// Load BAM header
bam_header = sam_hdr_read(samfh);
bamentry = bam_init1();
}
bufReader initBufReader2(const char*fname){
bufReader ret;
ret.fn = strdup(fname);
int newlen=strlen(fname);//<-just to avoid valgrind -O3 uninitialized warning
ret.fp = openBAM(ret.fn);
ret.isEOF =0;
ret.itr=NULL;
ret.idx=NULL;
ret.hdr = sam_hdr_read(ret.fp);
if(ret.hdr==NULL) {
fprintf(stderr, "[main_samview] fail to read the header from \"%s\".\n", ret.fn);
exit(0);
}
return ret;
}
int hts_for_each(string& filename, function<void(Alignment&)> lambda) {
samFile *in = hts_open(filename.c_str(), "r");
if (in == NULL) return 0;
bam_hdr_t *hdr = sam_hdr_read(in);
map<string, string> rg_sample;
parse_rg_sample_map(hdr->text, rg_sample);
bam1_t *b = bam_init1();
while (sam_read1(in, hdr, b) >= 0) {
Alignment a = bam_to_alignment(b, rg_sample);
lambda(a);
}
bam_destroy1(b);
bam_hdr_destroy(hdr);
hts_close(in);
return 1;
}
bam_hdr_t* hts_string_header(string& header,
map<string, int64_t>& path_length,
map<string, string>& rg_sample) {
stringstream hdr;
hdr << "@HD\tVN:1.5\tSO:unknown\n";
for (auto& p : path_length) {
hdr << "@SQ\tSN:" << p.first << "\t" << "LN:" << p.second << "\n";
}
for (auto& s : rg_sample) {
hdr << "@RG\tID:" << s.first << "\t" << "SM:" << s.second << "\n";
}
hdr << "@PG\tID:0\tPN:vg\n";
header = hdr.str();
string sam = "data:" + header;
samFile *in = sam_open(sam.c_str(), "r");
bam_hdr_t *h = sam_hdr_read(in);
sam_close(in);
return h;
}
static int view_sam(hFILE *hfp, const char *filename)
{
samFile *in = hts_hopen(hfp, filename, "r");
if (in == NULL) return 0;
samFile *out = dup_stdout("w");
bam_hdr_t *hdr = sam_hdr_read(in);
if (show_headers) sam_hdr_write(out, hdr);
if (mode == view_all) {
bam1_t *b = bam_init1();
while (sam_read1(in, hdr, b) >= 0)
sam_write1(out, hdr, b);
bam_destroy1(b);
}
bam_hdr_destroy(hdr);
hts_close(out);
hts_close(in);
return 1;
}
bwa_seqio_t *bwa_bam_open(const char *fn, int which)
{
bwa_seqio_t *bs;
#ifndef USE_HTSLIB
bam_header_t *h;
#endif
bs = (bwa_seqio_t*)calloc(1, sizeof(bwa_seqio_t));
bs->is_bam = 1;
bs->which = which;
#ifdef USE_HTSLIB
bs->fp = sam_open(fn, "rb");
#else
bs->fp = bam_open(fn, "r");
#endif
if (0 == bs->fp) err_fatal_simple("Couldn't open bam file");
#ifdef USE_HTSLIB
bs->h = sam_hdr_read(bs->fp);
#else
h = bam_header_read(bs->fp);
bam_header_destroy(h);
#endif
return bs;
}
int hts_for_each_parallel(string& filename, function<void(Alignment&)> lambda) {
samFile *in = hts_open(filename.c_str(), "r");
if (in == NULL) return 0;
bam_hdr_t *hdr = sam_hdr_read(in);
map<string, string> rg_sample;
parse_rg_sample_map(hdr->text, rg_sample);
int thread_count = get_thread_count();
vector<bam1_t*> bs; bs.resize(thread_count);
for (auto& b : bs) {
b = bam_init1();
}
bool more_data = true;
#pragma omp parallel shared(in, hdr, more_data, rg_sample)
{
int tid = omp_get_thread_num();
while (more_data) {
bam1_t* b = bs[tid];
#pragma omp critical (hts_input)
if (more_data) {
more_data = sam_read1(in, hdr, b) >= 0;
}
if (more_data) {
Alignment a = bam_to_alignment(b, rg_sample);
lambda(a);
}
}
}
for (auto& b : bs) bam_destroy1(b);
bam_hdr_destroy(hdr);
hts_close(in);
return 1;
}
BM_mappedRead * extractReads(char * bamFile,
char ** contigs,
int numContigs,
uint16_t * groups,
char * prettyName,
int headersOnly,
int minMapQual,
int maxMisMatches,
int ignoreSuppAlignments,
int ignoreSecondaryAlignments) {
//-----
// code uses the pattern outlined in samtools view (sam_view.c)
// thanks lh3!
//
int i = 0;
int result = -1;
int hh = 0;
int supp_check = 0x0; // include supp mappings
if (ignoreSuppAlignments) {
supp_check |= BAM_FSUPPLEMENTARY;
}
if (ignoreSecondaryAlignments) {
supp_check |= BAM_FSECONDARY;
}
// we need to let the users know if their pairings
// will be corrupted
int p_corrupt = 0;
// helper variables
samFile *in = 0;
bam_hdr_t *header = NULL;
bam1_t *b = bam_init1();
BM_mappedRead * root = 0;
BM_mappedRead * prev = 0;
// open file handlers
if ((in = sam_open(bamFile, "r")) == 0) {
fprintf(stderr,
"ERROR: Failed to open \"%s\" for reading.\n",
bamFile);
}
else {
// retrieve the header
if ((header = sam_hdr_read(in)) == 0) {
fprintf(stderr,
"ERROR: Failed to read the header from \"%s\".\n",
bamFile);
}
else {
// check the index is intact
hts_idx_t *idx = sam_index_load(in, bamFile); // load index
if (idx == 0) { // index is unavailable
fprintf(stderr,
"ERROR: Random retrieval only works "\
"for indexed files.\n");
}
else {
cfuhash_table_t *pair_buffer = \
cfuhash_new_with_initial_size(1000000);
cfuhash_set_flag(pair_buffer, CFUHASH_FROZEN_UNTIL_GROWS);
for (hh = 0; hh < numContigs; ++hh) {
// parse a region in the format like `chr2:100-200'
hts_itr_t *iter = sam_itr_querys(idx, header, contigs[hh]);
if (iter == NULL) { // reference name is not found
fprintf(stderr,
"WARNING: Could not find contig: "\
"[%s] in BAM: [%s].\n",
contigs[hh],
bamFile);
}
// fetch alignments
int line = 0;
while ((result = sam_itr_next(in, iter, b)) >= 0) {
bam1_core_t core = b->core;
line += 1;
// only high quality?, primary? mappings
if ( core.qual < minMapQual)
continue;
if ((core.flag & supp_check) != 0)
continue;
if(bam_aux2i(bam_aux_get(b, "NM")) > maxMisMatches) {
continue;
}
char * seqId = bam_get_qname(b);
char * seq = 0;
char * qual = 0;
int qual_len = 0;
int seq_len = 0;
// get sequence and quality
if(0 == headersOnly) {
// no point allocating unused space
seq = calloc(core.l_qseq+1, sizeof(char));
qual = calloc(core.l_qseq+1, sizeof(char));
uint8_t *s = bam_get_seq(b);
if (core.flag&BAM_FREVERSE) {
// reverse the read
int r = 0;
for (i = core.l_qseq-1; i >=0 ; --i) {
seq[r]="=TGKCYSBAWRDMHVN"[bam_seqi(s,
i)];
++r;
}
}
else {
for (i = 0; i < core.l_qseq; ++i) {
seq[i]="=ACMGRSVTWYHKDBN"[bam_seqi(s,
i)];
}
}
seq_len = core.l_qseq;
s = bam_get_qual(b);
if (s[0] != 0xff) {
qual_len = core.l_qseq;
for (i = 0; i < core.l_qseq; ++i) {
qual[i] = (char)(s[i] + 33);
}
}
else if (qual != 0) {
free(qual);
qual = 0;
}
}
// work out pairing information
uint8_t rpi = RPI_ERROR;
if (core.flag&BAM_FPAIRED) {
if(core.flag&BAM_FMUNMAP) {
if (core.flag&BAM_FREAD1) {
rpi = RPI_SNGL_FIR;
}
else if (core.flag&BAM_FREAD2) {
rpi = RPI_SNGL_SEC;
}
}
else {
if (core.flag&BAM_FREAD1) {
rpi = RPI_FIR;
}
else if (core.flag&BAM_FREAD2) {
rpi = RPI_SEC;
}
}
}
else {
rpi = RPI_SNGL;
}
// make the funky Id
#define MAX_SEQ_ID_LEN 80
char * seq_id = calloc(MAX_SEQ_ID_LEN,
sizeof(char));
// allocate the string to the buffer but check to
// ensure we're not cutting anything off
int id_len = snprintf(seq_id,
MAX_SEQ_ID_LEN,
"b_%s;c_%s;r_%s",
prettyName,
contigs[hh],
seqId);
if(id_len >= MAX_SEQ_ID_LEN) {
seq_id = calloc(id_len+1, sizeof(char));
snprintf(seq_id,
id_len+1, // don't forget the NULL!
