// currently, this function ONLY works if each read has one hit
void bam_mating_core(bamFile in, bamFile out)
{
bam_header_t *header;
bam1_t *b[2];
int curr, has_prev, pre_end = 0, cur_end;
kstring_t str;
str.l = str.m = 0; str.s = 0;
header = bam_header_read(in);
bam_header_write(out, header);
b[0] = bam_init1();
b[1] = bam_init1();
curr = 0; has_prev = 0;
while (bam_read1(in, b[curr]) >= 0) {
bam1_t *cur = b[curr], *pre = b[1-curr];
if (cur->core.tid < 0) continue;
cur_end = bam_calend(&cur->core, bam1_cigar(cur));
if (cur_end > (int)header->target_len[cur->core.tid]) cur->core.flag |= BAM_FUNMAP;
if (cur->core.flag & BAM_FSECONDARY) continue; // skip secondary alignments
if (has_prev) {
if (strcmp(bam1_qname(cur), bam1_qname(pre)) == 0) { // identical pair name
cur->core.mtid = pre->core.tid; cur->core.mpos = pre->core.pos;
pre->core.mtid = cur->core.tid; pre->core.mpos = cur->core.pos;
if (pre->core.tid == cur->core.tid && !(cur->core.flag&(BAM_FUNMAP|BAM_FMUNMAP))
&& !(pre->core.flag&(BAM_FUNMAP|BAM_FMUNMAP))) // 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 (pre->core.flag&BAM_FREVERSE) cur->core.flag |= BAM_FMREVERSE;
else cur->core.flag &= ~BAM_FMREVERSE;
if (cur->core.flag&BAM_FREVERSE) pre->core.flag |= BAM_FMREVERSE;
else pre->core.flag &= ~BAM_FMREVERSE;
if (cur->core.flag & BAM_FUNMAP) { pre->core.flag |= BAM_FMUNMAP; pre->core.flag &= ~BAM_FPROPER_PAIR; }
if (pre->core.flag & BAM_FUNMAP) { cur->core.flag |= BAM_FMUNMAP; cur->core.flag &= ~BAM_FPROPER_PAIR; }
bam_template_cigar(pre, cur, &str);
bam_write1(out, pre);
bam_write1(out, cur);
has_prev = 0;
} else { // unpaired or singleton
pre->core.mtid = -1; pre->core.mpos = -1; pre->core.isize = 0;
if (pre->core.flag & BAM_FPAIRED) {
pre->core.flag |= BAM_FMUNMAP;
pre->core.flag &= ~BAM_FMREVERSE & ~BAM_FPROPER_PAIR;
}
bam_write1(out, pre);
}
} else has_prev = 1;
curr = 1 - curr;
pre_end = cur_end;
}
if (has_prev) bam_write1(out, b[1-curr]);
bam_header_destroy(header);
bam_destroy1(b[0]);
bam_destroy1(b[1]);
free(str.s);
}
int bam_pad2unpad(bamFile in, bamFile out)
{
bam_header_t *h;
bam1_t *b;
kstring_t r, q;
uint32_t *cigar2 = 0;
int n2 = 0, m2 = 0, *posmap = 0;
h = bam_header_read(in);
bam_header_write(out, h);
b = bam_init1();
r.l = r.m = q.l = q.m = 0; r.s = q.s = 0;
while (bam_read1(in, b) >= 0) {
uint32_t *cigar = bam1_cigar(b);
n2 = 0;
if (b->core.pos == 0 && b->core.tid >= 0 && strcmp(bam1_qname(b), h->target_name[b->core.tid]) == 0) {
int i, k;
unpad_seq(b, &r);
write_cigar(cigar2, n2, m2, bam_cigar_gen(b->core.l_qseq, BAM_CMATCH));
replace_cigar(b, n2, cigar2);
posmap = realloc(posmap, r.m * sizeof(int));
for (i = k = 0; i < r.l; ++i) {
posmap[i] = k; // note that a read should NOT start at a padding
if (r.s[i]) ++k;
}
} else {
int i, k, op;
unpad_seq(b, &q);
if (bam_cigar_op(cigar[0]) == BAM_CSOFT_CLIP) write_cigar(cigar2, n2, m2, cigar[0]);
for (i = 0, k = b->core.pos; i < q.l; ++i, ++k)
