bed_file_t *bed_open(char *filename) {
size_t len;
char *data = mmap_file(&len, filename);
bed_file_t *bed_file = (bed_file_t *) malloc(sizeof(bed_file_t));
bed_file->filename = filename;
bed_file->data = data;
bed_file->data_len = len;
bed_file->header_entries = linked_list_new(COLLECTION_MODE_SYNCHRONIZED);
bed_file->records = linked_list_new(COLLECTION_MODE_SYNCHRONIZED);
return bed_file;
}
cal_t *convert_bwt_anchor_to_CAL(bwt_anchor_t *bwt_anchor, size_t read_start, size_t read_end) {
linked_list_t *linked_list = linked_list_new(COLLECTION_MODE_ASYNCHRONIZED);
seed_region_t *seed_region = seed_region_new(read_start, read_end,
bwt_anchor->start, bwt_anchor->end, 0, 0, 0);
linked_list_insert_first(seed_region, linked_list);
cal_t *cal = cal_new(bwt_anchor->chromosome + 1, bwt_anchor->strand,
bwt_anchor->start, bwt_anchor->end,
1, linked_list,
linked_list_new(COLLECTION_MODE_ASYNCHRONIZED));
return cal;
}
suffix_mng_t *suffix_mng_new(sa_genome3_t *genome) {
suffix_mng_t *p = (suffix_mng_t *) calloc(1, sizeof(suffix_mng_t));
int num_chroms = genome->num_chroms;
char *name;
Container *subject = (Container *) malloc(sizeof(Container));
bl_containerInit(subject, num_chroms, sizeof(char *));
linked_list_t **suffix_lists = (linked_list_t **) malloc (sizeof(linked_list_t *) * num_chroms);
for (int i = 0; i < num_chroms; i++) {
suffix_lists[i] = linked_list_new(COLLECTION_MODE_ASYNCHRONIZED);
name = calloc(64, sizeof(char));
sprintf(name, "%i", i);
bl_containerAdd(subject, &name);
}
p->num_seeds = 0;
p->num_chroms = num_chroms;
p->subject = subject;
p->suffix_lists = suffix_lists;
return p;
}
void test_insert()
{
// regular insert
linked_list* ll = linked_list_new();
linked_list_insert(ll, 2);
linked_list_insert(ll, 4);
linked_list_insert(ll, 6);
assert_equals(linked_list_size(ll), 3);
assert_true(linked_list_contains(ll, 2));
assert_true(linked_list_contains(ll, 4));
assert_true(linked_list_contains(ll, 6));
assert_false(linked_list_contains(ll, -1));
assert_false(linked_list_contains(ll, 0));
linked_list_clear(ll);
linked_list* ll = linked_list_new();
assert_equals(linked_list_size(ll), 0);
assert_false(linked_list_contains(ll, 1));
linked_list_insert(ll, 1);
assert_equals(linked_list_size(ll), 1);
assert_true(linked_list_contains(ll, 1));
linked_list_clear(ll);
// insert by index
// set by index
}
/**
* Initialize empty BAM framework data structure.
*/
void
bfwork_init(bam_fwork_t *fwork)
{
assert(fwork);
//Set all to zero
memset(fwork, 0, sizeof(bam_fwork_t));
//Create regions
fwork->regions_list = linked_list_new(COLLECTION_MODE_SYNCHRONIZED);
//Init locks
omp_init_lock(&fwork->regions_lock);
omp_init_lock(&fwork->free_slots);
omp_init_lock(&fwork->output_file_lock);
omp_init_lock(&fwork->reference_lock);
}
void test_lookup()
{
// get
linked_list* ll = linked_list_new();
int i;
for (i = 0; i < 100; i++)
{
linked_list_insert(ll, i * 2);
}
assert_equals(linked_list_get(ll, 0), 0);
assert_equals(linked_list_get(ll, 99), 99 * 2);
assert_equals(linked_list_get(ll, 50), 50 * 2);
// contains
assert_true(linked_list_contains(ll, 10));
assert_false(linked_list_contains(ll, 1));
// peek
//assert_equals(linked_list_first(ll), 0);
// last
//assert_equals(linked_list_last(ll), 99 * 2);
// index of
}
//====================================================================================
// apply_caling
//====================================================================================
int apply_caling(cal_seeker_input_t* input, batch_t *batch) {
mapping_batch_t *mapping_batch = batch->mapping_batch;
array_list_t *list = NULL;
size_t read_index, num_cals;
int min_seeds, max_seeds;
