void precalculate_aux_values_for_annotation(int calculate_stats, int calculate_dp, int calculate_mq, vcf_record_t *record,
variant_stats_t **variant_stats, file_stats_t *file_stats, list_t *stats_list,
int *dp, int *mq0, double *mq) {
// Variant statistics
if (calculate_stats) {
get_variants_stats(&record, 1, NULL, NULL,0, stats_list, file_stats);
list_item_t *item = list_remove_item(stats_list);
*variant_stats = item->data_p;
list_item_free(item);
}
// Read depth
if (calculate_dp) {
int dp_pos = -1;
*dp = 0;
for (int j = 0; j < record->samples->size; j++) {
char *aux = strndup(record->format, record->format_len);
dp_pos = get_field_position_in_format("DP", aux);
free(aux);
if (dp_pos >= 0) {
aux = strdup((char*) array_list_get(j, record->samples));
*dp += atoi(get_field_value_in_sample(aux, dp_pos));
free(aux);
}
}
}
// Mapping quality
if (calculate_mq) {
for (int j = 0; j < record->samples->size; j++) {
char *aux = strndup(record->format, record->format_len);
int mq_pos = get_field_position_in_format("GQ", aux);
free(aux);
if (mq_pos < 0) {
continue;
}
aux = strdup((char*) array_list_get(j, record->samples));
int cur_gq = atoi(get_field_value_in_sample(aux, mq_pos));
free(aux);
// printf("sample = %s\tmq_pos = %d\tvalue = %d\n", array_list_get(j, record->samples), mq_pos,
// atoi(get_field_value_in_sample(strdup((char*) array_list_get(j, record->samples)), mq_pos)));
if (cur_gq == 0) {
(*mq0)++;
} else {
*mq += cur_gq * cur_gq;
}
}
*mq = sqrt(*mq / record->samples->size);
}
}
void display_sr_lists(char *msg, mapping_batch_t *mapping_batch) {
fastq_read_t *read;
array_list_t *fq_batch = mapping_batch->fq_batch;
size_t read_index;
cal_t *cal;
array_list_t *cal_list = NULL;
size_t num_cals, num_targets = mapping_batch->num_targets;
seed_region_t *s;
linked_list_iterator_t* itr;
LOG_DEBUG_F("%s\n", msg);
// debugging....
for (size_t i = 0; i < num_targets; i++) {
read_index = mapping_batch->targets[i];
read = (fastq_read_t *) array_list_get(read_index, fq_batch);
LOG_DEBUG_F("Read %s\n", read->id);
cal_list = mapping_batch->mapping_lists[read_index];
num_cals = array_list_size(cal_list);
if (num_cals <= 0) continue;
// processing each CAL from this read
for(size_t j = 0; j < num_cals; j++) {
// get cal and read index
cal = array_list_get(j, cal_list);
LOG_DEBUG_F("\tCAL #%i of %i (strand %i, chr %i:%lu-%lu), sr_list size = %i\n",
j, num_cals, cal->strand, cal->chromosome_id - 1, cal->start, cal->end, cal->sr_list->size);
itr = linked_list_iterator_new(cal->sr_list);
s = (seed_region_t *) linked_list_iterator_curr(itr);
while (s != NULL) {
LOG_DEBUG_F("\t\t%s (dist. %i)\t[%i|%i - %i|%i]\n",
(s->info ? new_cigar_code_string((cigar_code_t *) s->info) : ">>>>>> gap"),
(s->info ? ((cigar_code_t *) s->info)->distance : -1),
s->genome_start, s->read_start, s->read_end, s->genome_end);
linked_list_iterator_next(itr);
s = linked_list_iterator_curr(itr);
}
linked_list_iterator_free(itr);
}
}
}
void write_mapped_read(array_list_t *array_list, bam_file_t *bam_file) {
size_t num_items = array_list_size(array_list);
alignment_t *alig;
bam1_t *bam1;
for (size_t j = 0; j < num_items; j++) {
alig = (alignment_t *) array_list_get(j, array_list);
//printf("\t******** %i(%i)\n", j, num_items);
//printf("is null alig->name %i\n", (alig->query_name == NULL));
//printf("name = %s\n", alig->query_name);
//printf("read = %s\n", alig->sequence);
//printf("\t-----> %s\n", alig->cigar);
LOG_DEBUG("writting bam..\n");
//alignment_print(alig);
//exit(-1);
if (alig != NULL) {
bam1 = convert_to_bam(alig, 33);
bam_fwrite(bam1, bam_file);
bam_destroy1(bam1);
alignment_free(alig);
} else {
LOG_FATAL_F("alig is NULL, num_items = %lu\n", num_items);
}
//printf("\t**************** %i(%i)\n", j, num_items);
}
if (array_list) { array_list_free(array_list, NULL); }
}
void assoc_test(enum ASSOC_task test_type, vcf_record_t **variants, int num_variants, individual_t **samples, int num_samples,
const void *opt_input, list_t *output_list) {
int tid = omp_get_thread_num();
vcf_record_t *record;
individual_t *individual;
char *sample_data;
int gt_position;
int allele1, allele2;
// Affection counts
int A1 = 0, A2 = 0, U1 = 0, U2 = 0;
// Perform analysis for each variant
for (int i = 0; i < num_variants; i++) {
record = variants[i];
// LOG_DEBUG_F("[%d] Checking variant %.*s:%ld\n", tid, record->chromosome_len, record->chromosome, record->position);
A1 = 0; A2 = 0;
U1 = 0; U2 = 0;
gt_position = get_field_position_in_format("GT", strndup(record->format, record->format_len));
// Count over individuals
for (int j = 0; j < num_samples; j++) {
individual = samples[j];
sample_data = strdup(array_list_get(j, record->samples));
if (!get_alleles(sample_data, gt_position, &allele1, &allele2)) {
assoc_count_individual(individual, record, allele1, allele2, &A1, &A2, &U1, &U2);
}
free(sample_data);
}
// Finished counting: now compute the statistics
if (test_type == CHI_SQUARE) {
double assoc_basic_chisq = assoc_basic_test(A1, U1, A2, U2);
assoc_basic_result_t *result = assoc_basic_result_new(record->chromosome, record->chromosome_len,
record->position,
record->reference, record->reference_len,
record->alternate, record->alternate_len,
A1, A2, U1, U2, assoc_basic_chisq);
list_item_t *output_item = list_item_new(tid, 0, result);
list_insert_item(output_item, output_list);
} else if (test_type == FISHER) {
double p_value = assoc_fisher_test(A1, A2, U1, U2, (double*) opt_input);
assoc_fisher_result_t *result = assoc_fisher_result_new(record->chromosome, record->chromosome_len,
record->position,
record->reference, record->reference_len,
record->alternate, record->alternate_len,
A1, A2, U1, U2, p_value);
list_item_t *output_item = list_item_new(tid, 0, result);
list_insert_item(output_item, output_list);
}
// LOG_DEBUG_F("[%d] after adding %.*s:%ld\n", tid, record->chromosome_len, record->chromosome, record->position);
} // next variant
}
array_list_t *fastq_filter(array_list_t *reads, array_list_t *passed, array_list_t *failed, fastq_filter_options_t *options) {
fastq_read_t *read;
fastq_read_stats_t *fq_read_stats = fastq_read_stats_new();
fastq_read_stats_options_t *fq_read_stats_options = fastq_read_stats_options_new(options->min_length, options->max_length, 4);
// #pragma omp parallel for schedule(dynamic, 500000)
for(size_t i=0; i<reads->size; i++) {
// fastq_read_stats_init(fq_read_stats);
read = array_list_get(i, reads);
if(read->length >= options->min_length && read->length <= options->max_length) {
fastq_read_stats_se(read, fq_read_stats_options, fq_read_stats);
// fastq_read_stats_print(fq_read_stats);
if(fq_read_stats->quality_average >= options->min_quality && fq_read_stats->quality_average <= options->max_quality && fq_read_stats->Ns < options->max_Ns) {
array_list_insert(read, passed);
}else {
array_list_insert(read, failed);
}
}else {
// get read stats
array_list_insert(read, failed);
}
}
fastq_read_stats_free(fq_read_stats);
fastq_read_stats_options_free(fq_read_stats_options);
return passed;
}
int apply_bwt(bwt_server_input_t* input, batch_t *batch) {
//printf("APPLY BWT SERVER...\n");
mapping_batch_t *mapping_batch = batch->mapping_batch;
size_t num_reads = array_list_size(mapping_batch->fq_batch);
size_t num_mappings;
array_list_t *list;
size_t *unmapped_indices = mapping_batch->targets;
size_t num_unmapped = 0;
size_t num_anchors;
for (int i = 0; i < num_reads; i++) {
fastq_read_t *read = array_list_get(i, mapping_batch->fq_batch);
//printf("BWT: %s\n", read->id);
list = mapping_batch->mapping_lists[i];
array_list_set_flag(0, list);
num_mappings = bwt_map_inexact_read(read,
input->bwt_optarg_p,
input->bwt_index_p,
list);
if (array_list_get_flag(list) != 2) { //If flag 2, the read exceded the max number of mappings
if (array_list_get_flag(list) == 1) {
if (num_mappings > 0) {
num_anchors = bwt_search_pair_anchors(list, read->length);
if (num_anchors == 0) {
array_list_set_flag(NOT_ANCHORS, list);
}
} else {
array_list_set_flag(NOT_ANCHORS, list);
}
//printf("tot anchors found %i %s\n", num_anchors, read->id);
unmapped_indices[num_unmapped++] = i;
} else if (num_mappings <= 0) {
array_list_set_flag(0, list);
//printf("Read NO Mapped %i %s\n", num_anchors, read->id);
unmapped_indices[num_unmapped++] = i;
}
} else {
array_list_set_flag(ALIGNMENTS_EXCEEDED, list);
}
}
// array_list flag: 0 -> Not BWT Anchors found
// 1 -> One BWT Anchors found
// 2 -> Pair BWT Anchors found
// 3 -> Alignments found
// 4 -> Alignments exceded
mapping_batch->num_targets = num_unmapped;
if (batch->mapping_batch->num_targets > 0) {
return CAL_STAGE;
}
return DNA_POST_PAIR_STAGE;
}
char *get_annotation_non_missing_samples(vcf_record_t *record, char *empty_sample, int *output_len) {
char *result = calloc(16, sizeof(char));
int ns = 0;
for (int j = 0; j < record->samples->size; j++) {
if (strcmp(array_list_get(j, record->samples), empty_sample)) {
ns++;
}
}
sprintf(result, "NS=%d", ns);
*output_len = strlen(result);
return result;
}
int main( int argc, char **args ) {
//create a new array list.