q.s[i] = q.s[i]? (r.s[k]? BAM_CMATCH : BAM_CINS) : (r.s[k]? BAM_CDEL : BAM_CPAD);
for (i = k = 1, op = q.s[0]; i < q.l; ++i) {
if (op != q.s[i]) {
write_cigar(cigar2, n2, m2, bam_cigar_gen(k, op));
op = q.s[i]; k = 1;
} else ++k;
}
write_cigar(cigar2, n2, m2, bam_cigar_gen(k, op));
if (bam_cigar_op(cigar[b->core.n_cigar-1]) == BAM_CSOFT_CLIP) write_cigar(cigar2, n2, m2, cigar[b->core.n_cigar-1]);
for (i = 2; i < n2; ++i)
if (bam_cigar_op(cigar2[i]) == BAM_CMATCH && bam_cigar_op(cigar2[i-1]) == BAM_CPAD && bam_cigar_op(cigar2[i-2]) == BAM_CMATCH)
cigar2[i] += cigar2[i-2], cigar2[i-2] = cigar2[i-1] = 0;
for (i = k = 0; i < n2; ++i)
if (cigar2[i]) cigar2[k++] = cigar2[i];
n2 = k;
replace_cigar(b, n2, cigar2);
b->core.pos = posmap[b->core.pos];
}
bam_write1(out, b);
}
free(r.s); free(q.s); free(posmap);
bam_destroy1(b);
bam_header_destroy(h);
return 0;
}
int samwrite(samfile_t *fp, const bam1_t *b)
{
if (fp == 0 || (fp->type & TYPE_READ)) return -1; // not open for writing
if (fp->type & TYPE_BAM) return bam_write1(fp->x.bam, b);
else {
char *s = bam_format1_core(fp->header, b, fp->type>>2&3);
int l = strlen(s);
fputs(s, fp->x.tamw); fputc('\n', fp->x.tamw);
free(s);
return l + 1;
}
}
int main(int argc, char** argv)
{
if(argc < 3) {
printf("No input nor output files provided");
return -1;
}
bamFile in = bam_open(argv[1], "r");
bam_header_t* header;
if (in == NULL) {
printf("opening input file failed");
return -1;
}
bam1_t* b = bam_init1();
bamFile out = bam_open(argv[2], "w");
if (out == NULL) {
printf("opening input file failed");
return -1;
}
header = bam_header_read(in);
if(bam_header_write(out, header) < 0) {
printf("writing header failed");
}
long nextPrunedId;
if(!scanf ("%lu", &nextPrunedId)) {
printf("warning: no ids provided");
return -1;
}
long id = 0;
while (bam_read1(in, b) >= 0) {
// write BAM back
if (nextPrunedId != id++) {
bam_write1(out, b);
} else {
// fprintf(stderr, "pruning: id: %lu, pos: %d, length: %d\n", nextPrunedId, b->core.pos, b->core.l_qseq);
if(!scanf ("%lu", &nextPrunedId)) {
break;
}
}
}
// closing all resources
bam_header_destroy(header);
bam_close(in);
bam_close(out);
bam_destroy1(b);
return 0;
}
int samwrite(samfile_t *fp, const bam1_t *b)
{
if (fp == 0 || (fp->type & TYPE_READ)) return -1; // not open for writing
if (fp->type & TYPE_BAM) return bam_write1(fp->x.bam, b);
else {
char *s = bam_format1_core(fp->header, b, fp->type>>2&3);
int l = strlen(s);
fp->x.tamw.writer(fp->x.tamw.writer_data, (uint8_t*) s, strlen(s));
fp->x.tamw.writer(fp->x.tamw.writer_data, (uint8_t*) "\n", 1);
free(s);
return l + 1;
}
}
void convert_sam_to_bam(char* sam_input, char* bam_input) {
bam1_t* bam_p = bam_init1();
LOG_DEBUG("CONVERT-START: sam to bam\n");
//open SAM file for read
if (time_flag) {
start_timer(t1_convert);
}
tamFile sam_fd = sam_open(sam_input);
//open BAM file for write
bam_file_t* bam_file_p = bam_fopen_mode(bam_input, NULL, "w");