cal_t *cal;
array_list_t *cal_list;
fastq_read_t *read;
size_t num_chromosomes = input->genome->num_chromosomes + 1;
size_t num_targets = mapping_batch->num_targets;
size_t *targets = mapping_batch->targets;
size_t new_num_targets = 0;
array_list_t *region_list;
bwt_anchor_t *bwt_anchor_back, *bwt_anchor_forw;
linked_list_t *linked_list;
int anchor_nt, gap_nt;
seed_region_t *seed_region_start, *seed_region_end;
//max_seeds = input->cal_optarg->num_seeds;
// size_t *new_targets = (size_t *) calloc(num_targets, sizeof(size_t));
// set to zero
mapping_batch->num_to_do = 0;
for (size_t i = 0; i < num_targets; i++) {
read_index = targets[i];
read = array_list_get(read_index, mapping_batch->fq_batch);
region_list = mapping_batch->mapping_lists[read_index];
// for debugging
// LOG_DEBUG_F("%s\n", ((fastq_read_t *) array_list_get(read_index, mapping_batch->fq_batch))->id);
if (!list) {
list = array_list_new(1000,
1.25f,
COLLECTION_MODE_ASYNCHRONIZED);
}
if (array_list_get_flag(region_list) == 0 ||
array_list_get_flag(region_list) == 2) {
//We have normal and extend seeds (anchors)
max_seeds = (read->length / 15)*2 + 10;
num_cals = bwt_generate_cal_list_linked_list(region_list,
input->cal_optarg,
&min_seeds, &max_seeds,
num_chromosomes,
list, read->length,
input->cal_optarg->min_cal_size, 0);
} else {
//We have double anchors with smaller distance between they
//printf("Easy case... Two anchors and same distance between read gap and genome distance\n");
num_cals = 0;
for (int a = array_list_size(region_list) - 1; a >= 0; a -= 2) {
max_seeds = 2;
min_seeds = 2;
bwt_anchor_back = array_list_remove_at(a, region_list);
bwt_anchor_forw = array_list_remove_at(a - 1, region_list);
linked_list = linked_list_new(COLLECTION_MODE_ASYNCHRONIZED);
//Seed for the first anchor
anchor_nt = bwt_anchor_forw->end - bwt_anchor_forw->start;
//printf("\t seed0[%i-%i][%lu-%lu]\n", 0, anchor_nt - 1,
// bwt_anchor_forw->start, bwt_anchor_forw->end);
seed_region_start = seed_region_new(0, anchor_nt - 1,
bwt_anchor_forw->start, bwt_anchor_forw->end, 0, 0, 0);
//Seed for the first anchor
gap_nt = read->length - (anchor_nt + (bwt_anchor_back->end - bwt_anchor_back->start));
//printf("\t gap_nt = %i, anchor_nt = %i\n", gap_nt, anchor_nt);
//printf("\t seed1[%i-%i][%lu-%lu]\n", anchor_nt + gap_nt, read->length - 1,
// bwt_anchor_back->start + 1, bwt_anchor_back->end);
seed_region_end = seed_region_new(anchor_nt + gap_nt, read->length - 1,
bwt_anchor_back->start + 1, bwt_anchor_back->end, 1, 0, 0);
//The reference distance is 0 and the read distance not
//The read distance is 0 and the reference distance not
//if (seed_region_start->genome_end > seed_region_end->genome_start ||
// seed_region_start->read_end > seed_region_end->read_start) {
//array_list_clear(region_list, NULL);
//continue;
if (seed_region_end->genome_start - seed_region_start->genome_end < 5 ||
seed_region_end->read_start - seed_region_start->read_end < 5) {
seed_region_start->genome_end -= 5;
seed_region_start->read_end -= 5;
seed_region_end->genome_start += 5;
seed_region_end->read_start += 5;
}
linked_list_insert(seed_region_start, linked_list);
linked_list_insert_last(seed_region_end, linked_list);
cal = cal_new(bwt_anchor_forw->chromosome + 1,
bwt_anchor_forw->strand,
bwt_anchor_forw->start,
bwt_anchor_back->end + 1,
2,
linked_list,
linked_list_new(COLLECTION_MODE_ASYNCHRONIZED));
array_list_insert(cal, list);
num_cals++;
}
}
// for debugging
LOG_DEBUG_F("read %s : num. cals = %i, min. seeds = %i, max. seeds = %i\n",
read->id, num_cals, min_seeds, max_seeds);
/* if (num_cals == 0) {
int seed_size = 24;
//First, Delete old regions