friso_array_t array = new_array_list();
fstring keys[] = {
"chenmanwen", "yangqinghua",
"chenxin", "luojiangyan", "xiaoyanzi", "bibi",
"zhangrenfang", "yangjian",
"liuxiao", "pankai",
"chenpei", "liheng", "zhangzhigang", "zhgangyishao", "yangjiangbo",
"caizaili", "panpan", "xiaolude", "yintanwen"
};
int j, idx = 2, len = sizeof( keys ) / sizeof( fstring );
for ( j = 0; j < len; j++ ) {
array_list_add( array, keys[j] );
}
printf("length=%d, allocations=%d\n", array->length, array->allocs );
array_list_trim( array );
printf("after tirm length=%d, allocations=%d\n", array->length, array->allocs );
printf("idx=%d, value=%s\n", idx, ( fstring ) array_list_get( array, idx ) );
printf("\nAfter set %dth item.\n", idx );
array_list_set( array, idx, "chenxin__" );
printf("idx=%d, value=%s\n", idx, ( fstring ) array_list_get( array, idx ) );
printf("\nAfter remove %dth item.\n", idx );
array_list_remove( array, idx );
printf("length=%d, allocations=%d\n", array->length, array->allocs );
printf("idx=%d, value=%s\n", idx, ( fstring ) array_list_get( array, idx ) );
printf("\nInsert a item at %dth\n", idx );
array_list_insert( array, idx, "*chenxin*" );
printf("idx=%d, value=%s\n", idx, ( fstring ) array_list_get( array, idx ) );
free_array_list( array );
return 0;
}
int vcf_batch_print(FILE *fd, vcf_batch_t *batch) {
assert(fd);
assert(batch);
vcf_record_t *first_record = array_list_get(0, batch->records);
if (first_record) {
return fprintf(fd, "Batch with %zu/%zu records: begin in chr%s:%ld\n",
batch->records->size, batch->records->capacity,
first_record->chromosome, first_record->position);
} else {
return fprintf(fd, "Batch with %zu/%zu records (empty)\n",
batch->records->size, batch->records->capacity);
}
}
int bam1s_stats(array_list_t *bam1s,
bam_stats_options_t *opts,
array_list_t *bam1s_stats) {
bam1_t *bam1;
bam_stats_t *stats;
size_t num_items = array_list_size(bam1s);
for (int i = 0; i < num_items; i++) {
bam1 = array_list_get(i, bam1s);
stats = bam1_stats(bam1, opts);
array_list_insert(stats, bam1s_stats);
}
return array_list_size(bam1s_stats);
}
void test_array_list_get_and_indexof ()
{
array_list *list = create_array_list_simple ();
TEST_ASSERT_NOT_NULL (list);
char *strs1[] = {"test", "content", "random", "word", "code",
"free", "software", "no", "ideas"};
array_list_append_all (list, (void **) strs1, 9);
// try to get the index of an item
TEST_ASSERT_EQUAL_INT (array_list_index_of (list, strs1[2]), 2);
// try to get a item by its index
TEST_ASSERT_EQUAL_STR (array_list_get (list, 5), strs1[5]);
delete_array_list (list);
}
int insert_vcf_query_fields_list(array_list_t *list, sqlite3 *db) {
int rc;
sqlite3_stmt *stmt;
vcf_query_fields_t *fields;
char *error_message;
prepare_statement_vcf_query_fields(db, &stmt);
if (rc = sqlite3_exec(db, "BEGIN TRANSACTION", NULL, NULL, &error_message)) {
LOG_DEBUG_F("Stats databases failed: %s (%d)\n", rc, error_message);
}
int num_items = array_list_size(list);
for (int i = 0; i < num_items; i++) {
fields = array_list_get(i, list);
sqlite3_bind_text(stmt, 1, fields->chromosome, strlen(fields->chromosome), SQLITE_STATIC);
sqlite3_bind_int64(stmt, 2, fields->position);
sqlite3_bind_text(stmt, 3, fields->allele_ref, strlen(fields->allele_ref), SQLITE_STATIC);
sqlite3_bind_text(stmt, 4, fields->allele_maf, strlen(fields->allele_maf), SQLITE_STATIC);
sqlite3_bind_text(stmt, 5, fields->genotype_maf, strlen(fields->genotype_maf), SQLITE_STATIC);
sqlite3_bind_double(stmt, 6, fields->allele_maf_freq);
sqlite3_bind_double(stmt, 7, fields->genotype_maf_freq);
sqlite3_bind_int(stmt, 8, fields->missing_alleles);
sqlite3_bind_int(stmt, 9, fields->missing_genotypes);
sqlite3_bind_int(stmt, 10, fields->mendelian_errors);
sqlite3_bind_int(stmt, 11, fields->is_indel);
sqlite3_bind_double(stmt, 12, fields->cases_percent_dominant);
sqlite3_bind_double(stmt, 13, fields->controls_percent_dominant);
sqlite3_bind_double(stmt, 14, fields->cases_percent_recessive);
sqlite3_bind_double(stmt, 15, fields->controls_percent_recessive);
if (rc = sqlite3_step(stmt) != SQLITE_DONE) {
LOG_DEBUG_F("Stats databases failed: %s (%d)\n", sqlite3_errmsg(db), sqlite3_errcode(db));
}
sqlite3_reset(stmt);
}
if (rc = sqlite3_exec(db, "COMMIT TRANSACTION", NULL, NULL, &error_message)) {
LOG_DEBUG_F("Stats databases failed: %s (%d)\n", rc, error_message);
}
sqlite3_finalize(stmt);
}
int merge_interval(kh_pos_t* positions_read, char *max_chromosome_merged, unsigned long max_position_merged,
char **chromosome_order, int num_chromosomes, vcf_file_t **files,
shared_options_data_t *shared_options_data, merge_options_data_t *options_data, list_t *output_list) {
int num_entries = 0;
#pragma omp parallel for num_threads(shared_options_data->num_threads) reduction(+:num_entries)
for (int k = kh_begin(positions_read); k < kh_end(positions_read); k++) {
if (kh_exist(positions_read, k)) {
array_list_t *records_in_position = kh_value(positions_read, k);
assert(records_in_position);
vcf_record_t *record = ((vcf_record_file_link*) array_list_get(0, records_in_position))->record;
vcf_record_file_link **links = NULL;
int num_links = 0;
// Remove positions prior to the last chromosome:position to merge
int cmp_chrom = compare_chromosomes(record->chromosome, max_chromosome_merged, chromosome_order, num_chromosomes);
if (cmp_chrom < 0 || (cmp_chrom == 0 && compare_positions(record->position, max_position_merged) <= 0)) {
links = records_in_position->items;
num_links = records_in_position->size;
}
// Launch merge
if (num_links > 0) {
// printf("links[0] = %s:%ld in file %s\n", links[0]->record->chromosome, links[0]->record->position, links[0]->file->filename);
int err_code = 0;
vcf_record_t *merged = merge_position(links, num_links, files, options_data->num_files, options_data, &err_code);
if (!err_code) {
list_item_t *item = list_item_new(k, MERGED_RECORD, merged);
list_insert_item(item, output_list);
num_entries += 1;
}
// Free empty nodes (lists of records in the same position)
array_list_free(records_in_position, vcf_record_file_link_free);
kh_del(pos, positions_read, k);
}
} // End kh_exist
}
return num_entries;
}
vcf_record_t *vcf_record_copy(vcf_record_t *orig) {
vcf_record_t *record = (vcf_record_t*) calloc (1, sizeof(vcf_record_t));
record->chromosome = strndup(orig->chromosome, orig->chromosome_len);
record->chromosome_len = orig->chromosome_len;
record->position = orig->position;
record->id = strndup(orig->id, orig->id_len);
record->id_len = orig->id_len;
record->reference = strndup(orig->reference, orig->reference_len);
record->reference_len = orig->reference_len;
record->alternate = strndup(orig->alternate, orig->alternate_len);
record->alternate_len = orig->alternate_len;
record->filter = strndup(orig->filter, orig->filter_len);
record->filter_len = orig->filter_len;
record->info = strndup(orig->info, orig->info_len);
record->info_len = orig->info_len;
record->format = strndup(orig->format, orig->format_len);
record->format_len = orig->format_len;
record->samples = array_list_new(orig->samples->size + 1, 1.5, COLLECTION_MODE_ASYNCHRONIZED);
for (int i = 0; i < orig->samples->size; i++) {
array_list_insert(strdup(array_list_get(i, orig->samples)), record->samples);
}
return record;
}
size_t bwt_search_pair_anchors(array_list_t *list, unsigned int read_length) {
bwt_anchor_t *bwt_anchor;
int max_anchor_length = 0;
bwt_anchor_t *bwt_anchor_back, *bwt_anchor_forw;
int anchor_length_tmp, anchor_back, anchor_forw;
int strand = 0, type = 0;
int found_anchor = 0, found_double_anchor = 0;
const int MIN_ANCHOR = 25;
const int MIN_SINGLE_ANCHOR = 40;
//const int MIN_DOUBLE_ANCHOR = MIN_ANCHOR*2;
const int MAX_BWT_REGIONS = 50;
const int MAX_BWT_ANCHOR_DISTANCE = 500000;
array_list_t *anchor_list_tmp, *forward_anchor_list, *backward_anchor_list;
cal_t *cal;
int seed_size, gap_read, gap_genome;
array_list_t *backward_anchor_list_0 = array_list_new(MAX_BWT_REGIONS, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
array_list_t *forward_anchor_list_0 = array_list_new(MAX_BWT_REGIONS, 1.25f , COLLECTION_MODE_ASYNCHRONIZED);
array_list_t *backward_anchor_list_1 = array_list_new(MAX_BWT_REGIONS, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
array_list_t *forward_anchor_list_1 = array_list_new(MAX_BWT_REGIONS, 1.25f , COLLECTION_MODE_ASYNCHRONIZED);
array_list_t *big_anchor_list = array_list_new(MAX_BWT_REGIONS, 1.25f , COLLECTION_MODE_ASYNCHRONIZED);
//printf("Tot Anchors %i\n", array_list_size(list));
for (int i = 0; i < array_list_size(list); i++) {
bwt_anchor = array_list_get(i, list);
if (bwt_anchor->strand == 1) {
//printf("(-)bwt anchor %i:%lu-%lu (%i): \n", bwt_anchor->chromosome + 1, bwt_anchor->start, bwt_anchor->end, bwt_anchor->end - bwt_anchor->start + 1);
if (bwt_anchor->type == FORWARD_ANCHOR) {
array_list_insert(bwt_anchor, forward_anchor_list_1);
//printf("FORW\n");
} else {
array_list_insert(bwt_anchor, backward_anchor_list_1);
//printf("BACK\n");
}
} else {
//printf("(+)bwt anchor %i:%lu-%lu (%i): \n", bwt_anchor->chromosome + 1, bwt_anchor->start, bwt_anchor->end, bwt_anchor->end - bwt_anchor->start + 1);
if (bwt_anchor->type == FORWARD_ANCHOR) {
array_list_insert(bwt_anchor, forward_anchor_list_0);
//printf("FORW\n");
} else {
array_list_insert(bwt_anchor, backward_anchor_list_0);
//printf("BACK\n");
}
}
anchor_length_tmp = bwt_anchor->end - bwt_anchor->start + 1;
if (anchor_length_tmp > MIN_SINGLE_ANCHOR && anchor_length_tmp > max_anchor_length) {
max_anchor_length = anchor_length_tmp;
found_anchor = 1;
strand = bwt_anchor->strand;
type = bwt_anchor->type;
}
if (read_length - anchor_length_tmp < 16) {
array_list_insert(bwt_anchor, big_anchor_list);
}
}
array_list_clear(list, NULL);
if (array_list_size(big_anchor_list) > 0) {
for (int i = array_list_size(big_anchor_list) - 1; i >= 0; i--) {
//printf("Insert cal %i\n", i);
bwt_anchor = array_list_remove_at(i, big_anchor_list);
size_t seed_size = bwt_anchor->end - bwt_anchor->start;
if (bwt_anchor->type == FORWARD_ANCHOR) {
cal = convert_bwt_anchor_to_CAL(bwt_anchor, 0, seed_size);
} else {
cal = convert_bwt_anchor_to_CAL(bwt_anchor, read_length - seed_size - 1, read_length - 1);
}
array_list_insert(cal, list);
}
array_list_set_flag(SINGLE_ANCHORS, list);
goto exit;
}
for (int type = 1; type >= 0; type--) {
if (!type) {
forward_anchor_list = forward_anchor_list_1;
backward_anchor_list = backward_anchor_list_1;
//printf("Strand (+): %i-%i\n", array_list_size(forward_anchor_list), array_list_size(backward_anchor_list));
} else {
forward_anchor_list = forward_anchor_list_0;
backward_anchor_list = backward_anchor_list_0;
//printf("Strand (-): %i-%i\n", array_list_size(forward_anchor_list), array_list_size(backward_anchor_list));
}
int *set_forward = (int *)calloc(array_list_size(forward_anchor_list), sizeof(int));
int *set_backward = (int *)calloc(array_list_size(backward_anchor_list), sizeof(int));
//Associate Anchors (+)/(-)
for (int i = 0; i < array_list_size(forward_anchor_list); i++) {
if (set_forward[i]) { continue; }
bwt_anchor_forw = array_list_get(i, forward_anchor_list);
for (int j = 0; j < array_list_size(backward_anchor_list); j++) {
if (set_backward[j]) { continue; }
bwt_anchor_back = array_list_get(j, backward_anchor_list);
anchor_forw = (bwt_anchor_forw->end - bwt_anchor_forw->start + 1);
anchor_back = (bwt_anchor_back->end - bwt_anchor_back->start + 1);
anchor_length_tmp = anchor_forw + anchor_back;
//printf("\tCommpare %i:%lu-%lu with %i:%lu-%lu\n", bwt_anchor_forw->chromosome + 1,
// bwt_anchor_forw->start, bwt_anchor_forw->end, bwt_anchor_back->chromosome + 1,
// bwt_anchor_back->start, bwt_anchor_back->end);
if (bwt_anchor_forw->chromosome == bwt_anchor_back->chromosome &&
abs(bwt_anchor_back->start - bwt_anchor_forw->end) <= MAX_BWT_ANCHOR_DISTANCE &&
anchor_forw >= MIN_ANCHOR && anchor_back >= MIN_ANCHOR) {
if (bwt_anchor_back->start < bwt_anchor_forw->end) { continue; }
gap_read = read_length - (anchor_forw + anchor_back);
gap_genome = bwt_anchor_back->start - bwt_anchor_forw->end;
//printf("anchor_forw = %i, anchor_back = %i, gap_read = %i, gap_genome = %i\n",
// anchor_forw, anchor_back, gap_read, gap_genome);
int apply_flank = 0;
if (gap_read < 2 || gap_genome < 2) {
int gap;
if (gap_read < 0 && gap_genome < 0) {
gap = abs(gap_read) > abs(gap_genome) ? abs(gap_read) : abs(gap_genome);
} else if (gap_read < 0) {
gap = abs(gap_read);
} else if (gap_genome < 0) {
gap = abs(gap_genome);
} else {
gap = 2;
}
int flank = 5;
apply_flank = 1;
if (abs(gap) >= flank*2) {
//Solve read overlap
flank = abs(gap)/2 + flank/2;
}
//printf("\tgap = %i, flank = %i\n", gap, flank);
if (flank >= anchor_forw) {
bwt_anchor_forw->end -= anchor_forw/2;
} else {
bwt_anchor_forw->end -= flank;
}
if (flank >= anchor_back) {
bwt_anchor_back->start += anchor_back/2;
} else {
bwt_anchor_back->start += flank;
}
}
cal = convert_bwt_anchor_to_CAL(bwt_anchor_forw, 0, bwt_anchor_forw->end - bwt_anchor_forw->start);
//printf("INSERT-1 (%i)[%i:%lu-%lu]\n", cal->strand, cal->chromosome_id, cal->start, cal->end);
array_list_insert(cal, list);
seed_size = bwt_anchor_back->end - bwt_anchor_back->start + 1;
//if (bwt_anchor_forw->end + read_length >= bwt_anchor_back->start) {
//seed_region_t *seed_region = seed_region_new(read_length - seed_size, read_length - 1,
//bwt_anchor_back->start, bwt_anchor_back->end, 1);
//cal->end = bwt_anchor_back->end;
//linked_list_insert_last(seed_region, cal->sr_list);
//} else {
cal = convert_bwt_anchor_to_CAL(bwt_anchor_back, read_length - seed_size, read_length - 1);
//printf("INSERT-2 (%i)[%i:%lu-%lu]\n", cal->strand, cal->chromosome_id, cal->start, cal->end);
array_list_insert(cal, list);
if (array_list_size(list) > 5) {
free(set_backward);
free(set_forward);
goto exit;
}
array_list_set_flag(DOUBLE_ANCHORS, list);
found_double_anchor = 1;
set_forward[i] = 1;
set_backward[j] = 1;
break;
}
}
}
free(set_backward);
free(set_forward);
}
if (!found_double_anchor && found_anchor) {
//Not Double anchor found but one Yes!!