//read header from SAM file
bam_header_t* bam_header_p = sam_header_read(sam_fd);
//write header to BAM file
bam_header_write(bam_file_p->bam_fd, bam_header_p);
//write alignments to BAM file
while (sam_read1(sam_fd, bam_header_p, bam_p) > 0) {
bam_write1(bam_file_p->bam_fd, bam_p);
num_alignments++;
}
//close BAM and SAM files, free bam alignment and bam file object
bam_fclose(bam_file_p);
sam_close(sam_fd);
bam_header_destroy(bam_header_p);
bam_destroy1(bam_p);
if (time_flag) {
stop_timer(t1_convert, t2_convert, convert_time);
}
//number_of_batchs = 1, convention value for statistics (not real batch)
number_of_batchs = 1;
}
static int uniform_fetch_func(bam1_t *b, void *data)
{
uint8_t *to_delete;
data_t_uniform *tmp = (data_t_uniform*)data;
bam1_core_t *c = &b->core;
char *iq;
char *dq;
iq = malloc((c->l_qseq+1) * sizeof(char));
memset(iq, tmp->iq, c->l_qseq);
iq[c->l_qseq] = '\0';
to_delete = bam_aux_get(b, BI_TAG);
if (to_delete) {
bam_aux_del(b, to_delete);
}
bam_aux_append(b, BI_TAG, 'Z', c->l_qseq+1, (uint8_t*) iq);
dq = malloc((c->l_qseq+1) * sizeof(char));
memset(dq, tmp->dq, c->l_qseq);
dq[c->l_qseq] = '\0';
to_delete = bam_aux_get(b, BD_TAG);
if (to_delete) {
bam_aux_del(b, to_delete);
}
bam_aux_append(b, BD_TAG, 'Z', c->l_qseq+1, (uint8_t*) dq);
bam_write1(tmp->out, b);
free(iq);
free(dq);
return 0;
}
static int dindel_fetch_func(bam1_t *b, void *data)
{
data_t_dindel *tmp = (data_t_dindel*)data;
bam1_core_t *c = &b->core;
int rlen;
uint8_t *to_delete;
/* don't change reads failing default mask: BAM_FUNMAP | BAM_FSECONDARY | BAM_FQCFAIL | BAM_FDUP */
if (c->flag & BAM_DEF_MASK) {
/* fprintf(stderr, "skipping read: %s at pos %d\n", bam1_qname(b), c->pos); */
bam_write1(tmp->out, b);
return 0;
}
/* get the reference sequence and compute homopolymer array */
if (tmp->tid != c->tid) {
/*fprintf(stderr, "fetching reference sequence %s\n",
tmp->in->header->target_name[c->tid]); */
char *ref = fai_fetch(tmp->fai, tmp->in->header->target_name[c->tid], &rlen);
strtoupper(ref);/* safeguard */
int rlen = strlen(ref);
tmp->tid = c->tid;
if (tmp->hpcount) free(tmp->hpcount);
tmp->hpcount = (int*)malloc(rlen*sizeof(int));
find_homopolymers(ref, tmp->hpcount, rlen);
free(ref);
tmp->rlen = rlen;
/* fprintf(stderr, "fetched reference sequence\n");*/
}
/* parse the cigar string */
uint32_t *cigar = bam1_cigar(b);
uint8_t indelq[c->l_qseq+1];
/* fprintf(stderr, "l_qseq:%d\n", c->l_qseq); */
int i;
int x = c->pos; /* coordinate on reference */
int y = 0; /* coordinate on query */
for (i = 0; i < c->n_cigar; ++i) {
int j, oplen = cigar[i]>>4, op = cigar[i]&0xf;
if (op == BAM_CMATCH || op == BAM_CEQUAL || op == BAM_CDIFF) {
for (j = 0; j < oplen; j++) {
/*fprintf(stderr, "query:%d, ref:%d, count:%d\n",
y, x, tmp->hpcount[x+1]); */
/* FIXME clang complains: The left operand of '>' is a garbage value */
indelq[y] = (x > tmp->rlen-2) ? DINDELQ[0] : (tmp->hpcount[x+1]>18 ?