"b_%s;c_%s;r_%s",
prettyName,
contigs[hh],
seqId);
}
// make the mapped read struct
prev = makeMappedRead(seq_id,
seq,
qual,
id_len,
seq_len,
qual_len,
rpi,
groups[hh],
prev);
if (0 == root) { root = prev; }
if(rpi == RPI_SNGL || \
rpi == RPI_SNGL_FIR || \
rpi == RPI_SNGL_SEC) {
// we can just add away
// indicate singleton reads by pointing the
// partner pointer to itself
prev->partnerRead = prev;
}
else {
// RPI_FIR or RPI_SEC
// work out pairing information using the hash
// we append a 1 or 2 to the end so that
// we don't accidentally pair 1's with 1's etc.
char * stripped_result;
if(rpi == RPI_FIR) {
stripped_result = \
pairStripper(seqId,
core.l_qname-1,
'2');
}
else {
stripped_result = \
pairStripper(seqId,
core.l_qname-1,
'1');
}
char * stripped = seqId;
if(stripped_result)
stripped = stripped_result;
//fprintf(stdout, "SEARCH %s\n", stripped);
// now stripped always holds a stripped value
// see if it is in the hash already
BM_mappedRead * stored_MR = \
cfuhash_get(pair_buffer,
stripped);
if (0 != stored_MR) {
// exists in the hash -> Add the pair info
if(rpi == RPI_FIR) {
prev->partnerRead = stored_MR;
}
else {
stored_MR->partnerRead = prev;
}
// delete the entry from the hash
cfuhash_delete(pair_buffer,
stripped);
}
else {
// we should put it in the hash
// make sure to change it into something
// we will find next time
if(rpi == RPI_FIR)
stripped[strlen(stripped)-1] = '1';
else
stripped[strlen(stripped)-1] = '2';
// check to make sure we're not overwriting
// anything important. cfuhash overwrites
// duplicate entries, so we need to grab
// it and put it to "SNGL_XXX" before we
// lose the pointer
BM_mappedRead * OWMMR = \
cfuhash_put(pair_buffer,
stripped, prev);
if(OWMMR) {
if(OWMMR->rpi == RPI_FIR)
OWMMR->rpi = RPI_SNGL_FIR;
else
OWMMR->rpi = RPI_SNGL_SEC;
OWMMR->partnerRead = OWMMR;
printPairCorruptionWarning(p_corrupt);
p_corrupt = 1;
}
}
if(stripped_result != 0) { // free this!
free(stripped_result);
stripped_result = 0;
}
}
}
hts_itr_destroy(iter);
if (result < -1) {
fprintf(stderr, "ERROR: retrieval of reads from "\
"contig: \"%s\" failed due to "\
"truncated file or corrupt BAM index "\
"file\n", header->target_name[hh]);
break;
}
}
// any entries left in the hash are pairs whose mates did
// not meet quality standards
size_t key_size = 0;
char * key;
BM_mappedRead * LOMMR;
size_t pr_size = 1;
if(cfuhash_each_data(pair_buffer,
(void**)&key,
&key_size,
(void**)&LOMMR,
&pr_size)) {
do {
// get the mapped read
// update it's pairing so we know it's really single
if (LOMMR->rpi == RPI_FIR)
LOMMR->rpi = RPI_SNGL_FIR;
else if (LOMMR->rpi == RPI_SEC)
LOMMR->rpi = RPI_SNGL_SEC;
// indicate singleton reads by pointing the
// partner pointer to itself
LOMMR->partnerRead = LOMMR;
} while(cfuhash_next_data(pair_buffer,
(void**)&key,
&key_size,
(void**)&LOMMR,
&pr_size));
}
cfuhash_clear(pair_buffer);
cfuhash_destroy(pair_buffer);
}
hts_idx_destroy(idx); // destroy the BAM index
}
}
// always do this
if (in) sam_close(in);
bam_destroy1(b);
if ( header ) bam_hdr_destroy(header);
return root;
}
static bool init(const parsed_opts_t* opts, state_t** state_out) {
state_t* retval = (state_t*) calloc(1, sizeof(state_t));
if (retval == NULL) {
fprintf(stderr, "[init] Out of memory allocating state struct.\n");
return false;
}
*state_out = retval;
// Open files
retval->input_file = sam_open_format(opts->input_name, "r", &opts->ga.in);
if (retval->input_file == NULL) {
fprintf(stderr, "[init] Could not open input file: %s\n", opts->input_name);
return false;
}
retval->input_header = sam_hdr_read(retval->input_file);
retval->output_header = bam_hdr_dup(retval->input_header);
retval->output_file = sam_open_format(opts->output_name == NULL?"-":opts->output_name, "w", &opts->ga.out);
if (retval->output_file == NULL) {
print_error_errno("addreplacerg", "Could not open output file: %s\n", opts->output_name);
return false;
}
if (opts->rg_line) {
// Append new RG line to header.
// Check does not already exist
if ( confirm_rg(retval->output_header, opts->rg_id) ) {
fprintf(stderr, "[init] ID of new RG line specified conflicts with that of an existing header RG line. Overwrite not yet implemented.\n");
return false;
}
retval->rg_id = strdup(opts->rg_id);
size_t new_len = strlen( retval->output_header->text ) + strlen( opts->rg_line ) + 2;
char* new_header = malloc(new_len);
if (!new_header) {
fprintf(stderr, "[init] Out of memory whilst writing new header.\n");
return false;
}
sprintf(new_header,"%s%s\n", retval->output_header->text, opts->rg_line);
free(retval->output_header->text);
retval->output_header->text = new_header;
retval->output_header->l_text = (int)new_len - 1;
} else {
if (opts->rg_id) {
// Confirm what has been supplied exists
if ( !confirm_rg(retval->output_header, opts->rg_id) ) {
fprintf(stderr, "RG ID supplied does not exist in header. Supply full @RG line with -r instead?\n");
return false;
}
retval->rg_id = strdup(opts->rg_id);
} else {
if ((retval->rg_id = get_first_rgid(retval->output_header)) == NULL ) {
fprintf(stderr, "No RG specified on command line or in existing header.\n");
return false;
}
}
}
switch (opts->mode) {
case overwrite_all:
retval->mode_func = &overwrite_all_func;
break;
case orphan_only:
retval->mode_func = &orphan_only_func;
break;
}
return true;
}
int main(int argc, char *argv[])
{
samFile *in;
char *fn_ref = 0;
int flag = 0, c, clevel = -1, ignore_sam_err = 0;
char moder[8];
bam_hdr_t *h;
bam1_t *b;
htsFile *out;
char modew[8];
int r = 0, exit_code = 0;
hts_opt *in_opts = NULL, *out_opts = NULL, *last = NULL;
int nreads = 0;
int benchmark = 0;
while ((c = getopt(argc, argv, "IbDCSl:t:i:o:N:B")) >= 0) {
switch (c) {
case 'S': flag |= 1; break;
case 'b': flag |= 2; break;
case 'D': flag |= 4; break;
case 'C': flag |= 8; break;
case 'B': benchmark = 1; break;
case 'l': clevel = atoi(optarg); flag |= 2; break;
case 't': fn_ref = optarg; break;
case 'I': ignore_sam_err = 1; break;
case 'i': if (add_option(&in_opts, optarg)) return 1; break;
case 'o': if (add_option(&out_opts, optarg)) return 1; break;
case 'N': nreads = atoi(optarg);
}
}
if (argc == optind) {
fprintf(stderr, "Usage: samview [-bSCSIB] [-N num_reads] [-l level] [-o option=value] <in.bam>|<in.sam>|<in.cram> [region]\n");
return 1;
}
strcpy(moder, "r");
if (flag&4) strcat(moder, "c");
else if ((flag&1) == 0) strcat(moder, "b");
in = sam_open(argv[optind], moder);
if (in == NULL) {
fprintf(stderr, "Error opening \"%s\"\n", argv[optind]);
return EXIT_FAILURE;
}
h = sam_hdr_read(in);
h->ignore_sam_err = ignore_sam_err;
b = bam_init1();
strcpy(modew, "w");
if (clevel >= 0 && clevel <= 9) sprintf(modew + 1, "%d", clevel);
if (flag&8) strcat(modew, "c");
else if (flag&2) strcat(modew, "b");
out = hts_open("-", modew);
if (out == NULL) {
fprintf(stderr, "Error opening standard output\n");
return EXIT_FAILURE;
}
/* CRAM output */
if (flag & 8) {
int ret;
// Parse input header and use for CRAM output
out->fp.cram->header = sam_hdr_parse_(h->text, h->l_text);
// Create CRAM references arrays
if (fn_ref)
ret = cram_set_option(out->fp.cram, CRAM_OPT_REFERENCE, fn_ref);
else
// Attempt to fill out a cram->refs[] array from @SQ headers
ret = cram_set_option(out->fp.cram, CRAM_OPT_REFERENCE, NULL);
if (ret != 0)
return EXIT_FAILURE;
}
// Process any options; currently cram only.