array_list_clear(mapping_batch->mapping_lists[read_index], region_bwt_free);
//Second, Create new regions with seed_size 24 and 1 Mismatch
bwt_map_inexact_seeds_seq(read->sequence, seed_size, seed_size/2,
bwt_optarg, bwt_index,
mapping_batch->mapping_lists[read_index]);
num_cals = bwt_generate_cal_list_linked_list(mapping_batch->mapping_lists[mapping_batch->targets[i]],
input->cal_optarg,
&min_seeds, &max_seeds,
num_chromosomes,
list, read->length);
}*/
/*
for (size_t j = 0; j < num_cals; j++) {
cal = array_list_get(j, list);
LOG_DEBUG_F("\tchr: %i, strand: %i, start: %lu, end: %lu, num_seeds = %i, num. regions = %lu\n",
cal->chromosome_id, cal->strand, cal->start, cal->end, cal->num_seeds, cal->sr_list->size);
}
*/
// printf("min_seeds = %i, max_seeds = %i, min_limit = %i, num_cals = %i\n",
// min_seeds, max_seeds, min_limit, array_list_size(list));
// filter incoherent CALs
int founds[num_cals], found = 0;
for (size_t j = 0; j < num_cals; j++) {
founds[j] = 0;
cal = array_list_get(j, list);
LOG_DEBUG_F("\tcal %i of %i: sr_list size = %i (cal->num_seeds = %i) %i:%lu-%lu\n",
j, num_cals, cal->sr_list->size, cal->num_seeds,
cal->chromosome_id, cal->start, cal->end);
if (cal->sr_list->size > 0) {
int start = 0;
for (linked_list_item_t *list_item = cal->sr_list->first; list_item != NULL; list_item = list_item->next) {
seed_region_t *s = list_item->item;
LOG_DEBUG_F("\t\t:: star %lu > %lu s->read_start\n", start, s->read_start);
if (start > s->read_start) {
LOG_DEBUG("\t\t\t:: remove\n");
found++;
founds[j] = 1;
}
start = s->read_end + 1;
}
} else {
found++;
founds[j] = 1;
}
}
if (found) {
min_seeds = 100000;
max_seeds = 0;
cal_list = array_list_new(MAX_CALS, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
for (size_t j = 0; j < num_cals; j++) {
if (!founds[j]) {
cal = array_list_get(j, list);
cal->num_seeds = cal->sr_list->size;
if (cal->num_seeds > max_seeds) max_seeds = cal->num_seeds;
if (cal->num_seeds < min_seeds) min_seeds = cal->num_seeds;
array_list_insert(cal, cal_list);
array_list_set(j, NULL, list);
}
}
array_list_free(list, (void *) cal_free);
num_cals = array_list_size(cal_list);
list = cal_list;
}
// LOG_FATAL_F("num. cals = %i, min. seeds = %i, max. seeds = %i\n", num_cals, min_seeds, max_seeds);
// filter CALs by the number of seeds
cal_list = list;
list = NULL;
/*
int min_limit = input->cal_optarg->min_num_seeds_in_cal;
if (min_limit < 0) min_limit = max_seeds;
// min_limit -= 3;
if (min_seeds == max_seeds || min_limit <= min_seeds) {
cal_list = list;
list = NULL;
} else {
cal_list = array_list_new(MAX_CALS, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
for (size_t j = 0; j < num_cals; j++) {
cal = array_list_get(j, list);
if (cal->num_seeds >= min_limit) {
array_list_insert(cal, cal_list);
array_list_set(j, NULL, list);
}
}
array_list_clear(list, (void *) cal_free);
num_cals = array_list_size(cal_list);
}
*/
if (num_cals > MAX_CALS) {
for (size_t j = num_cals - 1; j >= MAX_CALS; j--) {
cal = (cal_t *) array_list_remove_at(j, cal_list);
cal_free(cal);
}
num_cals = array_list_size(cal_list);
}
// LOG_DEBUG_F("num. cals = %i, MAX_CALS = %i\n", num_cals, MAX_CALS);
if (num_cals > 0 && num_cals <= MAX_CALS) {
array_list_set_flag(2, cal_list);
targets[new_num_targets++] = read_index;
/*
int count1 = 0, count2 = 0;
// count number of sw to do
// method #1
// printf("method #1\n");
seed_region_t *s, *prev_s;
linked_list_iterator_t* itr;
for (size_t j = 0; j < num_cals; j++) {
prev_s = NULL;
cal = array_list_get(j, cal_list);
itr = linked_list_iterator_new(cal->sr_list);
s = (seed_region_t *) linked_list_iterator_curr(itr);
while (s != NULL) {