if (strand == 1) {
if (type == FORWARD_ANCHOR) {
anchor_list_tmp = forward_anchor_list_1;
} else {
anchor_list_tmp = backward_anchor_list_1;
}
} else {
if (type == FORWARD_ANCHOR) {
anchor_list_tmp = forward_anchor_list_0;
} else {
anchor_list_tmp = backward_anchor_list_0;
}
}
//printf("LIST SIZE %i\n", array_list_size(anchor_list_tmp));
for (int i = 0; i < array_list_size(anchor_list_tmp); i++) {
bwt_anchor = array_list_get(i, anchor_list_tmp);
size_t seed_size = bwt_anchor->end - bwt_anchor->start;
//array_list_insert(bwt_anchor_new(bwt_anchor->strand, bwt_anchor->chromosome,
// bwt_anchor->start, bwt_anchor->end, bwt_anchor->type), anchor_list);
if (bwt_anchor->type == FORWARD_ANCHOR) {
//printf("------------------------> start %i\n", 0);
cal = convert_bwt_anchor_to_CAL(bwt_anchor, 0, seed_size);
} else {
//printf("------------------------> start %i\n", read_length - seed_size);
cal = convert_bwt_anchor_to_CAL(bwt_anchor, read_length - seed_size - 1, read_length - 1);
}
array_list_insert(cal, list);
}
array_list_set_flag(SINGLE_ANCHORS, list);
}
exit:
array_list_free(forward_anchor_list_1, (void *)bwt_anchor_free);
array_list_free(backward_anchor_list_1, (void *)bwt_anchor_free);
array_list_free(forward_anchor_list_0, (void *)bwt_anchor_free);
array_list_free(backward_anchor_list_0, (void *)bwt_anchor_free);
array_list_free(big_anchor_list, (void *)bwt_anchor_free);
return array_list_size(list);
}
void fill_gaps(mapping_batch_t *mapping_batch, sw_optarg_t *sw_optarg,
genome_t *genome, int min_gap, int min_distance) {
int sw_count = 0;
fastq_read_t *read;
array_list_t *fq_batch = mapping_batch->fq_batch;
size_t read_index, read_len;
cal_t *cal;
array_list_t *cal_list = NULL;
size_t num_cals, num_targets = mapping_batch->num_targets;
char *revcomp_seq = NULL;
seed_region_t *s, *prev_s, *new_s;
linked_list_iterator_t* itr;
cigar_code_t *cigar_code;
size_t start, end;
size_t gap_read_start, gap_read_end, gap_read_len;
size_t gap_genome_start, gap_genome_end, gap_genome_len;
int left_flank, right_flank;
sw_prepare_t *sw_prepare;
array_list_t *sw_prepare_list = array_list_new(1000, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
char *query, *ref;
int distance, first = 0, last = 0;
// LOG_DEBUG("\n\n P R E - P R O C E S S\n");
// initialize query and reference sequences to Smith-Waterman
for (size_t i = 0; i < num_targets; i++) {
read_index = mapping_batch->targets[i];
read = (fastq_read_t *) array_list_get(read_index, fq_batch);
cal_list = mapping_batch->mapping_lists[read_index];
num_cals = array_list_size(cal_list);
if (num_cals <= 0) continue;
read_len = read->length;
min_distance = read_len*0.2;
LOG_DEBUG_F(">>>>> read %s\n", read->id);
// printf(">>>>> read %s\n", read->id);
// processing each CAL from this read
for(size_t j = 0; j < num_cals; j++) {
// get cal and read index
cal = array_list_get(j, cal_list);
LOG_DEBUG_F("CAL #%i of %i (strand %i), sr_list size = %i, sr_duplicate_list size = %i\n",
j, num_cals, cal->strand, cal->sr_list->size, cal->sr_duplicate_list->size);
prev_s = NULL;
itr = linked_list_iterator_new(cal->sr_list);
s = (seed_region_t *) linked_list_iterator_curr(itr);
while (s != NULL) {
{
// for debugging
size_t start = s->genome_start;// + 1;
size_t end = s->genome_end;// + 1;
size_t len = end - start + 1;
// printf(":::::::::: %lu - %lu = %i ::::::::::::\n", end, start, len );
char *ref = (char *) malloc((len + 1) * sizeof(char));
genome_read_sequence_by_chr_index(ref, 0, cal->chromosome_id - 1,
&start, &end, genome);
ref[len] = '\0';
//
LOG_DEBUG_F("\tseed: [%i|%i - %i|%i] %s (len = %i)\n",
s->genome_start, s->read_start, s->read_end, s->genome_end, ref, len);
free(ref);
}
// set the cigar for the current region
gap_read_len = s->read_end - s->read_start + 1;
cigar_code = cigar_code_new();
cigar_code_append_op(cigar_op_new(gap_read_len, 'M'), cigar_code);
s->info = (void *) cigar_code;
cigar_code = NULL;
sw_prepare = NULL;
if ((prev_s == NULL && s->read_start != 0) || (prev_s != NULL)) {
distance = 0;
mapping_batch->num_gaps++;
if (prev_s == NULL) {
// gap at the first position
gap_read_start = 0;
gap_read_end = s->read_start - 1;
gap_genome_start = s->genome_start - s->read_start;
gap_genome_end = s->genome_start - 1;
gap_read_len = gap_read_end - gap_read_start + 1;
gap_genome_len = gap_genome_end - gap_genome_start + 1;
cal->start = gap_genome_start;
assert(gap_read_len != 0);
assert(gap_genome_len != 0);
if (gap_read_len > min_gap) {
// the gap is too big, may be there's another CAL to cover it
cigar_code = cigar_code_new();
cigar_code_append_op(cigar_op_new(gap_read_len, 'H'), cigar_code);
} else {
left_flank = 0;
right_flank = DOUBLE_FLANK;
}
} else {
assert(prev_s->read_end < s->read_start);
// gap in a middle position
gap_read_start = prev_s->read_end + 1;
gap_read_end = s->read_start - 1;
gap_genome_start = prev_s->genome_end + 1;
gap_genome_end = s->genome_start - 1;
gap_read_len = gap_read_end - gap_read_start + 1;
gap_genome_len = gap_genome_end - gap_genome_start + 1;
LOG_DEBUG_F("gap (read, genome) = (%i, %i)\n", gap_read_len, gap_genome_len);
if (gap_genome_len == 0) { printf("#@#: %s\n", read->id); }
assert(gap_genome_len != 0);
if (gap_read_len == 0) {
// there's a deletion just between two consecutives seeds
cigar_code = (cigar_code_t *)prev_s->info;
cigar_code_append_op(cigar_op_new(gap_genome_len, 'D'), cigar_code);
cigar_code->distance += gap_genome_len;
cigar_code_append_op(cigar_op_new(s->read_end - s->read_start + 1, 'M'), cigar_code);
cigar_code->distance += ((cigar_code_t *)s->info)->distance;
prev_s->read_end = s->read_end;
prev_s->genome_end = s->genome_end;
LOG_DEBUG_F("prev cigar = %s\n", new_cigar_code_string((cigar_code_t *)prev_s->info));
// continue loop...
linked_list_iterator_remove(itr);
s = linked_list_iterator_curr(itr);
continue;
}
left_flank = SINGLE_FLANK;
right_flank = SINGLE_FLANK;
}
if (!cigar_code) {
// we have to try to fill this gap and get a cigar
if (gap_read_len == gap_genome_len) {
// 1) first, for from begin -> end, and begin <- end
start = gap_genome_start;// + 1;
end = gap_genome_end;// + 1;
first = -1;
last = -1;
ref = (char *) malloc((gap_genome_len + 5) * sizeof(char));
genome_read_sequence_by_chr_index(ref, 0, cal->chromosome_id - 1,
&start, &end, genome);
// handle strand -
if (cal->strand) {
if (revcomp_seq == NULL) {
revcomp_seq = strdup(read->sequence);
seq_reverse_complementary(revcomp_seq, read_len);
}
query = &revcomp_seq[gap_read_start];
} else {
query = &read->sequence[gap_read_start];
}
for (int k = 0; k < gap_read_len; k++) {
if (query[k] != ref[k]) {
distance++;
if (first == -1) first = k;
last = k;
}
}
if (distance < min_distance) {
cigar_code = cigar_code_new();
cigar_code_append_op(cigar_op_new(gap_read_len, 'M'), cigar_code);
cigar_code_inc_distance(distance, cigar_code);
}
}
if (!cigar_code) {
// 2) second, prepare SW to run
// get query sequence, revcomp if necessary
size_t read_start = gap_read_start - left_flank;
size_t read_end = gap_read_end + right_flank;
int gap_read_len_ex = read_end - read_start + 1;
query = (char *) malloc((gap_read_len_ex + 1) * sizeof(char));
// handle strand -
if (cal->strand) {
if (revcomp_seq == NULL) {
revcomp_seq = strdup(read->sequence);
seq_reverse_complementary(revcomp_seq, read_len);
}
memcpy(query, &revcomp_seq[read_start], gap_read_len_ex);
} else {
memcpy(query, &read->sequence[read_start], gap_read_len_ex);
}
query[gap_read_len_ex] = '\0';
// get ref. sequence
size_t genome_start = gap_genome_start - left_flank;// + 1;
size_t genome_end = gap_genome_end + right_flank;// + 1;
int gap_genome_len_ex = genome_end - genome_start + 1;
ref = (char *) malloc((gap_genome_len_ex + 1) * sizeof(char));;
genome_read_sequence_by_chr_index(ref, 0, cal->chromosome_id - 1,
&genome_start, &genome_end, genome);
ref[gap_genome_len_ex] = '\0';
if (prev_s == NULL) {
sw_prepare = sw_prepare_new(query, ref, left_flank, right_flank, FIRST_SW);
} else {
sw_prepare = sw_prepare_new(query, ref, left_flank, right_flank, MIDDLE_SW);
}
array_list_insert(sw_prepare, sw_prepare_list);
// increase counter
sw_count++;
LOG_DEBUG_F("query: %s\n", query);
LOG_DEBUG_F("ref : %s\n", ref);
LOG_DEBUG_F("dist.: %i (min. %i) of %i (first = %i, last = %i)\n",
distance, min_distance, gap_read_len, first, last);
LOG_DEBUG_F("\tto SW (read %lu-%lu, genome %lu-%lu) = (%i, %i): read %s\n",
gap_read_start, gap_read_end, gap_genome_start, gap_genome_end,
gap_read_end - gap_read_start + 1, gap_genome_end - gap_genome_start + 1,
read->id);
}
}
// insert gap in the list
new_s = seed_region_new(gap_read_start, gap_read_end, gap_genome_start, gap_genome_end, 0, 0, 0);
new_s->info = (void *) cigar_code;
linked_list_iterator_insert(new_s, itr);
if (sw_prepare) {
sw_prepare->seed_region = new_s;
sw_prepare->cal = cal;
sw_prepare->read = read;
}
}
// continue loop...
prev_s = s;
linked_list_iterator_next(itr);
s = linked_list_iterator_curr(itr);
}
// check for a gap at the last position
sw_prepare = NULL;
if (prev_s != NULL && prev_s->read_end < read_len - 1) {
cigar_code = NULL;
mapping_batch->num_gaps++;
// mapping_batch->num_sws++;
// mapping_batch->num_ext_sws++;
// gap at the last position
gap_read_start = prev_s->read_end + 1;
gap_read_end = read_len - 1;
gap_read_len = gap_read_end - gap_read_start + 1;
assert(gap_read_len != 0);
gap_genome_len = gap_read_len;
gap_genome_start = prev_s->genome_end + 1;
gap_genome_end = gap_genome_start + gap_genome_len - 1;
cal->end = gap_genome_end;
assert(gap_genome_len != 0);
// LOG_DEBUG_F("\t\tgap_read_len = %i, gap_genome_len = %i\n", gap_read_len, gap_genome_len);
// LOG_DEBUG_F("\t\t%i : [%lu|%lu - %lu|%lu]\n",
// sw_count, gap_genome_start, gap_read_start, gap_read_end, gap_genome_end);
if (gap_read_len > min_gap) {
// the gap is too big, may be there's another CAL to cover it
cigar_code = cigar_code_new();
cigar_code_append_op(cigar_op_new(gap_read_len, 'H'), cigar_code);
} else {
// we have to try to fill this gap and get a cigar
// 1) first, for from begin -> end, and begin <- end
start = gap_genome_start;// + 1;
end = gap_genome_end;// + 1;
first = -1;
last = -1;
ref = (char *) malloc((gap_genome_len + 1) * sizeof(char));;
genome_read_sequence_by_chr_index(ref, 0, cal->chromosome_id - 1,
&start, &end, genome);
// handle strand -
if (cal->strand) {
if (revcomp_seq == NULL) {
revcomp_seq = strdup(read->sequence);
seq_reverse_complementary(revcomp_seq, read_len);
}
query = &revcomp_seq[gap_read_start];
} else {
query = &read->sequence[gap_read_start];
}
distance = 0;
for (int k = 0; k < gap_read_len; k++) {
if (query[k] != ref[k]) {
distance++;
if (first == -1) first = k;
last = k;
}
}
if (distance < min_distance) {
cigar_code = cigar_code_new();
cigar_code_append_op(cigar_op_new(gap_read_len, 'M'), cigar_code);
cigar_code_inc_distance(distance, cigar_code);
} else {
// 2) second, prepare SW to run
left_flank = DOUBLE_FLANK;
right_flank = 0;
// get query sequence, revcomp if necessary
size_t read_start = gap_read_start - left_flank;
size_t read_end = gap_read_end + right_flank;
int gap_read_len_ex = read_end - read_start + 1;
query = (char *) malloc((gap_read_len_ex + 1) * sizeof(char));
// handle strand -
if (cal->strand) {
if (revcomp_seq == NULL) {
revcomp_seq = strdup(read->sequence);
seq_reverse_complementary(revcomp_seq, read_len);
}
memcpy(query, &revcomp_seq[read_start], gap_read_len_ex);
} else {
memcpy(query, &read->sequence[read_start], gap_read_len_ex);
}
query[gap_read_len_ex] = '\0';
// get ref. sequence
size_t genome_start = gap_genome_start - left_flank;// + 1;
size_t genome_end = gap_genome_end + right_flank;// + 1;
int gap_genome_len_ex = genome_end - genome_start + 1;
ref = (char *) malloc((gap_genome_len_ex + 1) * sizeof(char));;
genome_read_sequence_by_chr_index(ref, 0, cal->chromosome_id - 1,
&genome_start, &genome_end, genome);
query[gap_genome_len_ex] = '\0';
sw_prepare = sw_prepare_new(query, ref, left_flank, right_flank, LAST_SW);
array_list_insert(sw_prepare, sw_prepare_list);
// increase counter
sw_count++;
LOG_DEBUG_F("query: %s\n", query);
LOG_DEBUG_F("ref : %s\n", ref);
LOG_DEBUG_F("dist.: %i (min. %i) of %i (first = %i, last = %i)\n",
distance, min_distance, gap_read_len, first, last);
LOG_DEBUG_F("\tto SW (read %lu-%lu, genome %lu-%lu) = (%i, %i): read %s\n",
gap_read_start, gap_read_end, gap_genome_start, gap_genome_end,
gap_read_end - gap_read_start + 1, gap_genome_end - gap_genome_start + 1,
read->id);
}
}
// insert gap in the list