for (; in_opts; in_opts = (last=in_opts)->next, free(last)) {
hts_set_opt(in, in_opts->opt, in_opts->val);
if (in_opts->opt == CRAM_OPT_REFERENCE)
if (hts_set_opt(out, in_opts->opt, in_opts->val) != 0)
return EXIT_FAILURE;
}
for (; out_opts; out_opts = (last=out_opts)->next, free(last))
if (hts_set_opt(out, out_opts->opt, out_opts->val) != 0)
return EXIT_FAILURE;
if (!benchmark)
sam_hdr_write(out, h);
if (optind + 1 < argc && !(flag&1)) { // BAM input and has a region
int i;
hts_idx_t *idx;
if ((idx = sam_index_load(in, argv[optind])) == 0) {
fprintf(stderr, "[E::%s] fail to load the BAM index\n", __func__);
return 1;
}
for (i = optind + 1; i < argc; ++i) {
hts_itr_t *iter;
if ((iter = sam_itr_querys(idx, h, argv[i])) == 0) {
fprintf(stderr, "[E::%s] fail to parse region '%s'\n", __func__, argv[i]);
continue;
}
while ((r = sam_itr_next(in, iter, b)) >= 0) {
if (!benchmark && sam_write1(out, h, b) < 0) {
fprintf(stderr, "Error writing output.\n");
exit_code = 1;
break;
}
if (nreads && --nreads == 0)
break;
}
hts_itr_destroy(iter);
}
hts_idx_destroy(idx);
} else while ((r = sam_read1(in, h, b)) >= 0) {
if (!benchmark && sam_write1(out, h, b) < 0) {
fprintf(stderr, "Error writing output.\n");
exit_code = 1;
break;
}
if (nreads && --nreads == 0)
break;
}
if (r < -1) {
fprintf(stderr, "Error parsing input.\n");
exit_code = 1;
}
r = sam_close(out);
if (r < 0) {
fprintf(stderr, "Error closing output.\n");
exit_code = 1;
}
bam_destroy1(b);
bam_hdr_destroy(h);
r = sam_close(in);
if (r < 0) {
fprintf(stderr, "Error closing input.\n");
exit_code = 1;
}
return exit_code;
}
multiReader::multiReader(int argc,char**argv){
gz=Z_NULL;
myglf=NULL;myvcf=NULL;mpil=NULL;bglObj=NULL;
nLines=50;
fname=NULL;
intName=1;
minQ = MINQ;
nInd =0;
isSim =0;
args=NULL;
args = setArgStruct(argc,argv);
fprintf(args->argumentFile,"\t-> Command: \n");
for(int i=0;i<argc;i++)
fprintf(args->argumentFile,"%s ",argv[i]);
// fprintf(args->argumentFile,"\n\n");
if(args->argumentFile!=stderr)
fprintf(args->argumentFile,"\n\t-> angsd version: %s (htslib: %s) build(%s %s)\n",ANGSD_VERSION,hts_version(),__DATE__,__TIME__);
void printTime(FILE *fp);
printTime(args->argumentFile);
//type = args->inputtype;
if(args->argc==2) {
if((!strcasecmp(args->argv[1],"-beagle")) ||
(!strcasecmp(args->argv[1],"-glf")) ||
(!strcasecmp(args->argv[1],"-glf3")) ||
(!strcasecmp(args->argv[1],"-pileup")) ||
(!strcasecmp(args->argv[1],"-vcf-GL")) ||
(!strcasecmp(args->argv[1],"-vcf-pl")) ||
(!strcasecmp(args->argv[1],"-glf10_text")) ||
(!strcasecmp(args->argv[1],"-vcf-GP"))) {
printArg(stdout,args);
exit(0);
}else if ((!strcasecmp(args->argv[1],"-bam")) ||
(!strcasecmp(args->argv[1],"-b"))){
setArgsBam(args);
exit(0);
}else
return;
}
getOptions(args);
if(args->fai==NULL){
int printAndExit =0;
switch(args->inputtype)
{
case INPUT_GLF:
printAndExit=1;
break;
case INPUT_GLF10_TEXT:
printAndExit=1;
break;
case INPUT_GLF3:
printAndExit=1;
break;
case INPUT_BEAGLE:
printAndExit=1;
break;
case INPUT_PILEUP:
printAndExit=1;
break;
}
if(printAndExit){
fprintf(stderr,"\t-> Must supply -fai file\n");
exit(0);
}
}
if(args->fai){
if(!(args->hd=getHeadFromFai(args->fai)))
exit(0);
}else{
if(args->nams.size()==0){
fprintf(stderr,"\t-> Must choose inputfile -bam/-glf/-glf3/-pileup/-i/-vcf-gl/-vcf-gp/-vcf-pl/-glf10_text filename\n");
exit(0);
}
if(args->inputtype==INPUT_BAM){
htsFile *in=sam_open(args->nams[0],"r");
assert(in);
args->hd= sam_hdr_read(in);
hts_close(in);
}
}
if(!(INPUT_VCF_GL||INPUT_VCF_GP)){
if(args->hd==NULL){
fprintf(stderr,"For non-bams you should include -fai arguments\n");
exit(0);
}
}
if((args->inputtype==INPUT_PILEUP||args->inputtype==INPUT_GLF||args->inputtype==INPUT_GLF3||args->inputtype==INPUT_GLF10_TEXT)){
if(nInd==0){
fprintf(stderr,"\t-> Must supply -nInd when using -glf/-glf3/-pileup/-glf10_text files\n");
exit(0);
}
}else
args->nInd = args->nams.size();
if(args->inputtype==INPUT_VCF_GP||args->inputtype==INPUT_VCF_GL){
if(args->regions.size()>1){
fprintf(stderr,"\t-> Only one region can be specified with using bcf (i doubt more is needed) will exit\n");
exit(0);
}else if(args->regions.size()<=1){
myvcf = new vcfReader(args->infile,NULL);
args->hd=bcf_hdr_2_bam_hdr_t(myvcf->hs);
args->nInd = myvcf->hs->nsamples;
}
}
//make args->hd
revMap = buildRevTable(args->hd);
args->revMap = revMap;
setArgsBam(args);
if(args->inputtype==INPUT_VCF_GL){
if(args->regions.size()==1){
char tmp[1024];
int start=args->regions[0].start;
int stop=args->regions[0].stop;
int ref=args->regions[0].refID;
snprintf(tmp,1024,"%s:%d-%d",args->hd->target_name[ref],start+1,stop);
// fprintf(stderr,"tmp:%s\n",tmp);
// exit(0);
myvcf->seek(tmp);
}
}
if(fname==NULL)
return;
gz=Z_NULL;
gz=gzopen(fname,"r");
if(gz==Z_NULL){
fprintf(stderr,"\t-> Problem opening file: \'%s\'\n",fname);
exit(0);
}
switch(args->inputtype){
case INPUT_PILEUP:{
mpil = new mpileup(args->nInd,gz,args->revMap,minQ);
break;
}
case INPUT_GLF:{
myglf = new glfReader(args->nInd,gz,10,isSim);
break;
}
case INPUT_GLF3:{
isSim = 1; //Added by FGV on 22/02/2015: GLF3 is always simulated data until a better alternative can be found
myglf = new glfReader(args->nInd,gz,3,isSim);
break;
}
case INPUT_GLF10_TEXT:{
myglf_text = new glfReader_text(args->nInd,gz,args->revMap);
break;
}
case INPUT_BEAGLE:{
bglObj = new beagle_reader(gz,args->revMap,intName,args->nInd);
break;
}
default:{
break;
}
}
if(args->inputtype==INPUT_VCF_GL||args->inputtype==INPUT_VCF_GL){
fprintf(stderr,"\t-> VCF still beta. Remember that\n");
fprintf(stderr,"\t 1. indels are are discarded\n");
fprintf(stderr,"\t 2. will use chrom, pos PL columns\n");
fprintf(stderr,"\t 3. GL tags are interpreted as log10 and are scaled to ln (NOT USED)\n");
fprintf(stderr,"\t 4. GP tags are interpreted directly as unscaled post probs (spec says phredscaled...) (NOT USED)\n");
fprintf(stderr,"\t 5. FILTER column is currently NOT used (not sure what concensus is)\n");
fprintf(stderr,"\t 6. -sites does NOT work with vcf input but -r does\n");
fprintf(stderr,"\t 7. vcffilereading is still BETA, please report strange behaviour\n");
}
}
int ctx_calls2vcf(int argc, char **argv)
{
const char *in_path = NULL, *out_path = NULL, *out_type = NULL;
// Filtering parameters
int32_t min_mapq = -1, max_align_len = -1, max_allele_len = -1;
// Alignment parameters
int nwmatch = 1, nwmismatch = -2, nwgapopen = -4, nwgapextend = -1;
// ref paths
char const*const* ref_paths = NULL;
size_t nref_paths = 0;
// flank file
const char *sam_path = NULL;
//
// Things we figure out by looking at the input
//
bool isbubble = false;
// samples in VCF, (0 for bubble, does not include ref in breakpoint calls)
size_t i, kmer_size, num_samples;
//
// Reference genome
//
// Hash map of chromosome name -> sequence
ChromHash *genome;
ReadBuffer chroms;
// Arg parsing
char cmd[100];
char shortopts[300];
cmd_long_opts_to_short(longopts, shortopts, sizeof(shortopts));
int c;
// silence error messages from getopt_long
// opterr = 0;
while((c = getopt_long_only(argc, argv, shortopts, longopts, NULL)) != -1) {
cmd_get_longopt_str(longopts, c, cmd, sizeof(cmd));
switch(c) {
case 0: /* flag set */ break;
case 'h': cmd_print_usage(NULL); break;
case 'o': cmd_check(!out_path, cmd); out_path = optarg; break;
case 'O': cmd_check(!out_type, cmd); out_type = optarg; break;
case 'f': cmd_check(!futil_get_force(), cmd); futil_set_force(true); break;
case 'F': cmd_check(!sam_path,cmd); sam_path = optarg; break;
case 'Q': cmd_check(min_mapq < 0,cmd); min_mapq = cmd_uint32(cmd, optarg); break;
case 'A': cmd_check(max_align_len < 0,cmd); max_align_len = cmd_uint32(cmd, optarg); break;
