SAMPLE *get_sample_by_name(
char *sample_name
)
{
HASH_TABLE_ENTRY * hash_entry = hash_lookup_str(sample_name, ht_seq_names);
return(hash_entry != NULL ? (SAMPLE *) hash_get_entry_value(hash_entry) : NULL);
} // get_sample_by_name
/**
* Creates a motif for a given mod using a simple frequency matrix.
*/
void create_simple_motif(SUMMARY_T* summary,
MOMO_OPTIONS_T* options,
MOD_INFO_T * mod_info) {
int i;
int j;
const char* alph_letters = summary->alph_letters;
// Create the frequency matrix
MATRIX_T* freqs = NULL;
freqs = get_count_matrix(freqs, mod_info->seq_list, NULL, options, summary);
normalize_rows(0.0, freqs);
// Create the motif
MOTIF_INFO_T* motifinfo = mm_malloc(sizeof(MOTIF_INFO_T));
motifinfo->motif = allocate_motif(mod_info->mod_name, "", summary->alph, freqs, NULL);
motifinfo->seqs = arraylst_create();
for (i = 0; i < arraylst_size(mod_info->seq_list); ++i) {
SEQ_T* seqobject = options->eliminate_repeats ? hash_get_entry_value(arraylst_get(i, mod_info->seq_list)) : arraylst_get(i, mod_info->seq_list);
arraylst_add(get_raw_sequence(seqobject), motifinfo->seqs);
}
motifinfo->fg_size = arraylst_size(mod_info->seq_list);
arraylst_add(motifinfo, mod_info->motifinfos);
// clean up
free_matrix(freqs);
}
/**
* Remove sequences do not match a pattern from phospho and bg lists and update their respective count matrix
*/
void remove_sequences_and_update_matrix(char letter,
int pos,
ARRAYLST_T* seqs,
MOTIFX_STATUS_T** status_array,
int* num_active,
MATRIX_T* count,
SUMMARY_T* summary,
MOMO_OPTIONS_T* options) {
int i;
const char* alph_letters = summary->alph_letters;
// Look through phospho_seqs and remove sequences. Update phospho_seqs
for (i = 0; i < arraylst_size(seqs); ++i) {
char* curr_seq = get_raw_sequence((SEQ_T*) (options->eliminate_repeats ? hash_get_entry_value((HASH_TABLE_ENTRY*) arraylst_get(i, seqs)) : arraylst_get(i, seqs)));
// For anything active that does not match the pattern, turn it inactive.
MOTIFX_STATUS_T status = (*status_array)[i];
if (status == ACTIVE && curr_seq[pos] != letter) {
*num_active = *num_active - 1;
(*status_array)[i] = INACTIVE;
}
}
count = get_count_matrix(count, seqs, status_array, options, summary);
}
/**
* Recursive function. Creates and stores a motif using the motif-x
* algorithm until no more are left.
*/
void create_motifx_motif(ARRAYLST_T* phospho_seqs,
ARRAYLST_T* bg_seqs,
MOTIFX_STATUS_T** phospho_status,
MOTIFX_STATUS_T** bg_status,
MATRIX_T* phospho_count,
MATRIX_T* bg_count,
int* num_active,
int* num_bg_active,
char* modname,
MOD_INFO_T* mod_info,
MOMO_OPTIONS_T* options,
SUMMARY_T* summary) {
int i;
int j;
const char* alph_letters = summary->alph_letters;
// Initialize pattern, sequence count, bg sequence count, and overall score for this motif.