new_s = seed_region_new(gap_read_start, gap_read_end, gap_genome_start, gap_genome_end, 0, 0, 0);
new_s->info = (void *) cigar_code;
linked_list_insert_last(new_s, cal->sr_list);
if (sw_prepare) {
sw_prepare->seed_region = new_s;
sw_prepare->cal = cal;
sw_prepare->read = read;
}
}
linked_list_iterator_free(itr);
}
// free memory
if (revcomp_seq) {
free(revcomp_seq);
revcomp_seq = NULL;
}
}
// display_sr_lists("ATER pre-process in fill_gaps", mapping_batch);
LOG_DEBUG_F("\nR U N S W (sw_count = %i, sw_prepare_list size = %i)\n", sw_count, array_list_size(sw_prepare_list));
assert(sw_count == array_list_size(sw_prepare_list));
char *q[sw_count], *r[sw_count];
for (int i = 0; i < sw_count; i++) {
sw_prepare = array_list_get(i, sw_prepare_list);
q[i] = sw_prepare->query;
r[i] = sw_prepare->ref;
}
sw_multi_output_t *output = sw_multi_output_new(sw_count);
// run Smith-Waterman
smith_waterman_mqmr(q, r, sw_count, sw_optarg, 1, output);
LOG_DEBUG("P O S T - P R O C E S S\n");
cigar_op_t* cigar_op;
for (int i = 0; i < sw_count; i++) {
sw_prepare = array_list_get(i, sw_prepare_list);
s = sw_prepare->seed_region;
int read_gap_len = s->read_end - s->read_start + 1;
int genome_gap_len = s->genome_end - s->genome_start + 1;
int read_gap_len_ex = read_gap_len_ex + sw_prepare->left_flank + sw_prepare->right_flank;
int genome_gap_len_ex = genome_gap_len_ex + sw_prepare->left_flank + sw_prepare->right_flank;
LOG_DEBUG_F("\tgap (read %lu-%lu, genome %lu-%lu) = (%i, %i): read %s\n",
s->read_start, s->read_end, s->genome_start, s->genome_end,
read_gap_len, genome_gap_len, sw_prepare->read->id);
LOG_DEBUG_F("\tflanks (left, right) = (%i, %i)\n", sw_prepare->left_flank, sw_prepare->right_flank);
LOG_DEBUG_F("\tquery : %s\n", sw_prepare->query);
LOG_DEBUG_F("\tref : %s\n", sw_prepare->ref);
LOG_DEBUG_F("\tmquery: %s (start %i)\n", output->query_map_p[i], output->query_start_p[i]);
LOG_DEBUG_F("\tmref : %s (start %i)\n", output->ref_map_p[i], output->ref_start_p[i]);
cigar_code_t *cigar_c = generate_cigar_code(output->query_map_p[i], output->ref_map_p[i],
strlen(output->query_map_p[i]), output->query_start_p[i],
output->ref_start_p[i], read_gap_len, genome_gap_len,
&distance, sw_prepare->ref_type);
LOG_DEBUG_F("\tscore : %0.2f, cigar: %s (distance = %i)\n",
output->score_p[i], new_cigar_code_string(cigar_c), distance);
/*
if (output->query_start_p[i] > 0 && output->ref_start_p[i] > 0 &&
output->query_start_p[i] != output->ref_start_p[i]) {
LOG_DEBUG("both map start points > 0 and are different lengths");
exit(-1);
}
*/
// assert(output->query_start_p[i] == 0);
// assert(output->ref_start_p[i] == 0);
cigar_op = cigar_code_get_op(0, cigar_c);
if (cigar_op) {
if (cigar_op->name == 'H') {
if (output->ref_start_p[i] == 0) {
cigar_op->name = 'I';
} else {
cigar_op->name = 'M';
}
} else if (cigar_op->name == '=') cigar_op->name = 'M';
}
cigar_op = cigar_code_get_last_op(cigar_c);
if (cigar_op && cigar_op->name == 'H') cigar_op->name = 'I';
LOG_DEBUG_F("gap_read_len = %i, cigar_code_length (%s) = %i\n",
read_gap_len, new_cigar_code_string(cigar_c), cigar_code_nt_length(cigar_c));
assert(read_gap_len == cigar_code_nt_length(cigar_c));
/*
if (cigar_code_get_num_ops(cigar_c) > 2) {
if (sw_prepare->left_flank > 0) {
cigar_op = cigar_code_get_op(0, cigar_c);
assert(cigar_op->number >= sw_prepare->left_flank && cigar_op->name == 'M');
cigar_op->number -= sw_prepare->left_flank;
}
if (sw_prepare->right_flank > 0) {
cigar_op = cigar_code_get_last_op(cigar_c);
assert(cigar_op->number >= sw_prepare->right_flank && cigar_op->name == 'M');
cigar_op->number -= sw_prepare->right_flank;
}
init_cigar_string(cigar_c);
LOG_DEBUG_F("\tnew cigar: %s\n", new_cigar_code_string(cigar_c));
} else {
assert(cigar_code_get_num_ops(cigar_c) == 1);
if (sw_prepare->right_flank > 0) {
cigar_op = cigar_code_get_last_op(cigar_c);
assert(cigar_op->number >= sw_prepare->right_flank && cigar_op->name == 'M');
cigar_op->number -= (sw_prepare->left_flank + sw_prepare->right_flank);
if (cigar_op->number > read_gap_len) {
cigar_code_append_op(cigar_op_new(cigar_op->number - read_gap_len, 'D'), cigar_c);
} else if (cigar_op->number < read_gap_len) {
cigar_code_append_op(cigar_op_new(read_gap_len - cigar_op->number, 'I'), cigar_c);
} else{
init_cigar_string(cigar_c);
}
// LOG_DEBUG_F("\tnew cigar: %s\n", new_cigar_code_string(cigar_c));
}
}
*/
// and now set the cigar for this gap
s->info = (void *) cigar_c;
// free
sw_prepare_free(sw_prepare);
}
display_sr_lists("END of fill_gaps", mapping_batch);
// free memory
sw_multi_output_free(output);
array_list_free(sw_prepare_list, (void *) NULL);
}
static int filesystem_readdir(const char *path, void *buffer, fuse_fill_dir_t fill_dir,
off_t offset, struct fuse_file_info *file_info) {
int retstat = 0;
// DIR *dp;
// struct dirent *de;
logging_log("Filesystem", LOGGING_LEVEL_INFO,
"filesystem_readdir(path=\"%s\", buf=0x%08x, filler=0x%08x, offset=%ld, fi=0x%08x)...",
path, buffer, fill_dir, offset, file_info);
if(!strcmp(path, "/")) {
pthread_mutex_lock(filesystem_io_mutex);
ArrayList *search_dirs = searcher_get_searches();
pthread_mutex_unlock(filesystem_io_mutex);
if(search_dirs != NULL) {
for(size_t i = 0; i < array_list_get_length(search_dirs); i++) {
char *name;
array_list_get(search_dirs, (const void **)&name, i);
if(fill_dir(buffer, name, NULL, 0) != 0) {
array_list_free(search_dirs);
return -ENOMEM;
}
}
array_list_free(search_dirs);
}
} else {
path_t *path_parsed = path_parse(path);
logging_log("Filesystem", LOGGING_LEVEL_INFO, "path_parsed->parts_length: %lu...",
path_parsed->parts_length);
if(path_parsed->parts_length) {
pthread_mutex_lock(filesystem_io_mutex);
ALDictionary *results = searcher_get_search_results(path_parsed->parts[0]);
if(results != NULL) {
logging_log("Filesystem", LOGGING_LEVEL_INFO, "results != NULL...");
ALDictionaryEnumerator *e = al_dictionary_get_enumerator(results);
while(al_dictionary_enumerator_move_next(e)) {
ALDictionaryKeyValuePair *pair;
al_dictionary_enumerator_get_current(e, &pair);
logging_log("Filesystem", LOGGING_LEVEL_INFO, "pair->key: %s...",
(char*)pair->key);
if(fill_dir(buffer, (char*)pair->key, NULL, 0) != 0) {
al_dictionary_enumerator_free(e);
path_free(path_parsed);
pthread_mutex_unlock(filesystem_io_mutex);
return -ENOMEM;
}
}
al_dictionary_enumerator_free(e);
} else
retstat = -1;
pthread_mutex_unlock(filesystem_io_mutex);
} else
retstat = -1;
path_free(path_parsed);
}
return retstat;
}
static int filesystem_getattr(const char *path, struct stat *statbuf) {
int retstat;
logging_log("Filesystem", LOGGING_LEVEL_INFO,
"filesystem_getattr(path=\"%s\", statbuf=0x%08x)", path, statbuf);
struct timeval tv;
gettimeofday(&tv, NULL);
struct timespec ts;
ts.tv_nsec = tv.tv_usec * 1000;
ts.tv_sec = tv.tv_sec;
statbuf->st_uid = getuid();
statbuf->st_size = 0;
statbuf->st_rdev = 0;
statbuf->st_nlink = 0;
statbuf->st_mtime = ts.tv_sec;
statbuf->st_mtimensec = ts.tv_nsec;
statbuf->st_ino = 0;
statbuf->st_gid = getgid();
statbuf->st_dev = 0;
statbuf->st_ctime = ts.tv_sec;
statbuf->st_ctimensec = ts.tv_nsec;
statbuf->st_blocks = 0;
statbuf->st_blksize = 0;
statbuf->st_atime = ts.tv_sec;
statbuf->st_atimensec = ts.tv_nsec;
path_t *path_parsed = path_parse(path);
logging_log("Filesystem", LOGGING_LEVEL_INFO, "path_parsed->parts_length: %lu...",
path_parsed->parts_length);
switch(path_parsed->parts_length) {
case 0: {
logging_log("Filesystem", LOGGING_LEVEL_INFO, "filesystem_getattr() - /...");
statbuf->st_mode = configuration->filesystem_directory_mode;
retstat = 0;
break;
}
case 1: {
if(!strcmp(path_parsed->parts[0], "search")) {
retstat = 0;
statbuf->st_mode = configuration->filesystem_directory_mode;
break;
}
// if(path_parsed->parts[0][0] == '.')
// retstat = -1;
// else {
// statbuf->st_mode = FILESYSTEM_DIRECTORY_MODE;
// retstat = 0;
// }
pthread_mutex_lock(filesystem_io_mutex);
ArrayList *search_dirs = searcher_get_searches();
pthread_mutex_unlock(filesystem_io_mutex);
if(search_dirs != NULL) {
retstat = -1;
for(size_t i = 0; i < array_list_get_length(search_dirs); i++) {
char *name;
array_list_get(search_dirs, (const void **)&name, i);
if(!strcmp(path_parsed->parts[0], name)) {
retstat = 0;
statbuf->st_mode = configuration->filesystem_directory_mode;
goto for_end;
}
}
for_end:
;
array_list_free(search_dirs);
} else
retstat = -1;
break;
}
case 2: {
if(!strcmp(path_parsed->parts[0], "search"))
if(strlen(path_parsed->parts[0]) > 3) {
ALDictionary *value = searcher_file_name_url_dictionary_get(path_parsed->parts[1]);
pthread_mutex_lock(filesystem_io_mutex);
searcher_add_search(path_parsed->parts[1], value);
pthread_mutex_unlock(filesystem_io_mutex);
retstat = 0;
statbuf->st_mode = configuration->filesystem_file_mode;
break;
}
pthread_mutex_lock(filesystem_io_mutex);
ALDictionary *results = searcher_get_search_results(path_parsed->parts[0]);
if(results != NULL) {
logging_log("Filesystem", LOGGING_LEVEL_INFO, "results != NULL...");
char result;
char *url = (char*)al_dictionary_get(results, &result, path_parsed->parts[1]);
if(result) {
logging_log("Filesystem", LOGGING_LEVEL_INFO, "Invalid file %s...", path);
retstat = -1;
} else {
logging_log("Filesystem", LOGGING_LEVEL_INFO, "Url is %s...", url);
retstat = 0;
statbuf->st_mode = configuration->filesystem_file_mode;
statbuf->st_size = downloader_file_size_try_get(url);
}
} else
retstat = -1;
pthread_mutex_unlock(filesystem_io_mutex);
break;
}
default: {
break;
}
}
path_free(path_parsed);
logging_log("Filesystem", LOGGING_LEVEL_INFO, "filesystem_getattr() finished...");
return retstat;
}
Object array_get(Array * const list, int index) {
return array_list_get((ArrayList * const ) list, index);
}
int apply_caling_rna(cal_seeker_input_t* input, batch_t *batch) {
LOG_DEBUG("========= APPLY CALING RNA =========\n");
//if (time_on) { start_timer(start); }
bwt_optarg_t *bwt_optarg = input->bwt_optarg;
bwt_index_t *bwt_index = input->index;
cal_optarg_t *cal_optarg = input->cal_optarg;
mapping_batch_t *mapping_batch = batch->mapping_batch;
size_t num_cals, select_cals;
size_t total_reads = 0;
size_t num_targets, target_pos, total_targets, extra_target_pos;
fastq_read_t *read;
genome_t *genome = input->genome;
unsigned int num_chromosomes = genome->num_chromosomes;
int min_seeds, max_seeds;
int seed_size = input->cal_optarg->seed_size;
array_list_t *cal_list, *list;
cal_t *cal;
//array_list_t *region_list;
num_targets = mapping_batch->num_targets;
total_targets = 0;
extra_target_pos = 0;
total_reads += num_targets;
target_pos = 0;
mapping_batch->extra_stage_do = 1;
/* int t, target;
for (t = 0; t < num_targets; t++) {
target = mapping_batch->targets[t];
mapping_batch->mapping_lists[target]->size = 0;
}
return RNA_POST_PAIR_STAGE;
*/
array_list_t *region_list = array_list_new(1000,
1.25f,
COLLECTION_MODE_ASYNCHRONIZED);
//extern size_t TOTAL_READS_SEEDING, TOTAL_READS_SEEDING2;
//pthread_mutex_lock(&mutex_sp);
//TOTAL_READS_SEEDING += num_targets;
//pthread_mutex_unlock(&mutex_sp);
//printf("Num targets = %i\n", num_targets);
for (size_t i = 0; i < num_targets; i++) {
read = array_list_get(mapping_batch->targets[i], mapping_batch->fq_batch);
//printf("From CAL Seeker %s\n", read->id);
list = mapping_batch->mapping_lists[mapping_batch->targets[i]];
//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;
//printf("%i\n", input->cal_optarg->min_cal_size);
num_cals = bwt_generate_cals(read->sequence,
seed_size,
bwt_optarg,
cal_optarg,
bwt_index,
list,
num_chromosomes);
// if we want to seed with 24-length seeds,
if (num_cals == 0) {
//printf("No Cals seeding...\n");
//pthread_mutex_lock(&mutex_sp);
//extern size_t seeds_1err;
//seeds_1err++;
//pthread_mutex_unlock(&mutex_sp);
int seed_size = 24;
//First, Delete old regions
array_list_clear(region_list, (void *)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,
region_list);
max_seeds = (read->length / 15)*2 + 10;
//int prev_min_cal = input->cal_optarg->min_cal_size;
//input->cal_optarg->min_cal_size = seed_size + seed_size / 2;
//printf("NO CALS, new seeds %lu\n", array_list_size(region_list));
num_cals = bwt_generate_cal_list_linked_list(region_list,
input->cal_optarg,
&min_seeds, &max_seeds,
genome->num_chromosomes + 1,
list, read->length,
cal_optarg->min_cal_size,
0);
//input->cal_optarg->min_cal_size = prev_min_cal;
//pthread_mutex_lock(&mutex_sp);
//TOTAL_READS_SEEDING2++;
//pthread_mutex_unlock(&mutex_sp);
}
array_list_clear(region_list, (void *)region_bwt_free);
//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;