case 'L': cmd_check(max_allele_len < 0,cmd); max_allele_len = cmd_uint32(cmd, optarg); break;
case 'm': nwmatch = cmd_int32(cmd, optarg); break;
case 'M': nwmismatch = cmd_int32(cmd, optarg); break;
case 'g': nwgapopen = cmd_int32(cmd, optarg); break;
case 'G': nwgapextend = cmd_int32(cmd, optarg); break;
case ':': /* BADARG */
case '?': /* BADCH getopt_long has already printed error */
die("`"CMD" "SUBCMD" -h` for help. Bad option: %s", argv[optind-1]);
default: ctx_assert2(0, "shouldn't reach here: %c", c);
}
}
// Defaults for unset values
if(out_path == NULL) out_path = "-";
if(max_align_len < 0) max_align_len = DEFAULT_MAX_ALIGN;
if(max_allele_len < 0) max_allele_len = DEFAULT_MAX_ALLELE;
if(optind+2 > argc)
cmd_print_usage("Require <in.txt.gz> and at least one reference");
in_path = argv[optind++];
ref_paths = (char const*const*)argv + optind;
nref_paths = argc - optind;
// These functions call die() on error
gzFile gzin = futil_gzopen(in_path, "r");
// Read call file header
cJSON *json = json_hdr_load(gzin, in_path);
// Check we can handle the kmer size
kmer_size = json_hdr_get_kmer_size(json, in_path);
db_graph_check_kmer_size(kmer_size, in_path);
// Get format (bubble or breakpoint file)
cJSON *json_fmt = json_hdr_get(json, "file_format", cJSON_String, in_path);
if(strcmp(json_fmt->valuestring,"CtxBreakpoints") == 0) isbubble = false;
else if(strcmp(json_fmt->valuestring,"CtxBubbles") == 0) isbubble = true;
else die("Unknown format: '%s'", json_fmt->valuestring);
status("Reading %s in %s format", futil_inpath_str(in_path),
isbubble ? "bubble" : "breakpoint");
if(isbubble) {
// bubble specific
if(sam_path == NULL)
cmd_print_usage("Require -F <flanks.sam> with bubble file");
if(min_mapq < 0) min_mapq = DEFAULT_MIN_MAPQ;
}
else {
// breakpoint specific
if(min_mapq >= 0)
cmd_print_usage("-Q,--min-mapq <Q> only valid with bubble calls");
}
// Open flank file if it exists
htsFile *samfh = NULL;
bam_hdr_t *bam_hdr = NULL;
bam1_t *mflank = NULL;
if(sam_path)
{
if((samfh = hts_open(sam_path, "r")) == NULL)
die("Cannot open SAM/BAM %s", sam_path);
// Load BAM header
bam_hdr = sam_hdr_read(samfh);
if(bam_hdr == NULL) die("Cannot load BAM header: %s", sam_path);
mflank = bam_init1();
}
// Output VCF has 0 samples if bubbles file, otherwise has N where N is
// number of samples/colours in the breakpoint graph
size_t num_graph_samples = json_hdr_get_ncols(json, in_path);
size_t num_graph_nonref = json_hdr_get_nonref_ncols(json, in_path);
num_samples = 0;
if(!isbubble) {
// If last colour has "is_ref", drop number of samples by one
num_samples = num_graph_nonref < num_graph_samples ? num_graph_samples-1
: num_graph_samples;
}
//
// Open output file
//
if(!out_path) out_path = "-";
int mode = vcf_misc_get_outtype(out_type, out_path);
futil_create_output(out_path);
htsFile *vcffh = hts_open(out_path, modes_htslib[mode]);
status("[calls2vcf] Reading %s call file with %zu samples",
isbubble ? "Bubble" : "Breakpoint", num_graph_samples);
status("[calls2vcf] %zu sample output to: %s format: %s",
num_samples, futil_outpath_str(out_path), hsmodes_htslib[mode]);
if(isbubble) status("[calls2vcf] min. MAPQ: %i", min_mapq);
status("[calls2vcf] max alignment length: %i", max_align_len);
status("[calls2vcf] max VCF allele length: %i", max_allele_len);
status("[calls2vcf] alignment match:%i mismatch:%i gap open:%i extend:%i",
nwmatch, nwmismatch, nwgapopen, nwgapextend);
// Load reference genome
read_buf_alloc(&chroms, 1024);
genome = chrom_hash_init();
chrom_hash_load(ref_paths, nref_paths, &chroms, genome);
// convert to upper case
char *s;
for(i = 0; i < chroms.len; i++)
for(s = chroms.b[i].seq.b; *s; s++) *s = toupper(*s);
if(!isbubble) brkpnt_check_refs_match(json, genome, in_path);
bcf_hdr_t *vcfhdr = make_vcf_hdr(json, in_path, !isbubble, kmer_size,
ref_paths, nref_paths,
chroms.b, chroms.len);
if(bcf_hdr_write(vcffh, vcfhdr) != 0) die("Cannot write VCF header");
AlignedCall *call = acall_init();
CallDecomp *aligner = call_decomp_init(vcffh, vcfhdr);
scoring_t *scoring = call_decomp_get_scoring(aligner);
scoring_init(scoring, nwmatch, nwmismatch, nwgapopen, nwgapextend,
false, false, 0, 0, 0, 0);
CallFileEntry centry;
call_file_entry_alloc(¢ry);
char kmer_str[50];
sprintf(kmer_str, ";K%zu", kmer_size);
if(isbubble)
{
// Bubble calls
DecompBubble *bubbles = decomp_bubble_init();
// Set scoring for aligning 3' flank
scoring = decomp_bubble_get_scoring(bubbles);
scoring_init(scoring, nwmatch, nwmismatch, nwgapopen, nwgapextend,
true, true, 0, 0, 0, 0);
while(call_file_read(gzin, in_path, ¢ry)) {
do {
if(sam_read1(samfh, bam_hdr, mflank) < 0)
die("We've run out of SAM entries!");
} while(mflank->core.flag & (BAM_FSECONDARY | BAM_FSUPPLEMENTARY));
// Align call
strbuf_reset(&call->info);
decomp_bubble_call(bubbles, genome, kmer_size, min_mapq,
¢ry, mflank, bam_hdr, call);
strbuf_append_str(&call->info, kmer_str);
acall_decompose(aligner, call, max_align_len, max_allele_len);
}
// print bubble stats
DecompBubbleStats *bub_stats = ctx_calloc(1, sizeof(*bub_stats));
decomp_bubble_cpy_stats(bub_stats, bubbles);
print_bubble_stats(bub_stats);
ctx_free(bub_stats);
decomp_bubble_destroy(bubbles);
}
else
{
// Breakpoint calls
DecompBreakpoint *breakpoints = decomp_brkpt_init();
while(call_file_read(gzin, in_path, ¢ry)) {
strbuf_reset(&call->info);
decomp_brkpt_call(breakpoints, genome, num_samples, ¢ry, call);
strbuf_append_str(&call->info, kmer_str);
acall_decompose(aligner, call, max_align_len, max_allele_len);
}
// print bubble stats
DecompBreakpointStats *brk_stats = ctx_calloc(1, sizeof(*brk_stats));
decomp_brkpt_cpy_stats(brk_stats, breakpoints);
print_breakpoint_stats(brk_stats);
ctx_free(brk_stats);
decomp_brkpt_destroy(breakpoints);
}
// Print stats
DecomposeStats *astats = ctx_calloc(1, sizeof(*astats));
call_decomp_cpy_stats(astats, aligner);
print_acall_stats(astats);
ctx_free(astats);
call_file_entry_dealloc(¢ry);
call_decomp_destroy(aligner);
acall_destroy(call);
// Finished - clean up
cJSON_Delete(json);
gzclose(gzin);
bcf_hdr_destroy(vcfhdr);
hts_close(vcffh);
for(i = 0; i < chroms.len; i++) seq_read_dealloc(&chroms.b[i]);
read_buf_dealloc(&chroms);
chrom_hash_destroy(genome);
if(sam_path) {
hts_close(samfh);
bam_hdr_destroy(bam_hdr);
bam_destroy1(mflank);
}
return EXIT_SUCCESS;
}
int main_samview(int argc, char *argv[])
{
samFile *in;
char *fn_ref = 0;
int flag = 0, c, clevel = -1, ignore_sam_err = 0;
char moder[8];
bam_hdr_t *h;
bam1_t *b;
while ((c = getopt(argc, argv, "IbSl:t:")) >= 0) {
switch (c) {
case 'S': flag |= 1; break;
case 'b': flag |= 2; break;
case 'l': clevel = atoi(optarg); flag |= 2; break;
case 't': fn_ref = optarg; break;
case 'I': ignore_sam_err = 1; break;
}
}
if (argc == optind) {
fprintf(stderr, "Usage: samview [-bSI] [-l level] <in.bam>|<in.sam> [region]\n");
return 1;
}
strcpy(moder, "r");
if ((flag&1) == 0) strcat(moder, "b");
in = sam_open(argv[optind], moder, fn_ref);
h = sam_hdr_read(in);
h->ignore_sam_err = ignore_sam_err;
b = bam_init1();
if ((flag&4) == 0) { // SAM/BAM output
htsFile *out;
char modew[8];
strcpy(modew, "w");
if (clevel >= 0 && clevel <= 9) sprintf(modew + 1, "%d", clevel);
if (flag&2) strcat(modew, "b");
out = hts_open("-", modew, 0);
sam_hdr_write(out, h);
if (optind + 1 < argc && !(flag&1)) { // BAM input and has a region
int i;
hts_idx_t *idx;
if ((idx = bam_index_load(argv[optind])) == 0) {
fprintf(stderr, "[E::%s] fail to load the BAM index\n", __func__);
return 1;
}
for (i = optind + 1; i < argc; ++i) {
hts_itr_t *iter;
if ((iter = bam_itr_querys(idx, h, argv[i])) == 0) {
fprintf(stderr, "[E::%s] fail to parse region '%s'\n", __func__, argv[i]);