DINDELQ[0] : DINDELQ[tmp->hpcount[x+1]]);
x++;
y++;
}
} else if (op == BAM_CHARD_CLIP) { /* do nothing */
} else if (op == BAM_CDEL) {
x += oplen;
} else if (op == BAM_CINS || op == BAM_CSOFT_CLIP) {
for (j = 0; j < oplen; j++) {
/* fprintf(stderr, "query:%d, ref:%d\n", y, x); */
indelq[y] = DINDELQ[0];
y++;
}
} else {
LOG_FATAL("unknown op %d for read %s\n", op, bam1_qname(b));/* FIXME skip? seen this somewhere else properly handled */
exit(1);
}
}
indelq[y] = '\0';
to_delete = bam_aux_get(b, BI_TAG);
if (to_delete) {
bam_aux_del(b, to_delete);
}
bam_aux_append(b, BI_TAG, 'Z', c->l_qseq+1, indelq);
to_delete = bam_aux_get(b, BD_TAG);
if (to_delete) {
bam_aux_del(b, to_delete);
}
bam_aux_append(b, BD_TAG, 'Z', c->l_qseq+1, indelq);
bam_write1(tmp->out, b);
return 0;
}
void filterReads(char * inBamFile,
char * outBamFile,
int minMapQual,
int minLen,
int maxMisMatches,
float minPcId,
float minPcAln,
int ignoreSuppAlignments,
int ignoreSecondaryAlignments) {
//
int result = -1;
int outResult = -1;
int supp_check = 0x0;
if (ignoreSuppAlignments) {
supp_check |= BAM_FSUPPLEMENTARY;
}
if (ignoreSecondaryAlignments) {
supp_check |= BAM_FSECONDARY;
}
// helper variables
BGZF* in = 0;
BGZF* out = 0;
bam1_t *b = bam_init1();
bam_hdr_t *h;
// open bam
if ((in = bgzf_open(inBamFile, "r")) == 0) {
fprintf(stderr,
"ERROR: Failed to open \"%s\" for reading.\n",
inBamFile);
}
else if ((h = bam_hdr_read(in)) == 0) { // read header
fprintf(stderr,
"ERROR: Failed to read BAM header of file \"%s\".\n",
inBamFile);
}
else if ((out = bgzf_open(outBamFile, "w")) == 0) {
fprintf(stderr,
"ERROR: Failed to open \"%s\" for writing.\n",
outBamFile);
}
else {
// write and destroy header
bam_hdr_write(out, h);
bam_hdr_destroy(h);
int line = 0;
int matches, mismatches, qLen;
float pcAln, pcId;
int showStats = 0;
// fetch alignments
while ((result = bam_read1(in, b)) >= 0) {
line += 1;
// only primary mappings
if ((b->core.flag & supp_check) != 0) {
if (showStats)
fprintf(stdout, "Rejected %d, non-primary\n", line);
continue;
}
// only high quality
if (b->core.qual < minMapQual) {
if (showStats)
fprintf(stdout, "Rejected %d, quality: %d\n", line, b->core.qual);
continue;
}
// not too many absolute mismatches
mismatches = bam_aux2i(bam_aux_get(b, "NM"));
if (mismatches > maxMisMatches) {
if (showStats)
fprintf(stdout, "Rejected %d, mismatches: %d\n", line, mismatches);
continue;
}
// not too short
qLen = bam_cigar2qlen((&b->core)->n_cigar, bam_get_cigar(b));
if (qLen < minLen) {
if (showStats)
fprintf(stdout, "Rejected %d, length: %d\n", line, qLen);
continue;
}
// only high percent identity
matches = bam_cigar2matches((&b->core)->n_cigar, bam_get_cigar(b));
pcId = (matches - mismatches) / (float)matches; // percentage as float between 0 to 1
if (pcId < minPcId) {
if (showStats)
fprintf(stdout, "Rejected %d, identity pc: %.4f\n", line, pcId);
continue;
}
// only high percent alignment
pcAln = matches / (float)qLen; // percentage as float between 0 to 1
if (pcAln < minPcAln) {
if (showStats)
fprintf(stdout, "Rejected %d, alignment pc: %.4f\n", line, pcAln);
continue;
}
if ((outResult = bam_write1(out, b)) < -1) {
fprintf(stderr,
"ERROR: Attempt to write read no. %d to file \"%s\" failed with code %d.\n",
line, outBamFile, outResult);
}
}
if (result < -1) {
fprintf(stderr,
"ERROR: retrieval of read no. %d from file \"%s\" failed with code %d.\n",
line, inBamFile, result);
}
}
if (in) bgzf_close(in);
if (out) bgzf_close(out);
bam_destroy1(b);
}
int main(int argc, char *argv[])
{
short out2stdout=0;
hashtable ht=new_hashtable(HASHSIZE);
bamFile in,in2;
bamFile out;
int paired;//1 if not paired or pair read 1, 2 otherwise
index_mem=sizeof(hashtable)*sizeof(hashnode**)*HASHSIZE*2;
if (argc != 3) {
fprintf(stderr, "Usage: bam_fix_NH <in.bam> <out.bam or - for stdout>\n");
return 1;
}
// Open file and exit if error
in = bam_open(argv[1], "rb");
out2stdout = strcmp(argv[2], "-")? 0 : 1;
out = strcmp(argv[2], "-")? bam_open(argv[2], "w") : bam_dopen(fileno(stdout), "w");
if (in == 0 ) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[1]);
return 1;
}
if (out == 0) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[2]);
return 1;
}
unsigned long num_alns=0;
int ref;
// ***********
// Copy header
bam_header_t *header;
header = bam_header_read(in);
bam_header_write(out,header);
// sorted by name?