size_t genome_start = 0;
int first = 1;
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);
LOG_DEBUG_F("\t\t:: read_star %lu > read_end %lu \n", s->read_start, s->read_end);
if (start > s->read_start || s->read_start >= s->read_end) {
LOG_DEBUG("\t\t\t:: remove\n");
found++;
founds[j] = 1;
}
if (!first &&
((s->genome_start < genome_start) ||
(s->genome_start - genome_start) > 2*read->length)) {
//printf("Remove (genome_start = %i s->genome_start = %i)\n", genome_start, s->genome_start);
//cal_print(cal);
found++;
founds[j] = 1;
}
first = 0;
start = s->read_end + 1;
genome_start = s->genome_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;
}
mapping_batch->mapping_lists[mapping_batch->targets[i]] = list;
num_cals = array_list_size(list);
int max = 100;
if (num_cals > max) {
select_cals = num_cals - max;
for(int j = num_cals - 1; j >= max; j--) {
cal_free(array_list_remove_at(j, mapping_batch->mapping_lists[mapping_batch->targets[i]]));
}
}
//mapping_batch->targets[target_pos++] = mapping_batch->targets[i];
//} //else if (num_cals > 0) {
mapping_batch->targets[target_pos++] = mapping_batch->targets[i];
/* printf("<<<<<===== CAL SERVER =====>>>>>\n"); */
/* for (int c = 0; c < array_list_size(mapping_batch->mapping_lists[mapping_batch->targets[i]]); c++) { */
/* cal_t *cal_aux = array_list_get(c, mapping_batch->mapping_lists[mapping_batch->targets[i]]); */
/* cal_print(cal_aux); */
/* } */
/* printf("<<<<<===== CAL SERVER END =====>>>>>\n"); */
//printf("Total CALs %i\n", num_cals);
}
mapping_batch->num_targets = target_pos;
array_list_free(region_list, NULL);
//if (time_on) { stop_timer(start, end, time); timing_add(time, CAL_SEEKER, timing); }
LOG_DEBUG("========= APPLY CALING RNA END =========\n");
// return RNA_STAGE;
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 merge_seed_regions(mapping_batch_t *mapping_batch) {
cal_t *cal;
seed_region_t *s, *s_first;
cigar_code_t *cigar_code, *cigar_code_prev;
cigar_op_t *cigar_op;
int num_ops;
int op;
array_list_t *cals_list;
fastq_read_t *fq_read;
linked_list_iterator_t itr;
register size_t num_targets = mapping_batch->num_targets;
register size_t num_cals;
register int i;
register size_t t;
for (t = 0; t < num_targets; t++) {
cals_list = mapping_batch->mapping_lists[mapping_batch->targets[t]];
num_cals = array_list_size(cals_list);
fq_read = array_list_get(mapping_batch->targets[t], mapping_batch->fq_batch);
//LOG_DEBUG_F("Read %s\n", fq_read->id);
for (i = 0; i < num_cals; i++) {
cal = array_list_get(i, cals_list);
//LOG_DEBUG_F("\tCAL %i:\n", i);
linked_list_iterator_init(cal->sr_list, &itr);
s_first = linked_list_iterator_curr(&itr);
if (s_first) {
cigar_code_prev = (cigar_code_t *)s_first->info;
s = linked_list_iterator_next(&itr);
while (s) {
//LOG_DEBUG_F("\t\tItem [%lu|%i - %i|%lu]: \n", s->genome_start, s->read_start, s->read_end, s->genome_end);
cigar_code = (cigar_code_t *)s->info;
if (cigar_code) { //TODO: delete
num_ops = array_list_size(cigar_code->ops);
for (op = 0, cigar_op = array_list_get(op, cigar_code->ops);
op < num_ops;
op++, cigar_op = array_list_get(op, cigar_code->ops)) {
cigar_code_append_op(cigar_op, cigar_code_prev);
}
cigar_code_prev->distance += cigar_code->distance;
}
s_first->read_end = s->read_end;
s_first->genome_end = s->genome_end;
linked_list_iterator_remove(&itr);
s = linked_list_iterator_curr(&itr);
}
cal->info = (void *)cigar_code_prev;
} else {
cal->info = NULL;
//LOG_DEBUG("\t\tLINKED LIST EMPTY");
}
/*LOG_DEBUG_F("\t\tItem [%lu|%i - %i|%lu]: Distance(%i) %s\n", s->genome_start, s->read_start,
s->read_end, s->genome_end, cigar_code->distance, new_cigar_code_string(cigar_code));*/
}
}
}
void test_array_list_add_and_remove ()
{
array_list *list = create_array_list_simple ();
TEST_ASSERT_NOT_NULL (list);
// try to append an item and get it back later
char *str1 = "hello";
array_list_append (list, str1);
TEST_ASSERT_EQUAL_INT (list->item_count, 1);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 0), str1);
// try to append multiple items and get one back later
char *strs1[] = {"world", "!"};
array_list_append_all (list, (void **) strs1, 2);
TEST_ASSERT_EQUAL_INT (list->item_count, 3);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 1), strs1[0]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 2), strs1[1]);
// try to add an item at a specific position
char *str2 = " ";
array_list_add (list, 1, str2);
TEST_ASSERT_EQUAL_INT (list->item_count, 4);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 0), str1);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 1), str2);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 2), strs1[0]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 3), strs1[1]);
// try to add multiple items at a specific position
char *strs2[] = {"WORLD", "?", "\n", "HELLO "};
array_list_add_all (list, 2, (void **) strs2, 4);
TEST_ASSERT_EQUAL_INT (list->item_count, 8);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 0), str1);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 1), str2);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 2), strs2[0]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 3), strs2[1]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 4), strs2[2]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 5), strs2[3]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 6), strs1[0]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 7), strs1[1]);
// try to remove an item
array_list_remove (list, 6);
TEST_ASSERT_EQUAL_INT (list->item_count, 7);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 0), str1);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 1), str2);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 2), strs2[0]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 3), strs2[1]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 4), strs2[2]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 5), strs2[3]);
TEST_ASSERT_EQUAL_STR (array_list_get (list, 6), strs1[1]);
delete_array_list (list);
}
int sa_sam_writer(void *data) {
sa_wf_batch_t *wf_batch = (sa_wf_batch_t *) data;
sa_mapping_batch_t *mapping_batch = (sa_mapping_batch_t *) wf_batch->mapping_batch;
if (mapping_batch == NULL) {
printf("bam_writer1: error, NULL mapping batch\n");
return 0;
}
/*
for (int i = 0; i < NUM_COUNTERS; i++) {
counters[i] += mapping_batch->counters[i];
}
*/
#ifdef _TIMING
for (int i = 0; i < NUM_TIMING; i++) {
func_times[i] += mapping_batch->func_times[i];
}
#endif
int num_mismatches, num_cigar_ops;
size_t flag, pnext = 0, tlen = 0;
char *cigar_string, *cigar_M_string, *rnext = "*";
fastq_read_t *read;
array_list_t *read_list = mapping_batch->fq_reads;
array_list_t *mapping_list, *mate_list;
FILE *out_file = (FILE *) wf_batch->writer_input->bam_file;
sa_genome3_t *genome = wf_batch->sa_index->genome;
size_t num_reads, num_mappings, num_mate_mappings;
num_reads = mapping_batch->num_reads;
if (mapping_batch->options->pair_mode != SINGLE_END_MODE) {
// PAIR MODE
int len;
char *sequence, *quality;
char *seq, *opt_fields;
alignment_t *alig;
for (size_t i = 0; i < num_reads; i++) {
read = (fastq_read_t *) array_list_get(i, read_list);
// seq = read->sequence;
/*
if (i % 2 == 0) {
mate_list = mapping_batch->mapping_lists[i+1];
num_mate_mappings = array_list_size(mate_list);
} else {
mate_list = mapping_list;
num_mate_mappings = num_mappings;
}
*/
mapping_list = mapping_batch->mapping_lists[i];
num_mappings = array_list_size(mapping_list);
num_total_mappings += num_mappings;
#ifdef _VERBOSE
if (num_mappings > 1) {
num_dup_reads++;
num_total_dup_reads += num_mappings;
}
#endif
if (num_mappings > 0) {
num_mapped_reads++;
if (num_mappings > 1) {
num_multihit_reads++;
}
for (size_t j = 0; j < num_mappings; j++) {
alig = (alignment_t *) array_list_get(j, mapping_list);
/*
// update alignment
alig->secondary_alignment = 0;
if (num_mate_mappings != 1) {
alig->is_mate_mapped = 0;
alig->is_paired_end_mapped = 0;
alig->mate_strand = 0;
}
*/
if (alig->optional_fields) {
opt_fields = (char *) calloc(strlen(alig->optional_fields) + 100, sizeof(char));
sprintf(opt_fields, "NH:i:%i\t%s", num_mappings, alig->optional_fields);
// sprintf(opt_fields, "NH:i:%i\t%s\tXU:i:%i", num_mappings, alig->optional_fields, mapping_batch->status[i]);
} else {
opt_fields = (char *) calloc(100, sizeof(char));
sprintf(opt_fields, "NH:i:%i", num_mappings);
// sprintf(opt_fields, "NH:i:%i\tXU:i:%i", num_mappings, mapping_batch->status[i]);
}
/*
// update alignment
alig->secondary_alignment = 0;
if (num_mate_mappings != 1) {
alig->is_mate_mapped = 0;
alig->is_paired_end_mapped = 0;
alig->mate_strand = 0;
}
*/
flag = 0;
if (alig->is_paired_end) flag += BAM_FPAIRED;
if (alig->is_paired_end_mapped) flag += BAM_FPROPER_PAIR;
if (!alig->is_seq_mapped) flag += BAM_FUNMAP;
if ((!alig->is_mate_mapped) && (alig->is_paired_end)) flag += BAM_FMUNMAP;
if (alig->mate_strand) flag += BAM_FMREVERSE;
if (alig->pair_num == 1) flag += BAM_FREAD1;
if (alig->pair_num == 2) flag += BAM_FREAD2;
if (alig->secondary_alignment) flag += BAM_FSECONDARY;
if (alig->fails_quality_check) flag += BAM_FQCFAIL;
if (alig->pc_optical_duplicate) flag += BAM_FDUP;
if (alig->seq_strand) flag += BAM_FREVERSE;
fprintf(out_file, "%s\t%lu\t%s\t%i\t%i\t%s\t%s\t%i\t%i\t%s\t%s\t%s\n",
read->id,
flag,
genome->chrom_names[alig->chromosome],
alig->position + 1,
(num_mappings > 1 ? 0 : alig->mapq), //60, //(alig->map_quality > 3 ? 0 : alig->map_quality),
alig->cigar,
(alig->chromosome == alig->mate_chromosome ? "=" : genome->chrom_names[alig->mate_chromosome]),
alig->mate_position + 1,
alig->template_length,
alig->sequence,
alig->quality,
opt_fields
);
// free memory
free(opt_fields);
alignment_free(alig);
} // end for num_mappings
} else {
num_unmapped_reads++;
opt_fields = (char *) calloc(100, sizeof(char));
sprintf(opt_fields, "XM:i:%i XU:i:%i", num_mappings, mapping_batch->status[i]);
if (read->adapter) {
len = read->length + abs(read->adapter_length);
sequence = (char *) malloc(len + 1);
quality = (char *) malloc(len + 1);
if (read->adapter_length < 0) {
strcpy(quality, read->adapter_quality);
strcat(quality, read->quality);
} else {
strcpy(quality, read->quality);
strcat(quality, read->adapter_quality);
}
if ((read->adapter_strand == 0 && read->adapter_length < 0) ||
(read->adapter_strand == 1 && read->adapter_length > 0)) {
strcpy(sequence, read->adapter);
strcat(sequence, read->sequence);
} else {
strcpy(sequence, read->sequence);
strcat(sequence, read->adapter);
}
sequence[len] = 0;
quality[len] = 0;
} else {
sequence = read->sequence;
quality = read->quality;
}
fprintf(out_file, "%s\t4\t*\t0\t0\t*\t*\t0\t0\t%s\t%s\t%s\n",
read->id,
sequence,
quality,
opt_fields
);
free(opt_fields);
if (read->adapter) {
free(sequence);
free(quality);
}
}
array_list_free(mapping_list, (void *) NULL);
}
} else {
// SINGLE MODE
int len, mapq;
char *seq;
seed_cal_t *cal;
cigar_t *cigar;
char *sequence, *revcomp, *quality;
for (size_t i = 0; i < num_reads; i++) {
read = (fastq_read_t *) array_list_get(i, read_list);
mapping_list = mapping_batch->mapping_lists[i];
num_mappings = array_list_size(mapping_list);
num_total_mappings += num_mappings;
#ifdef _VERBOSE
if (num_mappings > 1) {
num_dup_reads++;
num_total_dup_reads += num_mappings;
}
#endif
if (num_mappings > 0) {
num_mapped_reads++;
if (num_mappings > 1) {
num_multihit_reads++;
}
for (size_t j = 0; j < num_mappings; j++) {
cal = (seed_cal_t *) array_list_get(j, mapping_list);
if (read->adapter) {
// sequences and cigar
len = read->length + abs(read->adapter_length);
sequence = (char *) malloc(len + 1);
revcomp = (char *) malloc(len + 1);
quality = (char *) malloc(len + 1);
cigar = cigar_new_empty();
if (read->adapter_length < 0) {
strcpy(quality, read->adapter_quality);
strcat(quality, read->quality);
} else {
strcpy(quality, read->quality);
strcat(quality, read->adapter_quality);
}
if ( (cal->strand == 1 &&
((read->adapter_strand == 0 && read->adapter_length > 0) ||
(read->adapter_strand == 1 && read->adapter_length < 0)))
||
(cal->strand == 0 &&
((read->adapter_strand == 0 && read->adapter_length < 0) ||
(read->adapter_strand == 1 && read->adapter_length > 0))) ) {
strcpy(sequence, read->adapter);
strcat(sequence, read->sequence);
strcpy(revcomp, read->adapter_revcomp);
strcat(revcomp, read->revcomp);
cigar_append_op(abs(read->adapter_length), 'S', cigar);
cigar_concat(&cal->cigar, cigar);
} else {
strcpy(sequence, read->sequence);
strcat(sequence, read->adapter);
strcpy(revcomp, read->revcomp);