if ((prev_s == NULL && s->read_start != 0) || (prev_s != NULL)) {
// printf("\t\t\tcase 1\n");
count1++;
}
prev_s = s;
linked_list_iterator_next(itr);
s = linked_list_iterator_curr(itr);
}
if (prev_s != NULL && prev_s->read_end < read->length - 1) {
count1++;
// printf("\t\t\tcase 2 (%i < %i)\n", prev_s->read_end, read->length - 1);
}
linked_list_iterator_free(itr);
}
// method #2
printf("method #2\n");
for (size_t j = 0; j < num_cals; j++) {
cal = array_list_get(j, cal_list);
printf("\t: %i\n", j);
if (cal->sr_list->size > 0) {
int start = 0;
for (linked_list_item_t *list_item = cal->sr_list->first; list_item != NULL; list_item = list_item->next) {
seed_region_t *s = list_item->item;
printf("\t\t[%i|%i - %i|%i]\n", s->genome_start, s->read_start, s->read_end, s->genome_end);
if (s->read_start != start) {
count2++;
}
start = s->read_end + 1;
}
if (start < read->length) {
count2++;
}
}
}
printf("count #1 = %i, count #2 = %i\n", count1, count2);
assert(count1 == count2);
mapping_batch->num_to_do += count1;
*/
// we have to free the region list
array_list_free(mapping_batch->mapping_lists[read_index], (void *) region_bwt_free);
mapping_batch->mapping_lists[read_index] = cal_list;
} else {
array_list_set_flag(0, mapping_batch->mapping_lists[read_index]);
// we have to free the region list
array_list_clear(mapping_batch->mapping_lists[read_index], (void *) region_bwt_free);
if (cal_list) array_list_free(cal_list, (void *) cal_free);
if (list) array_list_clear(list, (void *) cal_free);
}
/*
cal_list = list;
list = NULL;
array_list_set_flag(2, cal_list);
// mapping_batch->num_to_do += num_cals;
targets[new_num_targets++] = read_index;
// we have to free the region list
array_list_free(mapping_batch->mapping_lists[read_index], (void *) region_bwt_free);
mapping_batch->mapping_lists[read_index] = cal_list;
*/
/*
// filter CALs by the number of seeds
int min_limit = input->cal_optarg->min_num_seeds_in_cal;
if (min_limit < 0) min_limit = max_seeds;
printf("min_seeds = %i, max_seeds = %i, min_limit = %i, num_cals = %i\n",
min_seeds, max_seeds, min_limit, array_list_size(list));
if (min_seeds == max_seeds || min_limit <= min_seeds) {
cal_list = list;
list = NULL;
} else {
cal_list = array_list_new(MAX_CALS, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
for (size_t j = 0; j < num_cals; j++) {
cal = array_list_get(j, list);
if (cal->num_seeds >= min_limit) {
array_list_insert(cal, cal_list);
array_list_set(j, NULL, list);
}
}
array_list_clear(list, (void *) cal_free);
num_cals = array_list_size(cal_list);
printf("************, num_cals = %i\n", num_cals);
}
if (num_cals > MAX_CALS) {
for (size_t j = num_cals - 1; j >= MAX_CALS; j--) {
cal = (cal_t *) array_list_remove_at(j, cal_list);
cal_free(cal);
}
num_cals = array_list_size(cal_list);
}
if (num_cals > 0 && num_cals <= MAX_CALS) {
array_list_set_flag(2, cal_list);
mapping_batch->num_to_do += num_cals;
targets[new_num_targets++] = read_index;
// we have to free the region list
array_list_free(mapping_batch->mapping_lists[read_index], (void *) region_bwt_free);
mapping_batch->mapping_lists[read_index] = cal_list;
} else {
array_list_set_flag(0, mapping_batch->mapping_lists[read_index]);
// we have to free the region list
array_list_clear(mapping_batch->mapping_lists[read_index], (void *) region_bwt_free);
if (cal_list) array_list_free(cal_list, (void *) cal_free);
if (list) array_list_clear(list, (void *) cal_free);
}
*/
} // end for 0 ... num_targets
// update batch
mapping_batch->num_targets = new_num_targets;
// LOG_DEBUG_F("num. SW to do: %i\n", mapping_batch->num_to_do);
// exit(-1);
// free memory
if (list) array_list_free(list, NULL);
