/***********************************************************************
* Turn a given motif into its own reverse complement.
***********************************************************************/
void reverse_complement_motif
(MOTIF_T* a_motif)
{
ALPH_SIZE_T size;
int i, temp_trim;
ARRAY_T* left_freqs;
ARRAY_T* right_freqs;
ARRAY_T* temp_freqs; // Temporary space during swap.
assert(a_motif->alph == DNA_ALPH);
// Allocate space.
size = (a_motif->flags & MOTIF_HAS_AMBIGS ? ALL_SIZE : ALPH_SIZE);
temp_freqs = allocate_array(alph_size(a_motif->alph, size));
// Consider each row (position) in the motif.
for (i = 0; i < (int)((a_motif->length + 1) / 2); i++) {
left_freqs = get_matrix_row(i, a_motif->freqs);
right_freqs = get_matrix_row(a_motif->length - (i + 1), a_motif->freqs);
// Make a temporary copy of one row.
copy_array(left_freqs, temp_freqs);
// Compute reverse complements in both directions.
complement_dna_freqs(right_freqs, left_freqs);
complement_dna_freqs(temp_freqs, right_freqs);
}
free_array(temp_freqs);
if (a_motif->scores) {
// Allocate space.
temp_freqs = allocate_array(alph_size(a_motif->alph, ALPH_SIZE));
// Consider each row (position) in the motif.
for (i = 0; i < (int)((a_motif->length + 1) / 2); i++) {
left_freqs = get_matrix_row(i, a_motif->scores);
right_freqs = get_matrix_row(a_motif->length - (i + 1), a_motif->scores);
// Make a temporary copy of one row.
copy_array(left_freqs, temp_freqs);
// Compute reverse complements in both directions.
complement_dna_freqs(right_freqs, left_freqs);
complement_dna_freqs(temp_freqs, right_freqs);
}
free_array(temp_freqs);
}
//swap the trimming variables
temp_trim = a_motif->trim_left;
a_motif->trim_left = a_motif->trim_right;
a_motif->trim_right = temp_trim;
//swap the strand indicator
//this assumes a ? is equalivant to +
if (get_motif_strand(a_motif) == '-') {
set_motif_strand('+', a_motif);
} else {
set_motif_strand('-', a_motif);
}
}
void dxml_start_motif(void *ctx, char *id, char *seq, int length,
long num_sites, long p_hits, long n_hits,
double pvalue, double evalue, double uevalue) {
CTX_T *data;
MOTIF_T *motif;
data = (CTX_T*)ctx;
data->motif = (MOTIF_T*)mm_malloc(sizeof(MOTIF_T));
motif = data->motif;
memset(motif, 0, sizeof(MOTIF_T));
set_motif_id(seq, strlen(seq), motif);
set_motif_id2("", 0, motif);
set_motif_strand('+', motif);
motif->length = length;
motif->num_sites = num_sites;
motif->evalue = evalue;
// both DNA and RNA have 4 letters
motif->alph = data->fscope.alphabet;
motif->flags = MOTIF_BOTH_STRANDS; // DREME does not support the concept of single strand scanning (yet)
// allocate the matrix
motif->freqs = allocate_matrix(motif->length, alph_size(motif->alph, ALPH_SIZE));
motif->scores = NULL; // no scores in DREME xml
// no url in DREME
motif->url = strdup("");
// set by postprocessing
motif->complexity = -1;
motif->trim_left = 0;
motif->trim_right = 0;
}
/*****************************************************************************
* MEME > motifs > motif
* Construct the skeleton of a motif.
****************************************************************************/
void mxml_start_motif(void *ctx, char *id, char *name, char *alt, int width, double sites,
double llr, double ic, double re, double bayes_threshold,
double log10_evalue, double elapsed_time, char *url) {
CTX_T *data;
MOTIF_T *motif;
data = (CTX_T*)ctx;
data->mscope.motif = mm_malloc(sizeof(MOTIF_T));
motif = data->mscope.motif;
memset(motif, 0, sizeof(MOTIF_T));
set_motif_id(name, strlen(name), motif);
set_motif_id2(alt, sizeof(alt), motif);
set_motif_strand('+', motif);
motif->length = width;
motif->num_sites = sites;
motif->url = strdup(url);
motif->log_evalue = log10_evalue;
motif->evalue = pow(10.0, log10_evalue);
// calculate alphabet size
motif->alph = alph_hold(data->alph);
motif->flags = (data->fscope.strands == 2 ? MOTIF_BOTH_STRANDS : 0);
// allocate matricies
motif->freqs = allocate_matrix(motif->length, alph_size_core(motif->alph));
init_matrix(-1, motif->freqs);
motif->scores = allocate_matrix(motif->length, alph_size_core(motif->alph));
init_matrix(NO_SCORE, motif->scores);
// should be set by a post processing method
motif->complexity = -1;
motif->trim_left = 0;
motif->trim_right = 0;
// cache motif position
if (data->options & SCANNED_SITES) {
rbtree_put(data->motif_lookup, id, &(data->current_motif));
}
}
/***********************************************************************
* Create two copies of each motif. The new IDs are preceded by "+"
* and "-", and the "-" version is the reverse complement of the "+"
* version.
*
* The reverse complement is always listed directly after the original.
***********************************************************************/
void add_reverse_complements
(ARRAYLST_T* motifs)
{
int i, count;
char motif_id[MAX_MOTIF_ID_LENGTH + 1]; // Name of the current motif;
count = arraylst_size(motifs);
//double the array size
arraylst_preallocate(count*2, motifs);
arraylst_add_n(count, NULL, motifs);
//move and reverse complement the original motifs
for (i = count-1; i >= 0; --i) {
//move the motif to its new position
arraylst_swap(i, 2*i, motifs);
//get the motif and the one that will become the reverse complement
MOTIF_T *orig = arraylst_get(2*i, motifs);
if (get_motif_strand(orig) == '?') set_motif_strand('+', orig);
//copy and reverse complement the motif to the position after
MOTIF_T *rc = dup_rc_motif(orig);
arraylst_set(2*i + 1, rc, motifs);
}
assert(arraylst_size(motifs) == (2 * count));
}