continue;
}
while (bam_itr_next((BGZF*)in->fp, iter, b) >= 0) sam_write1(out, h, b);
hts_itr_destroy(iter);
}
hts_idx_destroy(idx);
} else while (sam_read1(in, h, b) >= 0) sam_write1(out, h, b);
sam_close(out);
}
bam_destroy1(b);
bam_hdr_destroy(h);
sam_close(in);
return 0;
}
int main_bedcov(int argc, char *argv[])
{
gzFile fp;
kstring_t str;
kstream_t *ks;
hts_idx_t **idx;
aux_t **aux;
int *n_plp, dret, i, n, c, min_mapQ = 0;
int64_t *cnt;
const bam_pileup1_t **plp;
int usage = 0;
sam_global_args ga = SAM_GLOBAL_ARGS_INIT;
static const struct option lopts[] = {
SAM_OPT_GLOBAL_OPTIONS('-', 0, '-', '-', 0),
{ NULL, 0, NULL, 0 }
};
while ((c = getopt_long(argc, argv, "Q:", lopts, NULL)) >= 0) {
switch (c) {
case 'Q': min_mapQ = atoi(optarg); break;
default: if (parse_sam_global_opt(c, optarg, lopts, &ga) == 0) break;
/* else fall-through */
case '?': usage = 1; break;
}
if (usage) break;
}
if (usage || optind + 2 > argc) {
fprintf(pysam_stderr, "Usage: samtools bedcov [options] <in.bed> <in1.bam> [...]\n\n");
fprintf(pysam_stderr, " -Q INT Only count bases of at least INT quality [0]\n");
sam_global_opt_help(pysam_stderr, "-.--.");
return 1;
}
memset(&str, 0, sizeof(kstring_t));
n = argc - optind - 1;
aux = calloc(n, sizeof(aux_t*));
idx = calloc(n, sizeof(hts_idx_t*));
for (i = 0; i < n; ++i) {
aux[i] = calloc(1, sizeof(aux_t));
aux[i]->min_mapQ = min_mapQ;
aux[i]->fp = sam_open_format(argv[i+optind+1], "r", &ga.in);
if (aux[i]->fp)
idx[i] = sam_index_load(aux[i]->fp, argv[i+optind+1]);
if (aux[i]->fp == 0 || idx[i] == 0) {
fprintf(pysam_stderr, "ERROR: fail to open index BAM file '%s'\n", argv[i+optind+1]);
return 2;
}
// TODO bgzf_set_cache_size(aux[i]->fp, 20);
aux[i]->header = sam_hdr_read(aux[i]->fp);
if (aux[i]->header == NULL) {
fprintf(pysam_stderr, "ERROR: failed to read header for '%s'\n",
argv[i+optind+1]);
return 2;
}
}
cnt = calloc(n, 8);
fp = gzopen(argv[optind], "rb");
ks = ks_init(fp);
n_plp = calloc(n, sizeof(int));
plp = calloc(n, sizeof(bam_pileup1_t*));
while (ks_getuntil(ks, KS_SEP_LINE, &str, &dret) >= 0) {
char *p, *q;
int tid, beg, end, pos;
bam_mplp_t mplp;
for (p = q = str.s; *p && *p != '\t'; ++p);
if (*p != '\t') goto bed_error;
*p = 0; tid = bam_name2id(aux[0]->header, q); *p = '\t';
if (tid < 0) goto bed_error;
for (q = p = p + 1; isdigit(*p); ++p);
if (*p != '\t') goto bed_error;
*p = 0; beg = atoi(q); *p = '\t';
for (q = p = p + 1; isdigit(*p); ++p);
if (*p == '\t' || *p == 0) {
int c = *p;
*p = 0; end = atoi(q); *p = c;
} else goto bed_error;
for (i = 0; i < n; ++i) {
if (aux[i]->iter) hts_itr_destroy(aux[i]->iter);
aux[i]->iter = sam_itr_queryi(idx[i], tid, beg, end);
}
mplp = bam_mplp_init(n, read_bam, (void**)aux);
bam_mplp_set_maxcnt(mplp, 64000);
memset(cnt, 0, 8 * n);
while (bam_mplp_auto(mplp, &tid, &pos, n_plp, plp) > 0)
if (pos >= beg && pos < end)
for (i = 0; i < n; ++i) cnt[i] += n_plp[i];
for (i = 0; i < n; ++i) {
kputc('\t', &str);
kputl(cnt[i], &str);
}
fputs(str.s, pysam_stdout) & fputc('\n', pysam_stdout);
bam_mplp_destroy(mplp);
continue;
bed_error:
fprintf(pysam_stderr, "Errors in BED line '%s'\n", str.s);
}
free(n_plp); free(plp);
ks_destroy(ks);
gzclose(fp);
free(cnt);
for (i = 0; i < n; ++i) {
if (aux[i]->iter) hts_itr_destroy(aux[i]->iter);
hts_idx_destroy(idx[i]);
bam_hdr_destroy(aux[i]->header);
sam_close(aux[i]->fp);
free(aux[i]);
}
free(aux); free(idx);
free(str.s);
sam_global_args_free(&ga);
return 0;
}
int bam_fillmd(int argc, char *argv[])
{
int c, flt_flag, tid = -2, ret, len, is_bam_out, is_uncompressed, max_nm, is_realn, capQ, baq_flag;
samFile *fp = NULL, *fpout = NULL;
bam_hdr_t *header = NULL;
faidx_t *fai = NULL;
char *ref = NULL, mode_w[8], *ref_file;
bam1_t *b = NULL;
sam_global_args ga = SAM_GLOBAL_ARGS_INIT;
static const struct option lopts[] = {
SAM_OPT_GLOBAL_OPTIONS('-', 0, 0, 0, 0),
{ NULL, 0, NULL, 0 }
};
flt_flag = UPDATE_NM | UPDATE_MD;
is_bam_out = is_uncompressed = is_realn = max_nm = capQ = baq_flag = 0;
strcpy(mode_w, "w");
while ((c = getopt_long(argc, argv, "EqreuNhbSC:n:Ad", lopts, NULL)) >= 0) {
switch (c) {
case 'r':
is_realn = 1;
break;
case 'e':
flt_flag |= USE_EQUAL;
break;
case 'd':
flt_flag |= DROP_TAG;
break;
case 'q':
flt_flag |= BIN_QUAL;
break;
case 'h':
flt_flag |= HASH_QNM;
break;
case 'N':
flt_flag &= ~(UPDATE_MD|UPDATE_NM);
break;
case 'b':
is_bam_out = 1;
break;
case 'u':
is_uncompressed = is_bam_out = 1;
break;
case 'S':
break;
case 'n':
max_nm = atoi(optarg);
break;
case 'C':
capQ = atoi(optarg);
break;
case 'A':
baq_flag |= 1;
break;
case 'E':
baq_flag |= 2;
break;
default:
if (parse_sam_global_opt(c, optarg, lopts, &ga) == 0) break;
fprintf(stderr, "[bam_fillmd] unrecognized option '-%c'\n\n", c);
/* else fall-through */
case '?':
return calmd_usage();
}
}
if (is_bam_out) strcat(mode_w, "b");
else strcat(mode_w, "h");
if (is_uncompressed) strcat(mode_w, "0");
if (optind + (ga.reference == NULL) >= argc)
return calmd_usage();
fp = sam_open_format(argv[optind], "r", &ga.in);
if (fp == NULL) {
print_error_errno("calmd", "Failed to open input file '%s'", argv[optind]);
return 1;
}
header = sam_hdr_read(fp);
if (header == NULL || header->n_targets == 0) {
fprintf(stderr, "[bam_fillmd] input SAM does not have header. Abort!\n");
goto fail;
}
fpout = sam_open_format("-", mode_w, &ga.out);
if (fpout == NULL) {
print_error_errno("calmd", "Failed to open output");
goto fail;
}
if (sam_hdr_write(fpout, header) < 0) {
print_error_errno("calmd", "Failed to write sam header");
goto fail;
}
ref_file = argc > optind + 1 ? argv[optind+1] : ga.reference;
fai = fai_load(ref_file);
if (!fai) {
print_error_errno("calmd", "Failed to open reference file '%s'", ref_file);
goto fail;
}
b = bam_init1();
if (!b) {
fprintf(stderr, "[bam_fillmd] Failed to allocate bam struct\n");
goto fail;
}
while ((ret = sam_read1(fp, header, b)) >= 0) {
if (b->core.tid >= 0) {
if (tid != b->core.tid) {
free(ref);
ref = fai_fetch(fai, header->target_name[b->core.tid], &len);
tid = b->core.tid;
if (ref == 0) { // FIXME: Should this always be fatal?
fprintf(stderr, "[bam_fillmd] fail to find sequence '%s' in the reference.\n",
header->target_name[tid]);
if (is_realn || capQ > 10) goto fail; // Would otherwise crash
}
}
if (is_realn) sam_prob_realn(b, ref, len, baq_flag);
if (capQ > 10) {
int q = sam_cap_mapq(b, ref, len, capQ);
if (b->core.qual > q) b->core.qual = q;
}
if (ref) bam_fillmd1_core(b, ref, len, flt_flag, max_nm);
}
if (sam_write1(fpout, header, b) < 0) {
print_error_errno("calmd", "failed to write to output file");
goto fail;
}
}
if (ret < -1) {
fprintf(stderr, "[bam_fillmd] Error reading input.\n");
goto fail;
}
bam_destroy1(b);
bam_hdr_destroy(header);
free(ref);
fai_destroy(fai);
sam_close(fp);
if (sam_close(fpout) < 0) {
fprintf(stderr, "[bam_fillmd] error when closing output file\n");
return 1;
}
return 0;
fail:
free(ref);
if (b) bam_destroy1(b);
if (header) bam_hdr_destroy(header);
if (fai) fai_destroy(fai);
if (fp) sam_close(fp);
if (fpout) sam_close(fpout);
return 1;
}
int main_depth(int argc, char *argv[])
{
int i, n, tid, reg_tid, beg, end, pos, *n_plp, baseQ = 0, mapQ = 0, min_len = 0;
int all = 0, status = EXIT_SUCCESS, nfiles, max_depth = -1;
const bam_pileup1_t **plp;
char *reg = 0; // specified region
void *bed = 0; // BED data structure
char *file_list = NULL, **fn = NULL;
bam_hdr_t *h = NULL; // BAM header of the 1st input
aux_t **data;
bam_mplp_t mplp;
int last_pos = -1, last_tid = -1, ret;
sam_global_args ga = SAM_GLOBAL_ARGS_INIT;
static const struct option lopts[] = {
SAM_OPT_GLOBAL_OPTIONS('-', 0, '-', '-', 0),
{ NULL, 0, NULL, 0 }