strcat(revcomp, read->adapter_revcomp);
cigar_concat(&cal->cigar, cigar);
cigar_append_op(read->adapter_length, 'S', cigar);
}
sequence[len] = 0;
revcomp[len] = 0;
quality[len] = 0;
} else {
// sequences and cigar
sequence = read->sequence;
revcomp = read->revcomp;
quality = read->quality;
cigar = &cal->cigar;
}
if (cal->strand) {
flag = 16;
seq = revcomp;
} else {
flag = 0;
seq = sequence;
}
/*
if (i == 0) {
flag += BAM_FSECONDARY;
}
*/
cigar_string = cigar_to_string(cigar);
cigar_M_string = cigar_to_M_string(&num_mismatches, &num_cigar_ops, cigar);
if (num_mappings > 1) {
cal->mapq = 0;
}
fprintf(out_file, "%s\t%i\t%s\t%i\t%i\t%s\t%s\t%lu\t%i\t%s\t%s\tNH:i:%i\tNM:i:%i\n",
read->id,
flag,
genome->chrom_names[cal->chromosome_id],
cal->start + 1,
(num_mappings == 1 ? cal->mapq : 0),
cigar_M_string,
rnext,
pnext,
tlen,
seq,
quality,
num_mappings,
num_mismatches
);
// free memory
free(cigar_M_string);
free(cigar_string);
seed_cal_free(cal);
if (read->adapter) {
free(sequence);
free(revcomp);
free(quality);
cigar_free(cigar);
}
}
} else {
num_unmapped_reads++;
if (read->adapter) {
// sequences and cigar
len = read->length + abs(read->adapter_length);
sequence = (char *) malloc(len + 1);
quality = (char *) malloc(len + 1);
if (read->adapter_length < 0) {
strcpy(quality, read->adapter_quality);
strcat(quality, read->quality);
} else {
strcpy(quality, read->quality);
strcat(quality, read->adapter_quality);
}
if ((read->adapter_strand == 0 && read->adapter_length < 0) ||
(read->adapter_strand == 1 && read->adapter_length > 0)) {
strcpy(sequence, read->adapter);
strcat(sequence, read->sequence);
} else {
strcpy(sequence, read->sequence);
strcat(sequence, read->adapter);
}
sequence[len] = 0;
quality[len] = 0;
} else {
// sequences
sequence = read->sequence;
quality = read->quality;
}
fprintf(out_file, "%s\t4\t*\t0\t0\t*\t*\t0\t0\t%s\t%s\n",
read->id,
sequence,
quality
);
if (read->adapter) {
free(sequence);
free(quality);
}
}
array_list_free(mapping_list, (void *) NULL);
} // end for num_reads
}
// free memory
sa_mapping_batch_free(mapping_batch);
if (wf_batch) sa_wf_batch_free(wf_batch);
return 0;
}
int apply_sw_bs_4nt(sw_server_input_t* input, batch_t *batch) {
mapping_batch_t *mapping_batch = batch->mapping_batch;
genome_t *genome1 = input->genome1_p;
genome_t *genome2 = input->genome2_p;
sw_optarg_t *sw_optarg = &input->sw_optarg;
{
char r[1024];
size_t start = 169312417;
size_t end = start + 99;
genome_read_sequence_by_chr_index(r, 0,
0, &start, &end, genome2);
printf("+++++++++++++ genome2 = %s \n", r);
genome_read_sequence_by_chr_index(r, 0,
0, &start, &end, genome1);
printf("+++++++++++++ genome1 = %s \n", r);
}
// fill gaps between seeds
fill_gaps_bs(mapping_batch, sw_optarg, genome2, genome1, 20, 5, 1);
merge_seed_regions_bs(mapping_batch, 1);
fill_end_gaps_bs(mapping_batch, sw_optarg, genome1, genome2, 20, 400, 1);
fill_gaps_bs(mapping_batch, sw_optarg, genome1, genome2, 20, 5, 0);
merge_seed_regions_bs(mapping_batch, 0);
fill_end_gaps_bs(mapping_batch, sw_optarg, genome2, genome1, 20, 400, 0);
// now we can create the alignments
fastq_read_t *read;
array_list_t *fq_batch = mapping_batch->fq_batch;
char *match_seq, *match_qual;
size_t read_index, read_len, match_len, match_start;
cal_t *cal;
array_list_t *cal_list = NULL;
size_t num_cals;
seed_region_t *s;
cigar_code_t *cigar_code;
cigar_op_t *first_op;
float score, norm_score, min_score = input->min_score;
alignment_t *alignment;
array_list_t *alignment_list;
char *p, *optional_fields;
int optional_fields_length, AS;
array_list_t **mapping_lists;
size_t num_targets;
size_t *targets;
for (int bs_id = 0; bs_id < 2; bs_id++) {
if (bs_id == 0) {
mapping_lists = mapping_batch->mapping_lists;
num_targets = mapping_batch->num_targets;
targets = mapping_batch->targets;
} else {
mapping_lists = mapping_batch->mapping_lists2;
num_targets = mapping_batch->num_targets2;
targets = mapping_batch->targets2;
}
for (size_t i = 0; i < num_targets; i++) {
read_index = targets[i];
read = (fastq_read_t *) array_list_get(read_index, fq_batch);
cal_list = mapping_lists[read_index];
num_cals = array_list_size(cal_list);
if (num_cals <= 0) continue;
read_len = read->length;
alignment_list = array_list_new(num_cals, 1.25f, COLLECTION_MODE_ASYNCHRONIZED);
// processing each CAL from this read
for(size_t j = 0; j < num_cals; j++) {
// get cal and read index
cal = array_list_get(j, cal_list);
if (cal->sr_list->size == 0) continue;
s = (seed_region_t *) linked_list_get_first(cal->sr_list);
cigar_code = (cigar_code_t *) s->info;
norm_score = cigar_code_get_score(read_len, cigar_code);
score = norm_score * 100; //read_len;
LOG_DEBUG_F("score = %0.2f\n", norm_score);
// filter by SW score
if (norm_score > min_score) {
// update cigar and sequence and quality strings
cigar_code_update(cigar_code);
LOG_DEBUG_F("\tcigar code = %s\n", new_cigar_code_string(cigar_code));
match_start = 0;
match_len = cigar_code_nt_length(cigar_code);
first_op = cigar_code_get_first_op(cigar_code);
match_start = (first_op && first_op->name == 'H' ? first_op->number : 0);
match_seq = (char *) malloc((match_len + 1)* sizeof(char));
memcpy(match_seq, &read->sequence[match_start], match_len);
match_seq[match_len] = 0;
match_qual = (char *) malloc((match_len + 1)* sizeof(char));
memcpy(match_qual, &read->quality[match_start], match_len);
match_qual[match_len] = 0;
// set optional fields
optional_fields_length = 100;
optional_fields = (char *) calloc(optional_fields_length, sizeof(char));
p = optional_fields;
AS = (int) norm_score * 100;
sprintf(p, "ASi");
p += 3;
memcpy(p, &AS, sizeof(int));
p += sizeof(int);
sprintf(p, "NHi");
p += 3;
memcpy(p, &num_cals, sizeof(int));
p += sizeof(int);
sprintf(p, "NMi");
p += 3;
memcpy(p, &cigar_code->distance, sizeof(int));
p += sizeof(int);
assert(read->length == cigar_code_nt_length(cigar_code));
// create an alignment and insert it into the list
alignment = alignment_new();
//read_id = malloc(read->length);
size_t header_len = strlen(read->id);
char *head_id = (char *) malloc(header_len + 1);
get_to_first_blank(read->id, header_len, head_id);
alignment_init_single_end(head_id, match_seq, match_qual,
cal->strand, cal->chromosome_id - 1, cal->start - 1,
new_cigar_code_string(cigar_code),
cigar_code_get_num_ops(cigar_code),
norm_score * 254, 1, (num_cals > 1),
optional_fields_length, optional_fields, alignment);
array_list_insert(alignment, alignment_list);
LOG_DEBUG_F("creating alignment (bs_id = %i)...\n", bs_id);
//alignment_print(alignment);
}
}
// free the cal list, and update the mapping list with the alignment list
array_list_free(cal_list, (void *) cal_free);
mapping_lists[read_index] = alignment_list;
}
}
// go to the next stage
return BS_POST_PAIR_STAGE;
}
int apply_sw_bs(sw_server_input_t* input, batch_t *batch) {
int sw_3_nucleotides = 0;
/*
sw_optarg_t *sw_optarg2 = &input->sw_optarg;
printf("Matrix Table\n\tA\tC\tG\tT\tN\nA\t%+.0f\t%+.0f\t%+.0f\t%+.0f\t%+.0f\nC\t%+.0f\t%+.0f\t%+.0f\t%+.0f\t%+.0f\nG\t%+.0f\t%+.0f\t%+.0f\t%+.0f\t%+.0f\nT\t%+.0f\t%+.0f\t%+.0f\t%+.0f\t%+.0f\nN\t%+.0f\t%+.0f\t%+.0f\t%+.0f\t%+.0f\n\n",
sw_optarg2->subst_matrix['A']['A'],
sw_optarg2->subst_matrix['C']['A'],
sw_optarg2->subst_matrix['G']['A'],
sw_optarg2->subst_matrix['T']['A'],
sw_optarg2->subst_matrix['N']['A'],
sw_optarg2->subst_matrix['A']['C'],
sw_optarg2->subst_matrix['C']['C'],
sw_optarg2->subst_matrix['G']['C'],
sw_optarg2->subst_matrix['T']['C'],
sw_optarg2->subst_matrix['N']['C'],
sw_optarg2->subst_matrix['A']['G'],
sw_optarg2->subst_matrix['C']['G'],
sw_optarg2->subst_matrix['G']['G'],
sw_optarg2->subst_matrix['T']['G'],
sw_optarg2->subst_matrix['N']['G'],
sw_optarg2->subst_matrix['A']['T'],
sw_optarg2->subst_matrix['C']['T'],
sw_optarg2->subst_matrix['G']['T'],
sw_optarg2->subst_matrix['T']['T'],
sw_optarg2->subst_matrix['N']['T'],
sw_optarg2->subst_matrix['A']['N'],
sw_optarg2->subst_matrix['C']['N'],
sw_optarg2->subst_matrix['G']['N'],
sw_optarg2->subst_matrix['T']['N'],
sw_optarg2->subst_matrix['N']['N']
);
*/
if (sw_3_nucleotides == 0) {
apply_sw_bs_4nt(input, batch);
} else {
//printf("START: apply_sw\n");
int tid = omp_get_thread_num();
mapping_batch_t *mapping_batch = batch->mapping_batch;
cal_t *cal = NULL;
array_list_t *cal_list = NULL, *mapping_list = NULL;
array_list_t *fq_batch = mapping_batch->fq_batch;
fastq_read_t *fq_read;
// added by PP for bisulfite
array_list_t *CT_fq_batch = mapping_batch->CT_fq_batch;
array_list_t *GA_fq_batch = mapping_batch->GA_fq_batch;
array_list_t *CT_rev_fq_batch = mapping_batch->CT_rev_fq_batch;
array_list_t *GA_rev_fq_batch = mapping_batch->GA_rev_fq_batch;
fastq_read_t *fq_read2;
// end added by PP for bisulfite
size_t start, end;
size_t start2, end2;
/*
genome_t *genome = input->genome_p;
*/
// added by PP for bisulfite
genome_t *genome1 = input->genome1_p;
genome_t *genome2 = input->genome2_p;
// end added by PP for bisulfite
size_t flank_length = input->flank_length;
// SIMD support for Smith-Waterman
float score, min_score = input->min_score;
sw_output_t *sw_output;
size_t read_index, num_cals;
size_t num_targets = mapping_batch->num_targets;
size_t new_num_targets = 0;
// added by PP for bisulfite
size_t num_targets2 = mapping_batch->num_targets2;
size_t new_num_targets2 = 0;
// added by PP for bisulfite
// added by PP for bisulfite
size_t sw_total1 = mapping_batch->num_to_do;
size_t sw_total2 = mapping_batch->num_to_do2;
size_t sw_total = sw_total1 + sw_total2;
// end added by PP for bisulfite
// set to zero
mapping_batch->num_to_do = 0;
// added by PP for bisulfite
mapping_batch->num_to_do2 = 0;
int g[sw_total];
// end added by PP for bisulfite
sw_optarg_t *sw_optarg = &input->sw_optarg;
sw_multi_output_t *output = sw_multi_output_new(sw_total);
char *q[sw_total], *r[sw_total];
uint8_t strands[sw_total], chromosomes[sw_total];
size_t starts[sw_total];
size_t sw_count = 0, read_indices[sw_total], sw_count2 = 0;
int read_len, ref_len, max_ref_len;
//printf("num of sw to do: %i\n", sw_total);
// initialize query and reference sequences to Smith-Waterman
for (size_t i = 0; i < num_targets; i++) {
// printf("sw_server: target #%i of %i\n", i, num_seqs);
read_index = mapping_batch->targets[i];
// to use with the three nucleotides searches
fq_read = (fastq_read_t *) array_list_get(read_index, GA_fq_batch);
fq_read2 = (fastq_read_t *) array_list_get(read_index, GA_rev_fq_batch);
//printf("read %lu = %s\n", read_index, fq_read->sequence);
//printf("read %lu = %s\n", read_index, fq_read2->sequence);
// printf("sw_server: read #%i\n", read_index);
cal_list = mapping_batch->mapping_lists[read_index];
num_cals = array_list_size(cal_list);
read_len = fq_read->length;
// max_ref_len = read_len + (read_len / 2);
//printf("sw_server: num_cals = %i cals\n", num_cals);
// processing each CAL from this read
for(size_t j = 0; j < num_cals; j++) {
// get cal and read index
cal = array_list_get(j, cal_list);
read_indices[sw_count] = read_index;
if (flank_length >= cal->start) {
start = 0;
} else {
start = cal->start - flank_length;
}
end = cal->end + flank_length;
if (end >= genome1->chr_size[cal->chromosome_id - 1]) {
end = genome1->chr_size[cal->chromosome_id - 1] - 1;
}
ref_len = end - start + 2;
// if (ref_len < max_ref_len) {
// query sequence, revcomp if necessary
q[sw_count] = (char *) calloc((read_len + 1), sizeof(char));
// to use with the three nucleotides searches
if (cal->strand == 0) {
memcpy(q[sw_count], fq_read->sequence, read_len);
//seq_reverse_complementary(q[sw_count], read_len);
} else {
memcpy(q[sw_count], fq_read2->sequence, read_len);
}
//q[sw_count] = &(fq_batch->seq[fq_batch->data_indices[index]]);
// reference sequence
//printf("\tSW: %d.[chromosome:%d]-[strand:%d]-[start:%d, end:%d]\n", j, cal->chromosome_id, cal->strand, cal->start, cal->end);
r[sw_count] = calloc(1, end - start + 2);
// to use with the three nucleotides searches
if (cal->strand == 0) {
genome_read_sequence_by_chr_index(r[sw_count], 0,
cal->chromosome_id - 1, &start, &end, genome1);
} else {
genome_read_sequence_by_chr_index(r[sw_count], 0,
cal->chromosome_id - 1, &start, &end, genome2);
/*
start2 = genome1->chr_size[cal->chromosome_id - 1] - 1 - end;
end2 = genome1->chr_size[cal->chromosome_id - 1] - 1 - start;
genome_read_sequence_by_chr_index(r[sw_count], 0,
cal->chromosome_id - 1, &start2, &end2, genome2);
*/
}
/*
genome_read_sequence_by_chr_index(r[sw_count], cal->strand,
cal->chromosome_id - 1, &start, &end, genome1);
*/
// save some stuff, we'll use them after...
