int inexact_pad_match(char *seq,
int seq_len,
char *string,
int string_len,
int mis_match,
int *match,
int *score,
int max_matches)
{
char *pos;
char *uppert;
int i;
int n_matches;
int n_mis;
/* remove any pads from the pattern search */
depad_seq(string, &string_len, NULL);
/* uppercase search string */
if (NULL == (uppert = (char *)xmalloc(string_len + 1)))
return -2;
uppert[string_len] = 0;
for (i = string_len-1; i >= 0; i--) {
uppert[i] = toupper(string[i]);
}
for (i = 0; i < seq_len; i++) {
seq[i] = toupper(seq[i]);
}
pos = NULL;
n_matches = 0;
pos = pstrnstr_inexact(seq,seq_len, uppert,string_len, mis_match, &n_mis);
while (pos) {
if (n_matches < max_matches) {
match[n_matches] = pos - seq;
score[n_matches] = string_len - n_mis;
n_matches++;
} else {
/* make positions start at 1 */
for (i=0; i < max_matches; i++) {
match[i]++;
}
return -1; /* out of match storage */
}
pos++;
pos = pstrnstr_inexact(pos, seq_len - (pos-seq),
uppert, string_len, mis_match, &n_mis);
}
/* make positions start at 1 */
for (i=0; i < n_matches; i++) {
match[i]++;
}
xfree(uppert);
return n_matches;
}
void dust(int len, char *s)
{
int i, j, l, from, to, a, b, v;
char *depadded = (char *)malloc(len);
int *depad_to_pad = (int *)calloc(len, sizeof(int));
int depadded_len;
if (!depadded || !depad_to_pad)
return;
memcpy(depadded, s, len);
depadded_len = len;
depad_seq(depadded, &depadded_len, depad_to_pad);
from = 0;
to = -1;
for (i=0; i < depadded_len; i += window2) {
from -= window2;
to -= window2;
l = (depadded_len > i+window) ? window : depadded_len-i;
v = wo(l, depadded+i, &a, &b);
for (j = from; j <= to; j++) {
if (isalpha(s[depad_to_pad[i+j]]))
s[depad_to_pad[i+j]] = '#';
}
if (v > level) {
for (j = a; j <= b && j < window2; j++) {
if (isalpha(s[depad_to_pad[i+j]]))
s[depad_to_pad[i+j]] = '#';
}
from = j;
to = b;
} else {
from = 0;
to = -1;
}
}
free(depadded);
free(depad_to_pad);
}
/*
* Main picking function.
* Picks primers from the right end of contig1 and the left end of contig2
* suitable for a PCR reaction.
* After calling the returned value is a an array of structures linking into
* the primer3 primer_pair structure along with gap4 sanitised copies holding
* padded position and length in each contig and the depadded sequence.
* The number of elements in this array can be fetched from pstate->npairs.
*
* Returns g4_primer_pair array pointer for success,
* NULL for failure
*/
static g4_primer_pair *pick_pcr_primers2(finish_t *fin, primlib_state *pstate,
int contig1, int contig2)
{
char *cons1 = NULL, *cons2 = NULL, *cons_joined = NULL;
int pos1l, pos1r, pos2l, pos2r;
int len1, len2;
int *depad1 = NULL, *depad2 = NULL;
char *upcons1 = NULL, *upcons2 = NULL;
g4_primer_pair *pp = NULL;
int i, j;
/* Compute contig ranges */
pos1l = MAX(1, io_clength(fin->io, contig1) - (fin->opts.pcr_offset1-1));
pos1r = MAX(1, io_clength(fin->io, contig1) - (fin->opts.pcr_offset2-1));
len1 = pos1r - pos1l + 1;
if (len1 < 25)
return NULL;
pos2l = MIN(io_clength(fin->io, contig2), fin->opts.pcr_offset2);
pos2r = MIN(io_clength(fin->io, contig2), fin->opts.pcr_offset1);
len2 = pos2r - pos2l + 1;
if (len2 < 25)
return NULL;
/* Get the depadded consensus */
cons1 = (char *)xmalloc(len1+1);
cons2 = (char *)xmalloc(len2+1);
if (!cons1 || !cons2)
goto error;
calc_consensus(contig1, pos1l, pos1r, CON_SUM, cons1, NULL,
NULL, NULL, consensus_cutoff, quality_cutoff,
database_info, (void *)fin->io);
calc_consensus(contig2, pos2l, pos2r, CON_SUM, cons2, NULL,
NULL, NULL, consensus_cutoff, quality_cutoff,
database_info, (void *)fin->io);
cons1[pos1r-pos1l+1] = 0;
cons2[pos2r-pos2l+1] = 0;
upcons1 = strdup(cons1);
upcons2 = strdup(cons2);
if (!(depad1 = (int *)xmalloc((len1+1)*sizeof(int))))
goto error;
if (!(depad2 = (int *)xmalloc((len2+1)*sizeof(int))))
goto error;
depad_seq(cons1, &len1, depad1);
depad_seq(cons2, &len2, depad2);
/* Filter low complexity data from the consensus */
finish_filter(fin, cons1, len1);
finish_filter(fin, cons2, len2);
/*
* For primer3 we join our two sequences together thus, with 20 Ns:
*
* <CONS1>NNNNNNNNNNNNNNNNNNNN<CONS2>
* ^ ^
* x y
*
* Points x and y define the target ("TARGET=x,y-x" in the normal
* boulder-io input file).