};
// parse the command line
while ((n = getopt_long(argc, argv, "r:b:q:Q:l:f:am:d:", lopts, NULL)) >= 0) {
switch (n) {
case 'l':
min_len = atoi(optarg);
break; // minimum query length
case 'r':
reg = strdup(optarg);
break; // parsing a region requires a BAM header
case 'b':
bed = bed_read(optarg); // BED or position list file can be parsed now
if (!bed) {
print_error_errno("depth", "Could not read file \"%s\"", optarg);
return 1;
}
break;
case 'q':
baseQ = atoi(optarg);
break; // base quality threshold
case 'Q':
mapQ = atoi(optarg);
break; // mapping quality threshold
case 'f':
file_list = optarg;
break;
case 'a':
all++;
break;
case 'd':
case 'm':
max_depth = atoi(optarg);
break; // maximum coverage depth
default:
if (parse_sam_global_opt(n, optarg, lopts, &ga) == 0) break;
/* else fall-through */
case '?':
return usage();
}
}
if (optind == argc && !file_list)
return usage();
// initialize the auxiliary data structures
if (file_list)
{
if ( read_file_list(file_list,&nfiles,&fn) ) return 1;
n = nfiles;
argv = fn;
optind = 0;
}
else
n = argc - optind; // the number of BAMs on the command line
data = calloc(n, sizeof(aux_t*)); // data[i] for the i-th input
reg_tid = 0;
beg = 0;
end = INT_MAX; // set the default region
for (i = 0; i < n; ++i) {
int rf;
data[i] = calloc(1, sizeof(aux_t));
data[i]->fp = sam_open_format(argv[optind+i], "r", &ga.in); // open BAM
if (data[i]->fp == NULL) {
print_error_errno("depth", "Could not open \"%s\"", argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
rf = SAM_FLAG | SAM_RNAME | SAM_POS | SAM_MAPQ | SAM_CIGAR | SAM_SEQ;
if (baseQ) rf |= SAM_QUAL;
if (hts_set_opt(data[i]->fp, CRAM_OPT_REQUIRED_FIELDS, rf)) {
fprintf(stderr, "Failed to set CRAM_OPT_REQUIRED_FIELDS value\n");
return 1;
}
if (hts_set_opt(data[i]->fp, CRAM_OPT_DECODE_MD, 0)) {
fprintf(stderr, "Failed to set CRAM_OPT_DECODE_MD value\n");
return 1;
}
data[i]->min_mapQ = mapQ; // set the mapQ filter
data[i]->min_len = min_len; // set the qlen filter
data[i]->hdr = sam_hdr_read(data[i]->fp); // read the BAM header
if (data[i]->hdr == NULL) {
fprintf(stderr, "Couldn't read header for \"%s\"\n",
argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
if (reg) { // if a region is specified
hts_idx_t *idx = sam_index_load(data[i]->fp, argv[optind+i]); // load the index
if (idx == NULL) {
print_error("depth", "can't load index for \"%s\"", argv[optind+i]);
status = EXIT_FAILURE;
goto depth_end;
}
data[i]->iter = sam_itr_querys(idx, data[i]->hdr, reg); // set the iterator
hts_idx_destroy(idx); // the index is not needed any more; free the memory
if (data[i]->iter == NULL) {
print_error("depth", "can't parse region \"%s\"", reg);
status = EXIT_FAILURE;
goto depth_end;
}
}
}
h = data[0]->hdr; // easy access to the header of the 1st BAM
if (reg) {
beg = data[0]->iter->beg; // and to the parsed region coordinates
end = data[0]->iter->end;
reg_tid = data[0]->iter->tid;
}
// the core multi-pileup loop
mplp = bam_mplp_init(n, read_bam, (void**)data); // initialization
if (0 < max_depth)
bam_mplp_set_maxcnt(mplp,max_depth); // set maximum coverage depth
n_plp = calloc(n, sizeof(int)); // n_plp[i] is the number of covering reads from the i-th BAM
plp = calloc(n, sizeof(bam_pileup1_t*)); // plp[i] points to the array of covering reads (internal in mplp)
while ((ret=bam_mplp_auto(mplp, &tid, &pos, n_plp, plp)) > 0) { // come to the next covered position
if (pos < beg || pos >= end) continue; // out of range; skip
if (tid >= h->n_targets) continue; // diff number of @SQ lines per file?
if (all) {
while (tid > last_tid) {
if (last_tid >= 0 && !reg) {
// Deal with remainder or entirety of last tid.
while (++last_pos < h->target_len[last_tid]) {
// Horribly inefficient, but the bed API is an obfuscated black box.
if (bed && bed_overlap(bed, h->target_name[last_tid], last_pos, last_pos + 1) == 0)
continue;
fputs(h->target_name[last_tid], stdout);
printf("\t%d", last_pos+1);
for (i = 0; i < n; i++)
putchar('\t'), putchar('0');
putchar('\n');
}
}
last_tid++;
last_pos = -1;
if (all < 2)
break;
}
// Deal with missing portion of current tid
while (++last_pos < pos) {
if (last_pos < beg) continue; // out of range; skip
if (bed && bed_overlap(bed, h->target_name[tid], last_pos, last_pos + 1) == 0)
continue;
fputs(h->target_name[tid], stdout);
printf("\t%d", last_pos+1);
for (i = 0; i < n; i++)
putchar('\t'), putchar('0');
putchar('\n');
}
last_tid = tid;
last_pos = pos;
}
if (bed && bed_overlap(bed, h->target_name[tid], pos, pos + 1) == 0) continue;
fputs(h->target_name[tid], stdout);
printf("\t%d", pos+1); // a customized printf() would be faster
for (i = 0; i < n; ++i) { // base level filters have to go here
int j, m = 0;
for (j = 0; j < n_plp[i]; ++j) {
const bam_pileup1_t *p = plp[i] + j; // DON'T modfity plp[][] unless you really know
if (p->is_del || p->is_refskip) ++m; // having dels or refskips at tid:pos
else if (bam_get_qual(p->b)[p->qpos] < baseQ) ++m; // low base quality
}
printf("\t%d", n_plp[i] - m); // this the depth to output
}
putchar('\n');
}
if (ret < 0) status = EXIT_FAILURE;
free(n_plp);
free(plp);
bam_mplp_destroy(mplp);
if (all) {
// Handle terminating region
if (last_tid < 0 && reg && all > 1) {
last_tid = reg_tid;
last_pos = beg-1;
}
while (last_tid >= 0 && last_tid < h->n_targets) {
while (++last_pos < h->target_len[last_tid]) {
if (last_pos >= end) break;
if (bed && bed_overlap(bed, h->target_name[last_tid], last_pos, last_pos + 1) == 0)
continue;
fputs(h->target_name[last_tid], stdout);
printf("\t%d", last_pos+1);
for (i = 0; i < n; i++)
putchar('\t'), putchar('0');
putchar('\n');
}
last_tid++;
last_pos = -1;
if (all < 2 || reg)
break;
}
}
depth_end:
for (i = 0; i < n && data[i]; ++i) {
bam_hdr_destroy(data[i]->hdr);
if (data[i]->fp) sam_close(data[i]->fp);
hts_itr_destroy(data[i]->iter);
free(data[i]);
}
free(data);
free(reg);
if (bed) bed_destroy(bed);
if ( file_list )
{
for (i=0; i<n; i++) free(fn[i]);
free(fn);
}
sam_global_args_free(&ga);
return status;
}
static int mpileup(mplp_conf_t *conf)
{
if (conf->nfiles == 0) {
fprintf(stderr,"[%s] no input file/data given\n", __func__);
exit(EXIT_FAILURE);
}
mplp_ref_t mp_ref = MPLP_REF_INIT;
conf->gplp = (mplp_pileup_t *) calloc(1,sizeof(mplp_pileup_t));
conf->mplp_data = (mplp_aux_t**) calloc(conf->nfiles, sizeof(mplp_aux_t*));
conf->plp = (const bam_pileup1_t**) calloc(conf->nfiles, sizeof(bam_pileup1_t*));
conf->n_plp = (int*) calloc(conf->nfiles, sizeof(int));
// Allow to run mpileup on multiple regions in one go. This comes at cost: the bai index
// must be kept in the memory for the whole time which can be a problem with many bams.
// Therefore if none or only one region is requested, we initialize the bam iterator as
// before and free the index. Only when multiple regions are queried, we keep the index.
int nregs = 0;
if ( conf->reg_fname )
{
if ( conf->reg_is_file )
{
conf->reg = regidx_init(conf->reg_fname,NULL,NULL,0,NULL);
if ( !conf->reg ) {
fprintf(stderr,"Could not parse the regions: %s\n", conf->reg_fname);
exit(EXIT_FAILURE);
}
}
else
{
conf->reg = regidx_init(NULL,regidx_parse_reg,NULL,sizeof(char*),NULL);
if ( regidx_insert_list(conf->reg,conf->reg_fname,',') !=0 ) {
fprintf(stderr,"Could not parse the regions: %s\n", conf->reg_fname);
exit(EXIT_FAILURE);
}
}
nregs = regidx_nregs(conf->reg);
conf->reg_itr = regitr_init(conf->reg);
regitr_loop(conf->reg_itr); // region iterator now positioned at the first region
}
// read the header of each file in the list and initialize data
// beware: mpileup has always assumed that tid's are consistent in the headers, add sanity check at least!