// Should not rely on the value in SO
bam1_t *aln=bam_init1();
bam1_t *prev=bam_init1();
if (!out2stdout) {
fprintf(stderr,"bam_fix_NH version %s\n",VERSION);
fprintf(stderr,"Processing %s\n",argv[1]);
fprintf(stderr,"Hashing...\n");fflush(stderr);
}
while(bam_read1(in,aln)>=0) { // read alignment
if (aln->core.tid < 0) continue;//ignore unaligned reads
if (aln->core.flag & BAM_FUNMAP) continue;
if (aln->core.flag & BAM_FREAD2) paired=2;
else paired=1;
++num_alns;
new_read_aln(ht,fix_read_name(bam1_qname(aln),paired));
if(!out2stdout) PRINT_ALNS_PROCESSED(num_alns);
}
bam_close(in);
if(!out2stdout) {
fprintf(stderr,"%s%lu\n",BACKLINE,num_alns);
fprintf(stderr,"Hashing complete (%lu alignments)\n",num_alns);
fprintf(stderr,"Memory used: %ld MB\n",index_mem/1024/1024);
fprintf(stderr,"Updating entries with NH and printing BAM...\n");
fflush(stderr);
}
// reopen
in2 = bam_open(argv[1], "rb");
if (in2 == 0 ) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[1]);
return 1;
}
header = bam_header_read(in2);
num_alns=0;
while(bam_read1(in2,aln)>=0) { // read alignment
paired=1;
if (aln->core.tid < 0) continue;//ignore unaligned reads
if (aln->core.flag & BAM_FUNMAP) continue;
if (aln->core.flag & BAM_FREAD2) paired=2;
++num_alns;
READ_ALN *r=get_read_aln(ht,fix_read_name(bam1_qname(aln),paired));
assert(r!=NULL);
// update the NH field
uint8_t *old_nh = bam_aux_get(aln, "NH");
int32_t nh=r->ctr;
if (old_nh) {
if (nh!=bam_aux2i(old_nh)) {
fprintf(stderr,"warning: value mismatch! replacing>%s %d->%d\n",bam1_qname(aln),bam_aux2i(old_nh),nh);
}
bam_aux_del(aln, old_nh);
bam_aux_append(aln, "NH", 'i', 4, (uint8_t*)&nh);
#ifdef DEBUG
// printf("!>%s %d\n",bam1_qname(aln),r->ctr);
#endif
}
if (!old_nh) { // add NH
bam_aux_append(aln, "NH", 'i', 4, (uint8_t*)&nh);
#ifdef DEBUG
fprintf(stderr,"!>%s %d\n",bam1_qname(aln),bam_aux2i(old_nh));
#endif
}
bam_write1(out,aln);
if(!out2stdout) PRINT_ALNS_PROCESSED(num_alns);
}
//
bam_destroy1(aln);
bam_close(in2);
bam_close(out);
if(!out2stdout) {
fprintf(stderr,"%s%lu\n",BACKLINE,num_alns);
fprintf(stderr,"Done.\n");
}
return 0;
}
int main(int argc, char *argv[])
{
hashtable ht=new_hashtable(HASHSIZE);
bamFile in,in2;
bamFile out;
if (argc != 3) {
fprintf(stderr, "Usage: bam_fix_NH <in.bam> <out.bam>\n");
return 1;
}
// Open file and exit if error
//in = strcmp(argv[1], "-")? bam_open(argv[1], "rb") : bam_dopen(fileno(stdin), "rb");
in = bam_open(argv[1], "rb");
out = strcmp(argv[2], "-")? bam_open(argv[2], "w") : bam_dopen(fileno(stdout), "w");
if (in == 0 ) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[1]);
return 1;
}
if (out == 0) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[2]);
return 1;
}
unsigned long num_alns=0;
int ref;
// ***********
// Copy header
bam_header_t *header;
header = bam_header_read(in);
bam_header_write(out,header);
// sorted by name?