if (batch->mapping_mode == RNA_MODE) {
return RNA_STAGE;
}
if (batch->pair_input->pair_mng->pair_mode != SINGLE_END_MODE) {
return PRE_PAIR_STAGE;
} else if (batch->mapping_batch->num_targets > 0) {
return SW_STAGE;
}
return DNA_POST_PAIR_STAGE;
}
void suffix_mng_create_cals(fastq_read_t *read, int min_area, int strand,
sa_index3_t *sa_index, array_list_t *cal_list,
suffix_mng_t *p) {
if (!p) return;
if (!p->suffix_lists) return;
if (p->num_seeds <= 0) return;
int read_area, chrom;
seed_t *seed;
seed_cal_t *cal;
linked_list_t *seed_list;
claspinfo_t info;
bl_claspinfoInit(&info);
// initialization
info.fragments = (Container *) malloc(sizeof(Container));
bl_containerInit(info.fragments, p->num_seeds, sizeof(slmatch_t));
info.subject = p->subject;
slmatch_t frag;
linked_list_t *suffix_list;
for (unsigned int i = 0; i < p->num_chroms; i++) {
suffix_list = p->suffix_lists[i];
if (suffix_list) {
for (linked_list_item_t *item = suffix_list->first;
item != NULL;
item = item->next) {
seed = item->item;
bl_slmatchInit(&frag, 0);
frag.i = seed->read_start;
frag.j = seed->read_end - seed->read_start + 1;
frag.p = seed->genome_start;
frag.q = seed->genome_end - seed->genome_start + 1;
frag.scr = seed->genome_end - seed->genome_start + 1;
frag.subject = seed->chromosome_id;
bl_containerAdd(info.fragments, &frag);
}
}
}
// sort fragments
qsort(info.fragments->contspace, bl_containerSize(info.fragments),
sizeof(slmatch_t), cmp_slmatch_qsort);
int begin = 0;
for (int i = 1; i <= bl_containerSize(info.fragments); i++){
// end of fragments list or different database sequence
// --> process fragment[begin]...fragment[i-1], write output
// and free chains (less memory consumption with large input files)
if (i == bl_containerSize(info.fragments) ||
((slmatch_t *) bl_containerGet(info.fragments, begin))->subject !=
((slmatch_t *) bl_containerGet(info.fragments, i))->subject){
if (info.chainmode == SOP){
// only use chaining without clustering if no ids are specified
bl_slClusterSop((slmatch_t *) info.fragments->contspace + begin, i - begin,
info.epsilon, info.lambda, info.maxgap);
}
else {
bl_slClusterLin((slmatch_t *) info.fragments->contspace + begin, i - begin,
info.epsilon, info.lambda, info.maxgap);
}
for (int j = begin; j < i; j++) {
slmatch_t *match = (slmatch_t *) bl_containerGet(info.fragments, j);
if (match->chain) {
slchain_t *chain = (slchain_t *) match->chain;
if (chain->scr >= info.minscore &&
bl_containerSize(chain->matches) >= info.minfrag) {
chrom = atoi(*(char **) bl_containerGet(info.subject, chain->subject));
read_area = 0;
seed_list = linked_list_new(COLLECTION_MODE_ASYNCHRONIZED);
for (int k = 0; k < bl_containerSize(chain->matches); k++){
slmatch_t *frag = *(slmatch_t **) bl_containerGet(chain->matches, k);
seed = seed_new(frag->i, frag->i + frag->j - 1, frag->p, frag->p + frag->q - 1);
seed->chromosome_id = chrom;
seed->strand = strand;
read_area += frag->j;
cigar_append_op(frag->j, '=', &seed->cigar);
linked_list_insert_last(seed, seed_list);
}
// extend seeds
cal = seed_cal_new(chrom, strand, chain->p, chain->p + chain->q - 1, seed_list);
cal->read = read;
extend_seeds(cal, sa_index);
seed_cal_update_info(cal);
if (cal->read_area >= min_area) {
array_list_insert(cal, cal_list);
} else {
seed_cal_free(cal);
}
}
bl_slchainDestruct(chain);
free(chain);
match->chain = NULL;
}
} // END OF for (j = begin; j < i; j++)
begin = i;
} // END OF if (i == bl_containerSize(info.fragments) ||
} // END OF for (i = 1; i <= bl_containerSize(info.fragments); i++)
// destruct everything
info.subject = NULL;
bl_claspinfoDestruct(&info);
// finally, clear suffix manager
suffix_mng_clear(p);
}