bam_hdr_t *hdr = NULL; // header of first file in input list
int i;
for (i = 0; i < conf->nfiles; ++i) {
bam_hdr_t *h_tmp;
conf->mplp_data[i] = (mplp_aux_t*) calloc(1, sizeof(mplp_aux_t));
conf->mplp_data[i]->fp = sam_open(conf->files[i], "rb");
if ( !conf->mplp_data[i]->fp )
{
fprintf(stderr, "[%s] failed to open %s: %s\n", __func__, conf->files[i], strerror(errno));
exit(EXIT_FAILURE);
}
if (hts_set_opt(conf->mplp_data[i]->fp, CRAM_OPT_DECODE_MD, 0)) {
fprintf(stderr, "Failed to set CRAM_OPT_DECODE_MD value\n");
exit(EXIT_FAILURE);
}
if (conf->fai_fname && hts_set_fai_filename(conf->mplp_data[i]->fp, conf->fai_fname) != 0) {
fprintf(stderr, "[%s] failed to process %s: %s\n",
__func__, conf->fai_fname, strerror(errno));
exit(EXIT_FAILURE);
}
conf->mplp_data[i]->conf = conf;
conf->mplp_data[i]->ref = &mp_ref;
h_tmp = sam_hdr_read(conf->mplp_data[i]->fp);
if ( !h_tmp ) {
fprintf(stderr,"[%s] fail to read the header of %s\n", __func__, conf->files[i]);
exit(EXIT_FAILURE);
}
conf->mplp_data[i]->h = i ? hdr : h_tmp; // for j==0, "h" has not been set yet
conf->mplp_data[i]->bam_id = bam_smpl_add_bam(conf->bsmpl,h_tmp->text,conf->files[i]);
if ( conf->mplp_data[i]->bam_id<0 )
{
// no usable readgroups in this bam, it can be skipped
sam_close(conf->mplp_data[i]->fp);
free(conf->mplp_data[i]);
bam_hdr_destroy(h_tmp);
free(conf->files[i]);
if ( i+1<conf->nfiles ) memmove(&conf->files[i],&conf->files[i+1],sizeof(*conf->files)*(conf->nfiles-i-1));
conf->nfiles--;
i--;
continue;
}
if (conf->reg) {
hts_idx_t *idx = sam_index_load(conf->mplp_data[i]->fp, conf->files[i]);
if (idx == NULL) {
fprintf(stderr, "[%s] fail to load index for %s\n", __func__, conf->files[i]);
exit(EXIT_FAILURE);
}
conf->buf.l = 0;
ksprintf(&conf->buf,"%s:%u-%u",conf->reg_itr->seq,conf->reg_itr->beg+1,conf->reg_itr->end+1);
conf->mplp_data[i]->iter = sam_itr_querys(idx, conf->mplp_data[i]->h, conf->buf.s);
if ( !conf->mplp_data[i]->iter )
{
conf->mplp_data[i]->iter = sam_itr_querys(idx, conf->mplp_data[i]->h, conf->reg_itr->seq);
if ( conf->mplp_data[i]->iter ) {
fprintf(stderr,"[E::%s] fail to parse region '%s'\n", __func__, conf->buf.s);
exit(EXIT_FAILURE);
}
fprintf(stderr,"[E::%s] the sequence \"%s\" not found: %s\n",__func__,conf->reg_itr->seq,conf->files[i]);
exit(EXIT_FAILURE);
}
if ( nregs==1 ) // no need to keep the index in memory
hts_idx_destroy(idx);
else
conf->mplp_data[i]->idx = idx;
}
if ( !hdr ) hdr = h_tmp; /* save the header of first file in list */
else {
// FIXME: check consistency between h and h_tmp
bam_hdr_destroy(h_tmp);
// we store only the first file's header; it's (alleged to be)
// compatible with the i-th file's target_name lookup needs
conf->mplp_data[i]->h = hdr;
}
}
// allocate data storage proportionate to number of samples being studied sm->n
bam_smpl_get_samples(conf->bsmpl, &conf->gplp->n);
conf->gplp->n_plp = (int*) calloc(conf->gplp->n, sizeof(int));
conf->gplp->m_plp = (int*) calloc(conf->gplp->n, sizeof(int));
conf->gplp->plp = (bam_pileup1_t**) calloc(conf->gplp->n, sizeof(bam_pileup1_t*));
fprintf(stderr, "[%s] %d samples in %d input files\n", __func__, conf->gplp->n, conf->nfiles);
// write the VCF header
conf->bcf_fp = hts_open(conf->output_fname?conf->output_fname:"-", hts_bcf_wmode(conf->output_type));
if (conf->bcf_fp == NULL) {
fprintf(stderr, "[%s] failed to write to %s: %s\n", __func__, conf->output_fname? conf->output_fname : "standard output", strerror(errno));
exit(EXIT_FAILURE);
}
if ( conf->n_threads ) hts_set_threads(conf->bcf_fp, conf->n_threads);
// BCF header creation
conf->bcf_hdr = bcf_hdr_init("w");
conf->buf.l = 0;
if (conf->record_cmd_line)
{
ksprintf(&conf->buf, "##bcftoolsVersion=%s+htslib-%s\n",bcftools_version(),hts_version());
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
conf->buf.l = 0;
ksprintf(&conf->buf, "##bcftoolsCommand=mpileup");
for (i=1; i<conf->argc; i++) ksprintf(&conf->buf, " %s", conf->argv[i]);
kputc('\n', &conf->buf);
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
}
if (conf->fai_fname)
{
conf->buf.l = 0;
ksprintf(&conf->buf, "##reference=file://%s\n", conf->fai_fname);
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
}
// Translate BAM @SQ tags to BCF ##contig tags
// todo: use/write new BAM header manipulation routines, fill also UR, M5
for (i=0; i<hdr->n_targets; i++)
{
conf->buf.l = 0;
ksprintf(&conf->buf, "##contig=<ID=%s,length=%d>", hdr->target_name[i], hdr->target_len[i]);
bcf_hdr_append(conf->bcf_hdr, conf->buf.s);
}
conf->buf.l = 0;
bcf_hdr_append(conf->bcf_hdr,"##ALT=<ID=*,Description=\"Represents allele(s) other than observed.\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=INDEL,Number=0,Type=Flag,Description=\"Indicates that the variant is an INDEL.\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=IDV,Number=1,Type=Integer,Description=\"Maximum number of reads supporting an indel\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=IMF,Number=1,Type=Float,Description=\"Maximum fraction of reads supporting an indel\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=DP,Number=1,Type=Integer,Description=\"Raw read depth\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=VDB,Number=1,Type=Float,Description=\"Variant Distance Bias for filtering splice-site artefacts in RNA-seq data (bigger is better)\",Version=\"3\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=RPB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Read Position Bias (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality Bias (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=BQB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Base Quality Bias (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQSB,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality vs Strand Bias (bigger is better)\">");
#if CDF_MWU_TESTS
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=RPB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Read Position Bias [CDF] (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality Bias [CDF] (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=BQB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Base Quality Bias [CDF] (bigger is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQSB2,Number=1,Type=Float,Description=\"Mann-Whitney U test of Mapping Quality vs Strand Bias [CDF] (bigger is better)\">");
#endif
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=SGB,Number=1,Type=Float,Description=\"Segregation based metric.\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=MQ0F,Number=1,Type=Float,Description=\"Fraction of MQ0 reads (smaller is better)\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=I16,Number=16,Type=Float,Description=\"Auxiliary tag used for calling, see description of bcf_callret1_t in bam2bcf.h\">");
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=QS,Number=R,Type=Float,Description=\"Auxiliary tag used for calling\">");
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=PL,Number=G,Type=Integer,Description=\"List of Phred-scaled genotype likelihoods\">");
if ( conf->fmt_flag&B2B_FMT_DP )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DP,Number=1,Type=Integer,Description=\"Number of high-quality bases\">");
if ( conf->fmt_flag&B2B_FMT_DV )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DV,Number=1,Type=Integer,Description=\"Number of high-quality non-reference bases\">");
if ( conf->fmt_flag&B2B_FMT_DPR )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DPR,Number=R,Type=Integer,Description=\"Number of high-quality bases observed for each allele\">");
if ( conf->fmt_flag&B2B_INFO_DPR )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=DPR,Number=R,Type=Integer,Description=\"Number of high-quality bases observed for each allele\">");
if ( conf->fmt_flag&B2B_FMT_DP4 )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=DP4,Number=4,Type=Integer,Description=\"Number of high-quality ref-fwd, ref-reverse, alt-fwd and alt-reverse bases\">");
if ( conf->fmt_flag&B2B_FMT_SP )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=SP,Number=1,Type=Integer,Description=\"Phred-scaled strand bias P-value\">");
if ( conf->fmt_flag&B2B_FMT_AD )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=AD,Number=R,Type=Integer,Description=\"Allelic depths\">");
if ( conf->fmt_flag&B2B_FMT_ADF )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=ADF,Number=R,Type=Integer,Description=\"Allelic depths on the forward strand\">");
if ( conf->fmt_flag&B2B_FMT_ADR )
bcf_hdr_append(conf->bcf_hdr,"##FORMAT=<ID=ADR,Number=R,Type=Integer,Description=\"Allelic depths on the reverse strand\">");
if ( conf->fmt_flag&B2B_INFO_AD )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=AD,Number=R,Type=Integer,Description=\"Total allelic depths\">");