// Should not rely on the value in SO
bam1_t *aln=bam_init1();
bam1_t *prev=bam_init1();
printf("Hashing...\n");flush(stdout);
while(bam_read1(in,aln)>=0) { // read alignment
if (aln->core.tid < 0) continue;//ignore unaligned reads
++num_alns;
new_read_aln(ht,bam1_qname(aln));
}
bam_close(in);
printf("Hashing complete (%lu alignments)\n",num_alns);
printf("Memory used in the hash: %ld MB\n",index_mem/1024/1024);
flush(stdout);
// reopen
in2 = bam_open(argv[1], "rb");
if (in2 == 0 ) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[1]);
return 1;
}
header = bam_header_read(in2);
while(bam_read1(in2,aln)>=0) { // read alignment
if (aln->core.tid < 0) continue;//ignore unaligned reads
++num_alns;
READ_ALN *r=get_read_aln(ht,bam1_qname(aln));
//assert(r!=NULL);
// update the NH field
uint8_t *old_nh = bam_aux_get(aln, "NH");
uint8_t nh=r->ctr;
if (old_nh) {
if (nh!=bam_aux2i(old_nh)) {
fprintf(stderr,"warning: value mismatch! replacing>%s %d->%d\n",bam1_qname(aln),bam_aux2i(old_nh),nh);
}
bam_aux_del(aln, old_nh);
bam_aux_append(aln, "NH", 'i', 4, (uint8_t*)&nh);
}
if (!old_nh) { // add NH
bam_aux_append(aln, "NH", 'i', 4, (uint8_t*)&nh);
#ifdef DEBUG
printf("!>%s %d\n",bam1_qname(aln),bam_aux2i(old_nh));
#endif
}
// in->header
// Also fix the XS:A tag
// BAM_FREAD1
// BAM_FREAD2
// BAM_FREVERSE the read is mapped to the reverse strand
//bam1_cigar(b)
//BAM_CREF_SKIP 3 CIGAR skip on the reference (e.g. spliced alignment)
//BAM_FREVERSE 16 the read is mapped to the reverse strand
if (aln->core.flag & BAM_FSECONDARY) continue; // skip secondary alignments
if (aln->core.flag & ! BAM_FPAIRED) continue; // not paired
if (aln->core.flag & ! BAM_FPROPER_PAIR) continue; // not a proper pair
if (aln->core.flag & ! BAM_FMUNMAP) continue; // the mate is mapped
if (aln->core.flag & BAM_FSECONDARY) continue; // secundary read
if (aln->core.flag & BAM_FREAD2) continue; // only count each pair once
// core.strand == 0 (f/+) 1 r/-
// flag
// bam1_qname(b)
bam_write1(out,aln);
}
//
bam_destroy1(aln);
bam_close(in2);
bam_close(out);
return 0;
/*
uint8_t *old_nm = bam_aux_get(b, "NM");
90 if (c->flag & BAM_FUNMAP) return;
91 if (old_nm) old_nm_i = bam_aux2i(old_nm);
92 if (!old_nm) bam_aux_append(b, "NM", 'i', 4, (uint8_t*)&nm);
93 else if (nm != old_nm_i) {
94 fprintf(stderr, "[bam_fillmd1] different NM for read '%s': %d -> %d\n", bam1_qname(b), old_nm_i, nm);
95 bam_aux_del(b, old_nm);
96 bam_aux_append(b, "NM", 'i', 4, (uint8_t*)&nm);
97 }
*/
}
int main(int argc, char *argv[])
{
short out2stdout=0;
bamFile in,in2;
bamFile out;
if (argc != 3) {
fprintf(stderr, "Usage: bam_fix_se_flag <in.bam> <out.bam or - for stdout>\n");
return 1;
}
// Open file and exit if error
in = bam_open(argv[1], "rb");
out2stdout = strcmp(argv[2], "-")? 0 : 1;
out = strcmp(argv[2], "-")? bam_open(argv[2], "w") : bam_dopen(fileno(stdout), "w");
if (in == 0 ) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[1]);
return 1;
}
if (out == 0) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[2]);
return 1;
}
unsigned long num_alns=0;
int ref;
// ***********
// Copy header
bam_header_t *header;
header = bam_header_read(in);
bam_header_write(out,header);
// sorted by name?
