/*
* Reads a file and outputs a new BAM file to fd with 'h' replaced as
* the header. No checks are made to the validity.
*/
int bam_reheader(BGZF *in, bam_hdr_t *h, int fd,
const char *arg_list, int add_PG)
{
BGZF *fp;
ssize_t len;
uint8_t *buf;
if (in->is_write) return -1;
buf = malloc(BUF_SIZE);
if (bam_hdr_read(in) == NULL) {
fprintf(stderr, "Couldn't read header\n");
free(buf);
return -1;
}
fp = bgzf_fdopen(fd, "w");
if (add_PG) {
// Around the houses, but it'll do until we can manipulate bam_hdr_t natively.
SAM_hdr *sh = sam_hdr_parse_(h->text, h->l_text);
if (sam_hdr_add_PG(sh, "samtools",
"VN", samtools_version(),
arg_list ? "CL": NULL,
arg_list ? arg_list : NULL,
NULL) != 0)
return -1;
free(h->text);
h->text = strdup(sam_hdr_str(sh));
h->l_text = sam_hdr_length(sh);
if (!h->text)
return -1;
sam_hdr_free(sh);
}
bam_hdr_write(fp, h);
if (in->block_offset < in->block_length) {
bgzf_write(fp, in->uncompressed_block + in->block_offset, in->block_length - in->block_offset);
bgzf_flush(fp);
}
while ((len = bgzf_raw_read(in, buf, BUF_SIZE)) > 0)
bgzf_raw_write(fp, buf, len);
free(buf);
fp->block_offset = in->block_offset = 0;
bgzf_close(fp);
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 bam_cat(int nfn, char * const *fn, const bam_hdr_t *h, const char* outbam)
{
BGZF *fp;
uint8_t *buf;
uint8_t ebuf[BGZF_EMPTY_BLOCK_SIZE];
const int es=BGZF_EMPTY_BLOCK_SIZE;
int i;
fp = strcmp(outbam, "-")? bgzf_open(outbam, "w") : bgzf_fdopen(fileno(stdout), "w");
if (fp == 0) {
fprintf(stderr, "[%s] ERROR: fail to open output file '%s'.\n", __func__, outbam);
return 1;
}
if (h) bam_hdr_write(fp, h);
buf = (uint8_t*) malloc(BUF_SIZE);
for(i = 0; i < nfn; ++i){
BGZF *in;
bam_hdr_t *old;
int len,j;
in = strcmp(fn[i], "-")? bgzf_open(fn[i], "r") : bgzf_fdopen(fileno(stdin), "r");
if (in == 0) {
fprintf(stderr, "[%s] ERROR: fail to open file '%s'.\n", __func__, fn[i]);
return -1;
}
if (in->is_write) return -1;
old = bam_hdr_read(in);
if (old == NULL) {
fprintf(stderr, "[%s] ERROR: couldn't read header for '%s'.\n",
__func__, fn[i]);
bgzf_close(in);
return -1;
}
if (h == 0 && i == 0) bam_hdr_write(fp, old);
if (in->block_offset < in->block_length) {
bgzf_write(fp, in->uncompressed_block + in->block_offset, in->block_length - in->block_offset);
bgzf_flush(fp);
}
j=0;
while ((len = bgzf_raw_read(in, buf, BUF_SIZE)) > 0) {
if(len<es){
int diff=es-len;
if(j==0) {
fprintf(stderr, "[%s] ERROR: truncated file?: '%s'.\n", __func__, fn[i]);
return -1;
}
bgzf_raw_write(fp, ebuf, len);
memcpy(ebuf,ebuf+len,diff);
memcpy(ebuf+diff,buf,len);
} else {
if(j!=0) bgzf_raw_write(fp, ebuf, es);
len-= es;
memcpy(ebuf,buf+len,es);
bgzf_raw_write(fp, buf, len);
}
j=1;
}
/* check final gzip block */
{
const uint8_t gzip1=ebuf[0];
const uint8_t gzip2=ebuf[1];
const uint32_t isize=*((uint32_t*)(ebuf+es-4));
if(((gzip1!=GZIPID1) || (gzip2!=GZIPID2)) || (isize!=0)) {
fprintf(stderr, "[%s] WARNING: Unexpected block structure in file '%s'.", __func__, fn[i]);
fprintf(stderr, " Possible output corruption.\n");
bgzf_raw_write(fp, ebuf, es);
}
}
bam_hdr_destroy(old);
bgzf_close(in);
}
free(buf);
bgzf_close(fp);
return 0;
}