if ( conf->fmt_flag&B2B_INFO_ADF )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=ADF,Number=R,Type=Integer,Description=\"Total allelic depths on the forward strand\">");
if ( conf->fmt_flag&B2B_INFO_ADR )
bcf_hdr_append(conf->bcf_hdr,"##INFO=<ID=ADR,Number=R,Type=Integer,Description=\"Total allelic depths on the reverse strand\">");
if ( conf->gvcf )
gvcf_update_header(conf->gvcf, conf->bcf_hdr);
int nsmpl;
const char **smpl = bam_smpl_get_samples(conf->bsmpl, &nsmpl);
for (i=0; i<nsmpl; i++)
bcf_hdr_add_sample(conf->bcf_hdr, smpl[i]);
bcf_hdr_write(conf->bcf_fp, conf->bcf_hdr);
conf->bca = bcf_call_init(-1., conf->min_baseQ);
conf->bcr = (bcf_callret1_t*) calloc(nsmpl, sizeof(bcf_callret1_t));
conf->bca->openQ = conf->openQ, conf->bca->extQ = conf->extQ, conf->bca->tandemQ = conf->tandemQ;
conf->bca->min_frac = conf->min_frac;
conf->bca->min_support = conf->min_support;
conf->bca->per_sample_flt = conf->flag & MPLP_PER_SAMPLE;
conf->bc.bcf_hdr = conf->bcf_hdr;
conf->bc.n = nsmpl;
conf->bc.PL = (int32_t*) malloc(15 * nsmpl * sizeof(*conf->bc.PL));
if (conf->fmt_flag)
{
assert( sizeof(float)==sizeof(int32_t) );
conf->bc.DP4 = (int32_t*) malloc(nsmpl * sizeof(int32_t) * 4);
conf->bc.fmt_arr = (uint8_t*) malloc(nsmpl * sizeof(float)); // all fmt_flag fields, float and int32
if ( conf->fmt_flag&(B2B_INFO_DPR|B2B_FMT_DPR|B2B_INFO_AD|B2B_INFO_ADF|B2B_INFO_ADR|B2B_FMT_AD|B2B_FMT_ADF|B2B_FMT_ADR) )
{
// first B2B_MAX_ALLELES fields for total numbers, the rest per-sample
conf->bc.ADR = (int32_t*) malloc((nsmpl+1)*B2B_MAX_ALLELES*sizeof(int32_t));
conf->bc.ADF = (int32_t*) malloc((nsmpl+1)*B2B_MAX_ALLELES*sizeof(int32_t));
for (i=0; i<nsmpl; i++)
{
conf->bcr[i].ADR = conf->bc.ADR + (i+1)*B2B_MAX_ALLELES;
conf->bcr[i].ADF = conf->bc.ADF + (i+1)*B2B_MAX_ALLELES;
}
}
}
// init mpileup
conf->iter = bam_mplp_init(conf->nfiles, mplp_func, (void**)conf->mplp_data);
if ( conf->flag & MPLP_SMART_OVERLAPS ) bam_mplp_init_overlaps(conf->iter);
if ( (double)conf->max_depth * conf->nfiles > 1<<20)
fprintf(stderr, "Warning: Potential memory hog, up to %.0fM reads in the pileup!\n", (double)conf->max_depth*conf->nfiles);
if ( (double)conf->max_depth * conf->nfiles / nsmpl < 250 )
fprintf(stderr, "Note: The maximum per-sample depth with -d %d is %.1fx\n", conf->max_depth,(double)conf->max_depth * conf->nfiles / nsmpl);
bam_mplp_set_maxcnt(conf->iter, conf->max_depth);
conf->max_indel_depth = conf->max_indel_depth * nsmpl;
conf->bcf_rec = bcf_init1();
bam_mplp_constructor(conf->iter, pileup_constructor);
// Run mpileup for multiple regions
if ( nregs )
{
int ireg = 0;
do
{
// first region is already positioned
if ( ireg++ > 0 )
{
conf->buf.l = 0;
ksprintf(&conf->buf,"%s:%u-%u",conf->reg_itr->seq,conf->reg_itr->beg,conf->reg_itr->end);
for (i=0; i<conf->nfiles; i++)
{
hts_itr_destroy(conf->mplp_data[i]->iter);
conf->mplp_data[i]->iter = sam_itr_querys(conf->mplp_data[i]->idx, conf->mplp_data[i]->h, conf->buf.s);
if ( !conf->mplp_data[i]->iter )
{
conf->mplp_data[i]->iter = sam_itr_querys(conf->mplp_data[i]->idx, conf->mplp_data[i]->h, conf->reg_itr->seq);
if ( conf->mplp_data[i]->iter ) {
fprintf(stderr,"[E::%s] fail to parse region '%s'\n", __func__, conf->buf.s);
exit(EXIT_FAILURE);
}
fprintf(stderr,"[E::%s] the sequence \"%s\" not found: %s\n",__func__,conf->reg_itr->seq,conf->files[i]);
exit(EXIT_FAILURE);
}
bam_mplp_reset(conf->iter);
}
}
mpileup_reg(conf,conf->reg_itr->beg,conf->reg_itr->end);
}
while ( regitr_loop(conf->reg_itr) );
}
else
mpileup_reg(conf,0,0);
flush_bcf_records(conf, conf->bcf_fp, conf->bcf_hdr, NULL);
// clean up
free(conf->bc.tmp.s);
bcf_destroy1(conf->bcf_rec);
if (conf->bcf_fp)
{
hts_close(conf->bcf_fp);
bcf_hdr_destroy(conf->bcf_hdr);
bcf_call_destroy(conf->bca);
free(conf->bc.PL);
free(conf->bc.DP4);
free(conf->bc.ADR);
free(conf->bc.ADF);
free(conf->bc.fmt_arr);
free(conf->bcr);
}
if ( conf->gvcf ) gvcf_destroy(conf->gvcf);
free(conf->buf.s);
for (i = 0; i < conf->gplp->n; ++i) free(conf->gplp->plp[i]);
free(conf->gplp->plp); free(conf->gplp->n_plp); free(conf->gplp->m_plp); free(conf->gplp);
bam_mplp_destroy(conf->iter);
bam_hdr_destroy(hdr);
for (i = 0; i < conf->nfiles; ++i) {
if ( nregs>1 ) hts_idx_destroy(conf->mplp_data[i]->idx);
sam_close(conf->mplp_data[i]->fp);
if ( conf->mplp_data[i]->iter) hts_itr_destroy(conf->mplp_data[i]->iter);
free(conf->mplp_data[i]);
}
if ( conf->reg_itr ) regitr_destroy(conf->reg_itr);
free(conf->mplp_data); free(conf->plp); free(conf->n_plp);
free(mp_ref.ref[0]);
free(mp_ref.ref[1]);
return 0;
}
// currently, this function ONLY works if each read has one hit
static void bam_mating_core(samFile* in, samFile* out, int remove_reads, int proper_pair_check, int add_ct)
{
bam_hdr_t *header;
bam1_t *b[2];
int curr, has_prev, pre_end = 0, cur_end = 0;
kstring_t str;
str.l = str.m = 0; str.s = 0;
header = sam_hdr_read(in);
if (header == NULL) {
fprintf(stderr, "[bam_mating_core] ERROR: Couldn't read header\n");
exit(1);
}
// Accept unknown, unsorted, or queryname sort order, but error on coordinate sorted.
if ((header->l_text > 3) && (strncmp(header->text, "@HD", 3) == 0)) {
char *p, *q;
p = strstr(header->text, "\tSO:coordinate");
q = strchr(header->text, '\n');
// Looking for SO:coordinate within the @HD line only
// (e.g. must ignore in a @CO comment line later in header)
if ((p != 0) && (p < q)) {
fprintf(stderr, "[bam_mating_core] ERROR: Coordinate sorted, require grouped/sorted by queryname.\n");
exit(1);
}
}
sam_hdr_write(out, header);
b[0] = bam_init1();
b[1] = bam_init1();
curr = 0; has_prev = 0;
while (sam_read1(in, header, b[curr]) >= 0) {
bam1_t *cur = b[curr], *pre = b[1-curr];
if (cur->core.flag & BAM_FSECONDARY)
{
if ( !remove_reads ) sam_write1(out, header, cur);
continue; // skip secondary alignments
}
if (cur->core.flag & BAM_FSUPPLEMENTARY)
{
sam_write1(out, header, cur);
continue; // pass supplementary alignments through unchanged (TODO:make them match read they came from)
}
if (cur->core.tid < 0 || cur->core.pos < 0) // If unmapped set the flag
{
cur->core.flag |= BAM_FUNMAP;
}
if ((cur->core.flag&BAM_FUNMAP) == 0) // If mapped calculate end
{
cur_end = bam_endpos(cur);
// Check cur_end isn't past the end of the contig we're on, if it is set the UNMAP'd flag
if (cur_end > (int)header->target_len[cur->core.tid]) cur->core.flag |= BAM_FUNMAP;
}
if (has_prev) { // do we have a pair of reads to examine?
if (strcmp(bam_get_qname(cur), bam_get_qname(pre)) == 0) { // identical pair name
pre->core.flag |= BAM_FPAIRED;
cur->core.flag |= BAM_FPAIRED;
sync_mate(pre, cur);
if (pre->core.tid == cur->core.tid && !(cur->core.flag&(BAM_FUNMAP|BAM_FMUNMAP))
&& !(pre->core.flag&(BAM_FUNMAP|BAM_FMUNMAP))) // if safe set TLEN/ISIZE
{
uint32_t cur5, pre5;
cur5 = (cur->core.flag&BAM_FREVERSE)? cur_end : cur->core.pos;
pre5 = (pre->core.flag&BAM_FREVERSE)? pre_end : pre->core.pos;
cur->core.isize = pre5 - cur5; pre->core.isize = cur5 - pre5;
} else cur->core.isize = pre->core.isize = 0;
if (add_ct) bam_template_cigar(pre, cur, &str);
// TODO: Add code to properly check if read is in a proper pair based on ISIZE distribution
if (proper_pair_check && !plausibly_properly_paired(pre,cur)) {
pre->core.flag &= ~BAM_FPROPER_PAIR;
cur->core.flag &= ~BAM_FPROPER_PAIR;
}
// Write out result
if ( !remove_reads ) {
sam_write1(out, header, pre);
sam_write1(out, header, cur);
} else {
// If we have to remove reads make sure we do it in a way that doesn't create orphans with bad flags
if(pre->core.flag&BAM_FUNMAP) cur->core.flag &= ~(BAM_FPAIRED|BAM_FMREVERSE|BAM_FPROPER_PAIR);
if(cur->core.flag&BAM_FUNMAP) pre->core.flag &= ~(BAM_FPAIRED|BAM_FMREVERSE|BAM_FPROPER_PAIR);
if(!(pre->core.flag&BAM_FUNMAP)) sam_write1(out, header, pre);
if(!(cur->core.flag&BAM_FUNMAP)) sam_write1(out, header, cur);
}
has_prev = 0;
} else { // unpaired? clear bad info and write it out
if (pre->core.tid < 0 || pre->core.pos < 0 || pre->core.flag&BAM_FUNMAP) { // If unmapped
pre->core.flag |= BAM_FUNMAP;
pre->core.tid = -1;
pre->core.pos = -1;
}
pre->core.mtid = -1; pre->core.mpos = -1; pre->core.isize = 0;
pre->core.flag &= ~(BAM_FPAIRED|BAM_FMREVERSE|BAM_FPROPER_PAIR);
if ( !remove_reads || !(pre->core.flag&BAM_FUNMAP) ) sam_write1(out, header, pre);
}
} else has_prev = 1;
curr = 1 - curr;
pre_end = cur_end;
}
if (has_prev && !remove_reads) { // If we still have a BAM in the buffer it must be unpaired
bam1_t *pre = b[1-curr];
if (pre->core.tid < 0 || pre->core.pos < 0 || pre->core.flag&BAM_FUNMAP) { // If unmapped
pre->core.flag |= BAM_FUNMAP;
pre->core.tid = -1;
pre->core.pos = -1;
}
pre->core.mtid = -1; pre->core.mpos = -1; pre->core.isize = 0;
pre->core.flag &= ~(BAM_FPAIRED|BAM_FMREVERSE|BAM_FPROPER_PAIR);
sam_write1(out, header, pre);
}
bam_hdr_destroy(header);
bam_destroy1(b[0]);
bam_destroy1(b[1]);
free(str.s);
}