// Should not rely on the value in SO
bam1_t *aln=bam_init1();
bam1_t *prev=bam_init1();
if (!out2stdout) {
fprintf(stderr,"bam_fix_se_flag version %s\n",VERSION);
fprintf(stderr,"Processing %s\n",argv[1]);
}
// reopen
in2 = bam_open(argv[1], "rb");
if (in2 == 0 ) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[1]);
return 1;
}
header = bam_header_read(in2);
num_alns=0;
while(bam_read1(in2,aln)>=0) { // read alignment
if (aln->core.tid < 0) continue;//ignore unaligned reads
if (aln->core.flag & BAM_FUNMAP) continue;
if (aln->core.flag & BAM_FPAIRED ) { // PAIRED
} else { //SE
//turn off the other pair related flags
aln->core.flag&=~BAM_FPROPER_PAIR;
aln->core.flag&=~BAM_FMUNMAP;
aln->core.flag&=~BAM_FREAD1;
aln->core.flag&=~BAM_FREAD2;
fprintf(stderr, ".");
}
bam_write1(out,aln);
if(!out2stdout) PRINT_ALNS_PROCESSED(num_alns);
++num_alns;
}
//
bam_destroy1(aln);
bam_close(in2);
bam_close(out);
if(!out2stdout) {
fprintf(stderr,"%s%lu\n",BACKLINE,num_alns);
fprintf(stderr,"Done.\n");
}
return 0;
}
// FIX MRNM and unaligned reads
int main(int argc, char *argv[])
{
bamFile in;
long num_unmapped=0;
long num_alns_pe=0;
if (argc != 3) {
fprintf(stderr, "Usage: bam_tophat2_pe_fix <in.bam> <out.bam>\n");
return 1;
}
in = strcmp(argv[1], "-")? bam_open(argv[1], "rb") : bam_dopen(fileno(stdin), "rb");
if (in == 0) {
fprintf(stderr, "ERROR: Fail to open input BAM file %s\n", argv[1]);
return 1;
}
int ref;
unsigned long num_alns=0;
// counts
unsigned long unalign_mapq_fix=0;
unsigned long mtid_fix=0;
unsigned long mpos_fix=0;
bamFile out;
bam_header_t *header;
header = bam_header_read(in);
bam1_t *aln=bam_init1();
out = strcmp(argv[2], "-")? bam_open(argv[2], "w") : bam_dopen(fileno(stdout), "w");
if (out == 0) {
fprintf(stderr, "ERROR: Fail to open BAM file %s\n", argv[2]);
return 1;
}
bam_header_write(out,header);
while(bam_read1(in,aln)>=0) {
++num_alns;
if (aln->core.tid < 0) {
// unaligned reads
if ( aln->core.qual!=0 ) {
//fprintf(stderr, "ERROR: Unaligned read with quality > 0 in line %lu\n",num_alns);
aln->core.qual=0;
unalign_mapq_fix++;
}
}
//fprintf(stderr,"%s %c %d %d\n",bam1_qname(aln),(aln->core.tid<0?'U':'M'),aln->core.mtid,aln->core.mpos);
if ( aln->core.flag & BAM_FPAIRED ) {
//fprintf(stderr,"paired %d\n",(aln->core.flag & BAM_FMUNMAP));
// paired
if ( aln->core.mtid <0 && !(aln->core.flag & BAM_FMUNMAP) ) {
aln->core.flag |= BAM_FMUNMAP;
aln->core.mpos=-1;
mtid_fix++;
}
if ( aln->core.mpos <0 && !(aln->core.flag & BAM_FMUNMAP) ) {
aln->core.flag |= BAM_FMUNMAP;
aln->core.mtid=-1;
mpos_fix++;
}
}
bam_write1(out,aln);
}
bam_destroy1(aln);
bam_close(in);
bam_close(out);
//
fprintf(stderr,"unaligned MAPQ fixes: %lu\n",unalign_mapq_fix);
fprintf(stderr,"unaligned mtid fixes: %lu\n",mtid_fix);
fprintf(stderr,"unaligned mpos fixes: %lu\n",mpos_fix);
return 0;
}
// load a pair from a bam file
SR_Status SR_BamInStreamLoadPair(SR_BamNode** ppUpAlgn,
SR_BamNode** ppDownAlgn,
SR_BamInStream* pBamInStream,
bamFile* bam_writer_complete_bam)
{
khash_t(queryName)* pNameHashPrev = pBamInStream->pNameHashes[PREV_BIN];
khash_t(queryName)* pNameHashCurr = pBamInStream->pNameHashes[CURR_BIN];
int ret = 1;
while(ret > 0 && (ret = SR_BamInStreamLoadNext(pBamInStream)) > 0)
{
// exclude those reads who are non-paired-end, qc-fail, duplicate-marked, proper-paired?!,
// both aligned, secondary-alignment and no-name-specified.