strands[sw_count] = cal->strand;
chromosomes[sw_count] = cal->chromosome_id;
starts[sw_count] = start;
/*
printf("st = %lu\tend = %lu\n", cal->start, cal->end);
printf("1\nseq %s\ngen %s\nstrand %2lu chromo %lu start %lu end %lu\n",
q[sw_count], r[sw_count], cal->strand, cal->chromosome_id, start, end);
*/
// increase counter
sw_count++;
}
// free cal_list
array_list_clear(cal_list, (void *) cal_free);
// batch->mapping_lists[index] = NULL;
}
////////////////
sw_count2 = sw_count;
for (size_t i = 0; i < num_targets2; i++) {
// printf("sw_server: target #%i of %i\n", i, num_seqs);
read_index = mapping_batch->targets2[i];
// to use with the three nucleotides searches
fq_read = (fastq_read_t *) array_list_get(read_index, CT_fq_batch);
fq_read2 = (fastq_read_t *) array_list_get(read_index, CT_rev_fq_batch);
//printf("read %lu = %s\n", read_index, fq_read->sequence);
//printf("read %lu = %s\n", read_index, fq_read2->sequence);
// printf("sw_server: read #%i\n", read_index);
cal_list = mapping_batch->mapping_lists2[read_index];
num_cals = array_list_size(cal_list);
read_len = fq_read->length;
// max_ref_len = read_len + (read_len / 2);
//printf("sw_server: num_cals = %i cals\n", num_cals);
// processing each CAL from this read
for(size_t j = 0; j < num_cals; j++) {
// get cal and read index
cal = array_list_get(j, cal_list);
read_indices[sw_count] = read_index;
if (flank_length >= cal->start) {
start = 0;
} else {
start = cal->start - flank_length;
}
end = cal->end + flank_length;
if (end >= genome1->chr_size[cal->chromosome_id - 1]) {
end = genome1->chr_size[cal->chromosome_id - 1] - 1;
}
ref_len = end - start + 2;
// if (ref_len < max_ref_len) {
// query sequence, revcomp if necessary
q[sw_count] = (char *) calloc((read_len + 1), sizeof(char));
// to use with the three nucleotides searches
if (cal->strand == 0) {
memcpy(q[sw_count], fq_read->sequence, read_len);
//seq_reverse_complementary(q[sw_count], read_len);
} else {
memcpy(q[sw_count], fq_read2->sequence, read_len);
}
//q[sw_count] = &(fq_batch->seq[fq_batch->data_indices[index]]);
// reference sequence
//printf("\tSW: %d.[chromosome:%d]-[strand:%d]-[start:%d, end:%d]\n", j, cal->chromosome_id, cal->strand, cal->start, cal->end);
r[sw_count] = calloc(1, end - start + 2);
// to use with the three nucleotides searches
if (cal->strand == 0) {
genome_read_sequence_by_chr_index(r[sw_count], 0,
cal->chromosome_id - 1, &start, &end, genome2);
} else {
genome_read_sequence_by_chr_index(r[sw_count], 0,
cal->chromosome_id - 1, &start, &end, genome1);
/*
start2 = genome1->chr_size[cal->chromosome_id - 1] - 1 - end;
end2 = genome1->chr_size[cal->chromosome_id - 1] - 1 - start;
genome_read_sequence_by_chr_index(r[sw_count], 0,
cal->chromosome_id - 1, &start2, &end2, genome1);
*/
}
/*
genome_read_sequence_by_chr_index(r[sw_count], cal->strand,
cal->chromosome_id - 1, &start, &end, genome2);
*/
// save some stuff, we'll use them after...
strands[sw_count] = cal->strand;
chromosomes[sw_count] = cal->chromosome_id;
starts[sw_count] = start;
//printf("2\nseq %s\ngen %s\nstrand %2lu chromo %lu start %lu end %lu\n",
// q[sw_count], r[sw_count], cal->strand, cal->chromosome_id, start, end);
// increase counter
sw_count++;
}
// free cal_list
array_list_clear(cal_list, (void *) cal_free);
// batch->mapping_lists[index] = NULL;
}
//printf("before smith_waterman: sw_total = %i, sw_count = %i, sw_count2 = %i\n", sw_total, sw_count, sw_count2);
// run Smith-Waterman
// printf("before smith_waterman: sw_total = %i, sw_count = %i\n", sw_total, sw_count);
smith_waterman_mqmr(q, r, sw_count, sw_optarg, 1, output);
// printf("after smith_waterman\n");
for (size_t i = 0; i < sw_count; i++) {
LOG_DEBUG_F("cal: start = %lu, strand = %i\n", starts[i], strands[i]);
LOG_DEBUG_F("\tquery : %s\n", q[i]);
LOG_DEBUG_F("\tref. : %s\n", r[i]);
LOG_DEBUG_F("\tquery map: %s (start: %i)\n",
output->query_map_p[i], output->query_start_p[i]);
LOG_DEBUG_F("\tref. map : %s (start: %i)\n",
output->ref_map_p[i], output->ref_start_p[i]);
LOG_DEBUG("\n");
}
//size_t mapp = 0, mapp2 = 0;
double norm_score;
// filter alignments by min_score
for (size_t i = 0; i < sw_count2; i++) {
read_index = read_indices[i];
fq_read = (fastq_read_t *) array_list_get(read_index, GA_fq_batch);
fq_read2 = (fastq_read_t *) array_list_get(read_index, GA_rev_fq_batch);
read_len = fq_read->length;
norm_score = NORM_SCORE(output->score_p[i], read_len, input->match);
if (norm_score >= min_score) {
// valid mappings,
//insert in the list for further processing
mapping_list = mapping_batch->mapping_lists[read_index];
array_list_set_flag(0, mapping_list);
if (array_list_size(mapping_list) == 0) {
mapping_batch->targets[new_num_targets++] = read_index;
//mapp++;
}
sw_output = sw_output_new(strands[i],
chromosomes[i],
starts[i],
strlen(r[i]),
strlen(output->query_map_p[i]),
output->query_start_p[i],
output->ref_start_p[i],
output->score_p[i],
norm_score,
output->query_map_p[i],
output->ref_map_p[i]);
array_list_insert(sw_output, mapping_list);
mapping_batch->num_to_do++;
}
// free query and reference
free(q[i]);
free(r[i]);
}
mapping_batch->num_targets = new_num_targets;
for (size_t i = sw_count2; i < sw_count; i++) {
read_index = read_indices[i];
fq_read = (fastq_read_t *) array_list_get(read_index, CT_fq_batch);
fq_read2 = (fastq_read_t *) array_list_get(read_index, CT_rev_fq_batch);
read_len = fq_read->length;
norm_score = NORM_SCORE(output->score_p[i], read_len, input->match);
if (norm_score >= min_score) {
// valid mappings,
//insert in the list for further processing
mapping_list = mapping_batch->mapping_lists2[read_index];
array_list_set_flag(0, mapping_list);
if (array_list_size(mapping_list) == 0) {
mapping_batch->targets2[new_num_targets2++] = read_index;
//mapp2++;
}
sw_output = sw_output_new(strands[i],
chromosomes[i],
starts[i],
strlen(r[i]),
strlen(output->query_map_p[i]),
output->query_start_p[i],
output->ref_start_p[i],
output->score_p[i],
norm_score,
output->query_map_p[i],
output->ref_map_p[i]);
array_list_insert(sw_output, mapping_list);
mapping_batch->num_to_do2++;
}
// free query and reference
free(q[i]);
free(r[i]);
}
mapping_batch->num_targets2 = new_num_targets2;
// update counter
// thr_sw_items[tid] += sw_count;
// free
sw_multi_output_free(output);
// go to the next stage
/*
printf("3 SW1 \t%3lu\tmapp \t%3lu\tno map (discard) \t%3lu\n",
num_targets, mapp, num_targets - mapp);
printf("3 SW2 \t%3lu\tmapp \t%3lu\tno map (discard) \t%3lu\n",
num_targets2, mapp2, num_targets2 - mapp2);
*/
//printf("END: apply_sw, (%d Smith-Waterman)\n", sw_total);
}
//return CONSUMER_STAGE;
return BS_POST_PAIR_STAGE;
// printf("END: apply_sw, (%d Smith-Waterman, %d valids)\n", total, valids);
}
int sa_bam_writer(void *data) {
sa_wf_batch_t *wf_batch = (sa_wf_batch_t *) data;
sa_mapping_batch_t *mapping_batch = (sa_mapping_batch_t *) wf_batch->mapping_batch;
if (mapping_batch == NULL) {
printf("bam_writer1: error, NULL mapping batch\n");
return 0;
}
// for (int i = 0; i < NUM_COUNTERS; i++) {
// counters[i] += mapping_batch->counters[i];
// }
#ifdef _TIMING
for (int i = 0; i < NUM_TIMING; i++) {
func_times[i] += mapping_batch->func_times[i];
}
#endif
int flag, len;
char *sequence, *quality;
fastq_read_t *read;
array_list_t *read_list = mapping_batch->fq_reads;
bam1_t *bam1;
alignment_t *alig;
array_list_t *mapping_list;
bam_file_t *out_file = wf_batch->writer_input->bam_file;
sa_genome3_t *genome = wf_batch->sa_index->genome;
size_t num_reads, num_mappings, num_mate_mappings;
num_reads = mapping_batch->num_reads;
for (size_t i = 0; i < num_reads; i++) {
read = (fastq_read_t *) array_list_get(i, read_list);
mapping_list = mapping_batch->mapping_lists[i];
num_mappings = array_list_size(mapping_list);
num_total_mappings += num_mappings;
#ifdef _VERBOSE
if (num_mappings > 1) {
num_dup_reads++;
num_total_dup_reads += num_mappings;
}
#endif
if (num_mappings > 0) {
num_mapped_reads++;
if (num_mappings > 1) {
num_multihit_reads++;
}
for (size_t j = 0; j < num_mappings; j++) {
alig = (alignment_t *) array_list_get(j, mapping_list);
// update alignment
if (num_mappings > 1) {
alig->map_quality = 0;
} else {
alig->map_quality = alig->mapq;
}
bam1 = convert_to_bam(alig, 33);
bam_fwrite(bam1, out_file);
bam_destroy1(bam1);
alignment_free(alig);
}
} else {
num_unmapped_reads++;
if (read->adapter) {
// sequences and cigar
len = read->length + abs(read->adapter_length);
sequence = (char *) malloc(len + 1);
quality = (char *) malloc(len + 1);
if (read->adapter_length < 0) {
strcpy(quality, read->adapter_quality);
strcat(quality, read->quality);
} else {
strcpy(quality, read->quality);
strcat(quality, read->adapter_quality);
}
if ((read->adapter_strand == 0 && read->adapter_length < 0) ||
(read->adapter_strand == 1 && read->adapter_length > 0)) {
strcpy(sequence, read->adapter);
strcat(sequence, read->sequence);
} else {
strcpy(sequence, read->sequence);
strcat(sequence, read->adapter);
}
sequence[len] = 0;
quality[len] = 0;
} else {
// sequences
sequence = read->sequence;
quality = read->quality;
}
alig = alignment_new();
alignment_init_single_end(strdup(read->id), sequence, quality,
0, -1, -1, /*strdup(aux)*/"", 0, 0, 0, 0, 0, NULL, alig);
bam1 = convert_to_bam(alig, 33);
bam_fwrite(bam1, out_file);
// free memory
bam_destroy1(bam1);
alig->sequence = NULL;
alig->quality = NULL;
alig->cigar = NULL;
alignment_free(alig);
if (read->adapter) {
free(sequence);
free(quality);
}
}
array_list_free(mapping_list, (void *) NULL);
}
// free memory
sa_mapping_batch_free(mapping_batch);
if (wf_batch) sa_wf_batch_free(wf_batch);
return 0;
}
int main (int argc, char *argv[]) {
if(!strcmp("count-lines", argv[1])) {
fastq_file_t *file = fastq_fopen(argv[2]);
array_list_t *reads = array_list_new(2000000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
size_t nread = 1;
int count = 0;
while((nread = fastq_fread_se(reads, 100000, file)) != 0) {
count += nread;
for(int i=0; i<reads->size; i++) {
fastq_read_print(array_list_get(i, reads));
}
// printf("Size: %i, Capacity: %i\n", reads->size, reads->capacity);
array_list_clear(reads, fastq_read_free);
}
// printf("Total num reads: %i\n", reads->size);
// fastq_read_print(array_list_get(0, reads));
// fastq_read_print(array_list_get(reads->size-1, reads));
array_list_free(reads, fastq_read_free);
fastq_fclose(file);
}
if(!strcmp("count-lines-gz", argv[1])) {
fastq_gzfile_t *file = fastq_gzopen(argv[2]);
// printf("=>%i\n", file->ret);
array_list_t *reads = array_list_new(1000000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
size_t nread = 1;
int count = 0;
while((nread = fastq_gzread_se(reads, 100000, file)) != 0) {
// nread = fastq_gzread_se(reads, 1000000, file);
count += nread;
// printf("Size: %i, Capacity: %i, count = %i, nread: %i\n", reads->size, reads->capacity, count, nread);
for(int i=0; i<reads->size; i++) {
fastq_read_print(array_list_get(i, reads));
}
// fastq_read_print((fastq_read_t*)array_list_get(reads->size-1, reads));
array_list_clear(reads, fastq_read_free);
}
// printf("Total num reads: %i\n", count);
// fastq_read_print(array_list_get(0, reads));
array_list_free(reads, fastq_read_free);
fastq_gzclose(file);
}
if(!strcmp("count-bytes-gz", argv[1])) {
fastq_gzfile_t *file = fastq_gzopen(argv[2]);
// printf("=>%i\n", file->ret);
array_list_t *reads = array_list_new(1000000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
size_t nread = 1;
int count = 0;
while((nread = fastq_gzread_bytes_se(reads, 10000000, file)) != 0) {
// nread = fastq_gzread_bytes_se(reads, 100000, file);
count += reads->size;
// printf("Size: %i, Capacity: %i, count = %i, nread: %i\n", reads->size, reads->capacity, count, nread);
for(int i=0; i<reads->size; i++) {
fastq_read_print(array_list_get(i, reads));
}
// fastq_read_print(array_list_get(reads->size-1, reads));
array_list_clear(reads, fastq_read_free);
}
// printf("Total num reads: %i\n", count);
// fastq_read_print(array_list_get(0, reads));
array_list_free(reads, fastq_read_free);
fastq_gzclose(file);
}
if(!strcmp("filter", argv[1])) {
fastq_file_t *file = fastq_fopen(argv[2]);
fastq_filter_options_t *fastq_filter_options = fastq_filter_options_new(50,150, 30, 80, 2, 100);
array_list_t *reads = array_list_new(200000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
array_list_t *passed_reads ;//= array_list_new(200000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
array_list_t *failed_reads ;//= array_list_new(200000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
size_t nread = 1;
int count = 0;
while((nread = fastq_fread_se(reads, 1000000, file)) != 0) {
count += reads->size;
passed_reads = array_list_new(200000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
failed_reads = array_list_new(200000, 1.8, COLLECTION_MODE_SYNCHRONIZED);
// for(int i=0; i<reads->size; i++) {