char* pattern = mm_malloc(options->width + 1);
for (i = 0; i < options->width; ++i) {
pattern[i] = 'X';
}
pattern[options->width] = '\0';
int* num_active_copy = mm_malloc(sizeof(int));
*num_active_copy = *num_active;
int* num_bg_active_copy = mm_malloc(sizeof(int));
*num_bg_active_copy = *num_bg_active;
double* motif_score = mm_malloc(sizeof(double));
*motif_score = 0;
// Set the pattern, num active copy, num bg active copy, motif score, and get a count of the sequences
MATRIX_T* result_count_matrix = add_to_pattern(pattern, phospho_seqs, bg_seqs, phospho_status, bg_status, num_active_copy, num_bg_active_copy, phospho_count, bg_count, motif_score, summary, options);
// If any of the characters are not X, then we have found a pattern
BOOLEAN_T found_pattern = FALSE;
for (i = 0; i < options->width; ++i) {
if (pattern[i] != 'X') {
found_pattern = TRUE;
}
}
// If there is a pattern, store the pattern and call create_motifx_motif again.
if (found_pattern) {
// fill out the rest of the pattern (e.g. if you have pattern ..ASAAA, and realize the actual pattern is A.ASAAA
for (i = 0; i < options->width; i++) {
for (j = 0; j < strlen(alph_letters); j++) {
if ((int) get_matrix_cell_defcheck(i, j, result_count_matrix) == *num_active_copy) {
pattern[i] = alph_letters[j];
}
}
}
// create the pattern name
char* pattern_name = mm_malloc(strlen(pattern) + strlen(modname) + 3);
pattern_name[0] = '\0';
strncat(pattern_name, pattern, strlen(pattern)/2);
strncat(pattern_name, "_", 1);
strncat(pattern_name, modname, strlen(modname));
strncat(pattern_name, "_", 1);
strncat(pattern_name, pattern + strlen(pattern)/2 + 1, strlen(pattern)/2);
// convert this count matrix into frequencies
normalize_rows(0.0, result_count_matrix);
// Store this motif
MOTIF_INFO_T* motifinfo = mm_malloc(sizeof(MOTIF_INFO_T));
MOTIF_T* motif = allocate_motif(pattern_name, "", summary->alph, result_count_matrix, NULL);
set_motif_nsites(motif, *num_active_copy);
motifinfo->motif = motif;
motifinfo->seqs = arraylst_create();
motifinfo->score = *motif_score;
motifinfo->fg_match = *num_active_copy;
motifinfo->fg_size = *num_active;
motifinfo->bg_match = *num_bg_active_copy;
motifinfo->bg_size = *num_bg_active;
for (i = 0; i < arraylst_size(phospho_seqs); ++i) {
MOTIFX_STATUS_T status = (*phospho_status)[i];
if (status == ACTIVE) {
SEQ_T* active_sequence = (options->eliminate_repeats) ? hash_get_entry_value(arraylst_get(i, phospho_seqs)) : arraylst_get(i, phospho_seqs);
arraylst_add(get_raw_sequence(active_sequence), motifinfo->seqs);
}
}
arraylst_add(motifinfo, mod_info->motifinfos);
// delete the sequences from this motif. turn inactive into active.
delete_sequences(phospho_status, arraylst_size(phospho_seqs));
delete_sequences(bg_status, arraylst_size(bg_seqs));
// update the count of number of actives
*num_active = *num_active - *num_active_copy;
*num_bg_active = *num_bg_active - *num_bg_active_copy;
// recalculate phospho count and bg count.
phospho_count = get_count_matrix(phospho_count, phospho_seqs, phospho_status, options, summary);
bg_count = get_count_matrix(bg_count, bg_seqs, bg_status, options, summary);
// free up space
myfree(pattern);
myfree(num_active_copy);
myfree(num_bg_active_copy);
myfree(motif_score);
myfree(pattern_name);
// try to create another motif.
create_motifx_motif(phospho_seqs,
bg_seqs,
phospho_status,
bg_status,
phospho_count,
bg_count,
num_active,
num_bg_active,
modname,
mod_info,
options,
summary);
}
// free up space
myfree(pattern);
myfree(num_active_copy);
myfree(num_bg_active_copy);
myfree(motif_score);
}