SR_Bool shouldBeFiltered = pBamInStream->filterFunc(pBamInStream->pNewNode, pBamInStream->filterData);
if (shouldBeFiltered)
{
#ifdef VERBOSE_DEBUG
fprintf(stderr,"%s: filtered.\n", bam1_qname(&(pBamInStream->pNewNode->alignment)));
#endif
if (bam_writer_complete_bam != NULL) bam_write1(*bam_writer_complete_bam, &(pBamInStream->pNewNode->alignment));
SR_BamNodeFree(pBamInStream->pNewNode, pBamInStream->pMemPool);
pBamInStream->pNewNode = NULL;
continue;
} else {
#ifdef VERBOSE_DEBUG
fprintf(stderr,"%s: kept in buffer.\n", bam1_qname(&(pBamInStream->pNewNode->alignment)));
#endif
}
// update the current ref ID or position if the incoming alignment has a
// different value. The name hash and the bam array will be reset
if (pNameHashPrev != NULL
&& (pBamInStream->pNewNode->alignment.core.tid != pBamInStream->currRefID
|| pBamInStream->pNewNode->alignment.core.pos >= pBamInStream->currBinPos + 2 * pBamInStream->binLen))
{
if (pBamInStream->pNewNode->alignment.core.tid != pBamInStream->currRefID)
{
ret = SR_OUT_OF_RANGE; // different chromosome id
}
pBamInStream->currRefID = pBamInStream->pNewNode->alignment.core.tid;
pBamInStream->currBinPos = pBamInStream->pNewNode->alignment.core.pos;
// Clear the hash buffer
kh_clear(queryName, pNameHashPrev);
kh_clear(queryName, pNameHashCurr);
// Store alignments before releasing them
if (bam_writer_complete_bam != NULL) {
SR_BamNode* cur = pBamInStream->pAlgnLists[PREV_BIN].first;
for (int i = 0; i < pBamInStream->pAlgnLists[PREV_BIN].numNode; ++i) {
// if the cur is not NULL, store the cur in the complete bam
if (cur != NULL) bam_write1(*bam_writer_complete_bam, &(cur->alignment));
cur = cur->next;
} // end for
cur = pBamInStream->pAlgnLists[CURR_BIN].first;
for (int i = 0; i < pBamInStream->pAlgnLists[CURR_BIN].numNode; ++i) {
// if the cur is not NULL, store the cur in the complete bam
if (cur != NULL) bam_write1(*bam_writer_complete_bam, &(cur->alignment));
cur = cur->next;
} // end for
} // end if
SR_BamListReset(&(pBamInStream->pAlgnLists[PREV_BIN]), pBamInStream->pMemPool);
SR_BamListReset(&(pBamInStream->pAlgnLists[CURR_BIN]), pBamInStream->pMemPool);
}
else if (pBamInStream->pNewNode->alignment.core.pos >= pBamInStream->currBinPos + pBamInStream->binLen)
{
pBamInStream->currBinPos += pBamInStream->binLen;
kh_clear(queryName, pNameHashPrev);
SR_SWAP(pNameHashPrev, pNameHashCurr, khash_t(queryName)*);
// Store alignments before releasing them
if (bam_writer_complete_bam != NULL) {
SR_BamNode* cur = pBamInStream->pAlgnLists[PREV_BIN].first;
for (int i = 0; i < pBamInStream->pAlgnLists[PREV_BIN].numNode; ++i) {
// if the cur is not NULL, store the cur in the complete bam
if (cur != NULL) bam_write1(*bam_writer_complete_bam, &(cur->alignment));
cur = cur->next;
}
} // end if
SR_BamListReset(&(pBamInStream->pAlgnLists[PREV_BIN]), pBamInStream->pMemPool);
SR_SWAP(pBamInStream->pAlgnLists[PREV_BIN], pBamInStream->pAlgnLists[CURR_BIN], SR_BamList);
}
else
{
} // end if-elseif-else