// fastq_read_print(array_list_get(i, reads));
// }
fastq_filter(reads, passed_reads, failed_reads, fastq_filter_options);
fastq_read_print(array_list_get(0, passed_reads));
fastq_read_print(array_list_get(0, failed_reads));
printf("Total Reads: %lu, Passed Reads: %lu, Reads failed: %lu\n", reads->size, passed_reads->size, failed_reads->size);
array_list_clear(reads, fastq_read_free);
array_list_free(passed_reads, NULL);
array_list_free(failed_reads, NULL);
// fastq_read_print(array_list_get(0, passed_reads));
// fastq_read_print(array_list_get(0, failed_reads));
// printf("Total Reads: %lu, Passed Reads: %lu, Reads filter: %lu\n", reads->size, passed_reads->size, failed_reads->size);
}
// fastq_read_print(array_list_get(0, passed_reads));
// fastq_read_print(array_list_get(0, failed_reads));
// printf("Total Reads: %lu, Passed Reads: %lu, Reads filter: %lu\n", reads->size, passed_reads->size, failed_reads->size);
fastq_filter_options_free(fastq_filter_options);
array_list_free(reads, NULL);
// array_list_free(passed_reads, fastq_read_free);
// array_list_free(failed_reads, fastq_read_free);
fastq_fclose(file);
}
return 0;
}
int apply_bwt_rna(bwt_server_input_t* input, batch_t *batch) {
LOG_DEBUG("========= APPLY BWT RNA =========\n");
metaexons_t *metaexons = input->metaexons;
mapping_batch_t *mapping_batch = batch->mapping_batch;
size_t num_reads = array_list_size(mapping_batch->fq_batch);
size_t num_mappings;
array_list_t *list;
size_t *unmapped_indices = mapping_batch->targets;
size_t num_unmapped = 0;
size_t num_anchors;
extern pthread_mutex_t mutex_sp;
extern st_bwt_t st_bwt;
//pthread_mutex_lock(&mutex_sp);
//extern size_t total_reads;
//total_reads += num_reads;
//pthread_mutex_unlock(&mutex_sp);
for (int i = 0; i < num_reads; i++) {
fastq_read_t *read = array_list_get(i, mapping_batch->fq_batch);
//Rev-comp
fastq_read_revcomp(read);
//printf("BWT: %s\n", read->id);
list = mapping_batch->mapping_lists[i];
array_list_set_flag(1, list);
num_mappings = bwt_map_inexact_read(read,
input->bwt_optarg_p,
input->bwt_index_p,
list);
if (array_list_get_flag(list) != 2) { //If flag 2, the read exceded the max number of mappings
if (array_list_get_flag(list) == 1) {
if (num_mappings > 0) {
num_anchors = bwt_search_pair_anchors(list, read->length);
if (num_anchors == 0) {
array_list_set_flag(NOT_ANCHORS, list);
unmapped_indices[num_unmapped++] = i;
}
} else {
array_list_set_flag(NOT_ANCHORS, list);
unmapped_indices[num_unmapped++] = i;
}
//printf("tot anchors found %i %s\n", num_anchors, read->id);
} else if (num_mappings <= 0) {
array_list_set_flag(0, list);
//printf("Read NO Mapped %i %s\n", num_anchors, read->id);
unmapped_indices[num_unmapped++] = i;
} else {
//Read Map, Metaexon Actualization
array_list_set_flag(ALIGNMENTS_FOUND, list);
pthread_mutex_lock(&mutex_sp);
st_bwt.map_bwt++;
pthread_mutex_unlock(&mutex_sp);
for (int i = 0; i < num_mappings; i++) {
alignment_t *alignment = array_list_get(i, list);
metaexon_insert(0, alignment->chromosome,
alignment->position, alignment->position + read->length, 40,
METAEXON_NORMAL, NULL,
metaexons);
//alignment->alig_data = cigar_code_new_by_string(alignment->cigar);
}
}
} else {
array_list_set_flag(ALIGNMENTS_EXCEEDED, list);
}
if (array_list_get_flag(list) == DOUBLE_ANCHORS) {
//printf("DOUBLE ANCHORS\n");
for (int j = 0; j < array_list_size(list); j++) {
//bwt_anchor_t *bwt_anchor_prev = array_list_get(j, list);
cal_t *cal = array_list_get(j, list);
metaexon_insert(0/*cal->strand*/, cal->chromosome_id - 1,
cal->start, cal->end, 40,
METAEXON_NORMAL, NULL,
metaexons);
}
} else if (array_list_get_flag(list) == SINGLE_ANCHORS) {
for (int j = 0; j < array_list_size(list); j++) {
//bwt_anchor_t *bwt_anchor = array_list_get(j, list);
cal_t *cal = array_list_get(j, list);
metaexon_t *metaexon;
if (metaexon_search(0/*cal->strand*/, cal->chromosome_id - 1,
cal->start, cal->end, &metaexon,
metaexons)) {
metaexon_insert(0/*cal->strand*/, cal->chromosome_id - 1,
cal->start, cal->end, 40,
METAEXON_NORMAL, NULL,
metaexons);
}
}
}
}
// array_list flag: 0 -> Not BWT Anchors found
// 1 -> One BWT Anchors found
// 2 -> Pair BWT Anchors found
// 3 -> Alignments found
// 4 -> Alignments exceded
mapping_batch->num_targets = num_unmapped;
LOG_DEBUG("========= APPLY BWT RNA END =========\n");
if (batch->mapping_batch->num_targets > 0) {
return RNA_CAL_STAGE;
} else {
return RNA_STAGE;
}
}
int main(int argc, char *argv[]) {
if (argc != 4) {
printf("Error.\n");
printf("Usage: %s index-dirname seq num-errors\n", argv[0]);
exit(-1);
}
char *index_dirname = argv[1];
char *seq = argv[2];
int num_errors = atoi(argv[3]);
// initializations
initReplaceTable();
bwt_optarg_t *bwt_optarg = bwt_optarg_new(num_errors, 1, 10000, 1, 0, 0);
bwt_index_t *bwt_index = bwt_index_new(index_dirname);
// seq
{
array_list_t *mapping_list = array_list_new(100000, 1.25f,
COLLECTION_MODE_SYNCHRONIZED);
size_t num_mappings;
num_mappings = bwt_map_seq(seq, bwt_optarg,
bwt_index, mapping_list);
printf("seq: %s\n", seq);
printf("num_mappings = %lu\n", num_mappings);
for (size_t i = 0; i < num_mappings; i++) {
printf("%lu\t---------------------\n", i);
alignment_print(array_list_get(i, mapping_list));
}
}
// seed
{
array_list_t *mapping_list = array_list_new(100000, 1.25f,
COLLECTION_MODE_SYNCHRONIZED);
size_t num_mappings;
size_t len = strlen(seq);
char *code_seq = (char *) calloc(len + 10, sizeof(char));
replaceBases(seq, code_seq, len);
num_mappings = bwt_map_exact_seeds_seq(code_seq, 18, 16,
bwt_optarg, bwt_index, mapping_list);
region_t *region;
for (size_t i = 0; i < num_mappings; i++) {
region = array_list_get(i, mapping_list);
printf("Region: chr = %lu, strand = %d, start = %lu, end = %lu\n",
region->chromosome_id, region->strand, region->start, region->end);
}
}
printf("Done.\n");
}
void apply_sw(sw_server_input_t* input, aligner_batch_t *batch) {
// printf("START: apply_sw\n");
int tid = omp_get_thread_num();
cal_t *cal = NULL;
array_list_t *cal_list = NULL, *mapping_list = NULL;//, *old_list = NULL, *new_list = NULL;
fastq_batch_t *fq_batch = batch->fq_batch;
size_t start, end;
genome_t *genome = input->genome_p;
size_t flank_length = input->flank_length;
// SIMD support for Smith-Waterman
float score, min_score = input->min_score;
// size_t curr_depth = 0;
sw_output_t *sw_output;
// sw_simd_input_t *sw_sinput = sw_simd_input_new(SIMD_DEPTH);
// sw_simd_output_t *sw_soutput = sw_simd_output_new(SIMD_DEPTH);
//sw_simd_context_t *context = sw_simd_context_new(input->match, input->mismatch,
// input->gap_open, input->gap_extend);
// for tracking the current read, cal being processed using sw_channel_t
//sw_channel_t *channel;
//sw_channel_t sw_channels[SIMD_DEPTH];
//memset(sw_channels, 0, sizeof(sw_channels));
//size_t header_len, read_len;
//size_t strands[SIMD_DEPTH], chromosomes[SIMD_DEPTH], starts[SIMD_DEPTH];
size_t index, num_cals;
size_t total = 0, valids = 0;
size_t num_seqs = batch->num_targets;
// set to zero
batch->num_done = batch->num_to_do;
batch->num_to_do = 0;
size_t sw_total = batch->num_done;
/*
// for all seqs pending to process !!
size_t sw_total = 0;
for (size_t i = 0; i < num_seqs; i++) {
sw_total += array_list_size(batch->mapping_lists[batch->targets[i]]);
}
printf("number of sw to run: %d (vs num_done = %d)\n", sw_total, batch->num_done);
*/
sw_optarg_t *sw_optarg = &input->sw_optarg;
/*
sw_optarg_t sw_optarg; //= sw_optarg_new(gap_open, gap_extend, matrix_filename);
sw_optarg.gap_open = input->gap_open;
sw_optarg.gap_extend = input->gap_extend;
sw_optarg.subst_matrix['A']['A'] = input->match; sw_optarg.subst_matrix['C']['A'] = input->mismatch; sw_optarg.subst_matrix['T']['A'] = input->mismatch; sw_optarg.subst_matrix['G']['A'] = input->mismatch;
sw_optarg.subst_matrix['A']['C'] = input->mismatch; sw_optarg.subst_matrix['C']['C'] = input->match; sw_optarg.subst_matrix['T']['C'] = input->mismatch; sw_optarg.subst_matrix['G']['C'] = input->mismatch;
sw_optarg.subst_matrix['A']['G'] = input->mismatch; sw_optarg.subst_matrix['C']['T'] = input->mismatch; sw_optarg.subst_matrix['T']['T'] = input->match; sw_optarg.subst_matrix['G']['T'] = input->mismatch;
sw_optarg.subst_matrix['A']['T'] = input->mismatch; sw_optarg.subst_matrix['C']['G'] = input->mismatch; sw_optarg.subst_matrix['T']['G'] = input->mismatch; sw_optarg.subst_matrix['G']['G'] = input->match;
*/
sw_multi_output_t *output = sw_multi_output_new(sw_total);
char *q[sw_total], *r[sw_total];
uint8_t strands[sw_total], chromosomes[sw_total];
size_t starts[sw_total];
size_t sw_count = 0, read_indices[sw_total];
int read_len;
// debugging: to kown how many reads are not mapped by SW score
// int unmapped_by_score[fq_batch->num_reads];
// memset(unmapped_by_score, 0, fq_batch->num_reads * sizeof(int));
// printf("num of sw to do: %i\n", sw_total);
// initialize query and reference sequences to Smith-Waterman
for (size_t i = 0; i < num_seqs; i++) {
index = batch->targets[i];
cal_list = batch->mapping_lists[index];
num_cals = array_list_size(cal_list);
// printf("sw_server: read #%i with %i cals\n", index, num_cals);
// processing each CAL from this read
for(size_t j = 0; j < num_cals; j++) {
// get cal and read index
cal = array_list_get(j, cal_list);
read_indices[sw_count] = index;
// query sequence, revcomp if necessary
read_len = fq_batch->data_indices[index + 1] - fq_batch->data_indices[index];
q[sw_count] = (char *) calloc((read_len + 1), sizeof(char));
memcpy(q[sw_count], &(fq_batch->seq[fq_batch->data_indices[index]]), read_len);
if (cal->strand == 1) {
seq_reverse_complementary(q[sw_count], read_len);
}
//q[sw_count] = &(fq_batch->seq[fq_batch->data_indices[index]]);
// reference sequence
//printf("\tSW: %d.[chromosome:%d]-[strand:%d]-[start:%d, end:%d]\n", j, cal->chromosome_id, cal->strand, cal->start, cal->end);
start = cal->start - flank_length;
end = cal->end + flank_length;
r[sw_count] = calloc(1, end - start + 2);
genome_read_sequence_by_chr_index(r[sw_count], cal->strand,
cal->chromosome_id - 1, &start, &end, genome);
// save some stuff, we'll use them after...
strands[sw_count] = cal->strand;
chromosomes[sw_count] = cal->chromosome_id;
starts[sw_count] = start;
// printf("read #%i (sw #%i): query: %s (%i)\nref : %s (%i)\n\n", index, sw_count, q[sw_count], strlen(q[sw_count]), r[sw_count], strlen(r[sw_count]));
// increase counter
sw_count++;
}
// free cal_list
array_list_free(cal_list, (void *)cal_free);
batch->mapping_lists[index] = NULL;
}
// run Smith-Waterman
// printf("before smith_waterman: number of sw = %i\n", sw_total);
smith_waterman_mqmr(q, r, sw_total, sw_optarg, 1, output);
// printf("after smith_waterman\n");
/*
// debugging
{
FILE *fd = fopen("sw.out", "w");
sw_multi_output_save(sw_total, output, fd);
fclose(fd);
}
*/
size_t num_targets = 0;
// filter alignments by min_score
for (size_t i = 0; i < sw_total; i++) {
// score = output->score_p[i] / (strlen(output->query_map_p[i]) * input->match);
// if (score >= min_score) {
/*
printf("--------------------------------------------------------------\n");
printf("Smith-Waterman results:\n");
printf("id\t%s\n", &(batch->fq_batch->header[batch->fq_batch->header_indices[read_indices[i]]]));
printf("ref\n%s\n", r[i]);
printf("query\n%s\n", q[i]);
printf("map\n%s\n", output->ref_map_p[i]);
printf("ref: chr = %d, strand = %d, start = %d, len = %d\n", chromosomes[i], strands[i], starts[i], strlen(r[i]));
printf("query-map-start = %d, ref-map-start = %d\n",
output->query_start_p[i], output->ref_start_p[i]);
printf("score = %0.2f (min. score = %0.2f)\n", output->score_p[i], min_score);
printf("--------------------------------------------------------------\n");
*/
if (output->score_p[i] >= min_score) {
// valid mappings,
//insert in the list for further processing
index = read_indices[i];
if (batch->mapping_lists[index] == NULL) {
mapping_list = array_list_new(1000,
1.25f,
COLLECTION_MODE_ASYNCHRONIZED);
array_list_set_flag(0, mapping_list);
batch->mapping_lists[index] = mapping_list;
batch->targets[num_targets++] = index;
}
sw_output = sw_output_new(strands[i],
chromosomes[i],
starts[i],
strlen(r[i]),
strlen(output->query_map_p[i]),
output->query_start_p[i],
output->ref_start_p[i],
output->score_p[i],
score,
output->query_map_p[i],
output->ref_map_p[i]);
array_list_insert(sw_output, mapping_list);
batch->num_to_do++;
// debugging
//unmapped_by_score[index] = 1;
}
// free query and reference
free(q[i]);
free(r[i]);
}
batch->num_targets = num_targets;
/*
// debugging
for (size_t i = 0; i < fq_batch->num_reads; i++) {
if (unmapped_by_score[i] == 0) {
unmapped_by_score_counter[tid]++;
//printf("by score: %s\n", &(batch->fq_batch->header[batch->fq_batch->header_indices[index]]));
}
}
*/
// update counter
thr_sw_items[tid] += sw_count;
// free
sw_multi_output_free(output);
// printf("END: apply_sw, (%d Smith-Waterman, %d valids)\n", total, valids);
}