* We also need to redefine the product size range to be 20 to
* 20 + 2*(len2-len1) allowing for full flexibility of primer
* positioning within the two consensus fragments.
*
* PCR primers will then be chosen using one primer within <CONS1>
* and the other within <CONS2>.
*/
if (NULL == (cons_joined = (char *)xmalloc(2*(len1 + len2 +2)+20)))
goto error;
sprintf(cons_joined, "%sNNNNNNNNNNNNNNNNNNNN%s", cons1, cons2);
{
size_t l = strlen(cons_joined);
for (i = 0; i < l; i++)
if (cons_joined[i] != 'A' &&
cons_joined[i] != 'C' &&
cons_joined[i] != 'G' &&
cons_joined[i] != 'T')
cons_joined[i] = 'N';
}
puts(cons_joined);
printf("target = %"PRId64",%d\n", (uint64_t)strlen(cons1)+1, 20);
/* Tweak arguments */
pstate->p3args.primer_task = pick_pcr_primers;
pstate->p3args.num_return = 20;
/* Pick the primer pairs */
if (-1 == primlib_choose_pcr(pstate, cons_joined, strlen(cons1)+1, 20))
goto error;
if (!(pp = (g4_primer_pair *)xmalloc(pstate->npairs * sizeof(*pp))))
goto error;
/* Store the primer pairs in the return structures */
for (i = j = 0; i < pstate->npairs; i++) {
int p1, p2, len;
/*
* Only pick pairs that have not had one or both primers rejected
* by the secondary primer-site detection code.
*/
/*
if (pstate->pairs[i].left->excl ||
pstate->pairs[i].right->excl) {
continue;
}
*/
pp[j].pair = &pstate->pairs[i];
/* Compute padded start + length for these primers. */
p1 = depad1[pstate->pairs[i].left->start];
p2 = depad1[pstate->pairs[i].left->start +
pstate->pairs[i].left->length-1];
pp[j].contig[0] = contig1;
pp[j].pos[0] = pos1l + p1;
pp[j].len[0] = p2-p1+1;
p1 = depad2[pstate->pairs[i].right->start-
pstate->pairs[i].right->length+1 - len1 -20];
p2 = depad2[pstate->pairs[i].right->start - len1 -20];
pp[j].contig[1] = contig2;
pp[j].pos[1] = pos2l + p1;
pp[j].len[1] = p2-p1+1;
/* Copy over depadded primer sequence */
len = MIN(pstate->pairs[i].left->length, MAX_PRIMER_LEN);
strncpy(pp[j].seq[0], &cons_joined[pstate->pairs[i].left->start], len);
pp[j].seq[0][len] = '\0';
len = MIN(pstate->pairs[i].right->length, MAX_PRIMER_LEN);
strncpy(pp[j].seq[1],
&cons_joined[pstate->pairs[i].right->start-
pstate->pairs[i].right->length+1],
len);
pp[j].seq[1][len] = '\0';
complement_seq(pp[j].seq[1], len);
/*
* Check if left/right primers have secondary binding sites, caching
* the result (in primer_rec.excl) to avoid subsequent searches.
*/
if (pstate->pairs[i].left->excl == 0) {
if (filter_primers(fin, 0, pp[j].seq[0]))
pstate->pairs[i].left->excl = 1;
else
pstate->pairs[i].left->excl = -1;
}
if (pstate->pairs[i].right->excl == 0) {
if (filter_primers(fin, 1, pp[j].seq[1]))
pstate->pairs[i].right->excl = 1;
else
pstate->pairs[i].right->excl = -1;
}
/* Use only if both L & R have no 2ndary match */
if (pstate->pairs[i].left->excl == -1 &&
pstate->pairs[i].right->excl == -1)
j++;
}
pstate->npairs = j;
if (!pstate->npairs) {
xfree(pp);
pp = NULL;
}
xfree(cons1);
xfree(cons2);
xfree(upcons1);
xfree(upcons2);
xfree(cons_joined);
xfree(depad1);
xfree(depad2);
return pp;
error:
if (cons1)
xfree(cons1);
if (cons2)
xfree(cons2);
if (upcons1)
xfree(upcons1);
if (upcons2)
xfree(upcons2);
if (cons_joined)
xfree(cons_joined);
if (depad1)
xfree(depad1);
if (depad2)
xfree(depad2);
if (pp)
xfree(pp);
return NULL;
}
int edview_search_consensus(edview *xx, int dir, int strand, char *value) {
int mismatches = 0; /* exact match */
int where = 2; /* consensus */
char *p;
int start, end;
char cons[WIN_WIDTH+1];
int patlen;
char *uppert, *upperb;
int found = 0, at_end = 0;
tg_rec fseq;
int fpos, i, j;
contig_t *c;
/*
* Parse value search string. It optionally includes two extra params
* separated by #. Ie:
* <string>#<N.mismatches>#<where>.
* <where> is 1 for readings, 2 for consensus, 3 for both.
*/
if (p = strchr(value, '#')) {
mismatches = atoi(p+1);
*p = 0;
if (p = strchr(p+1, '#'))
where = atoi(p+1);
}
/* uppercase search string, remove pads, and store fwd/rev copies */
patlen = strlen(value);
depad_seq(value, &patlen, NULL);
if (NULL == (uppert = (char *)xmalloc(patlen + 1)))
return 0;
if (NULL == (upperb = (char *)xmalloc(patlen + 1)))
return 0;
uppert[patlen] = upperb[patlen] = 0;
for (i = patlen-1; i >= 0; i--) {
upperb[i] = uppert[i] = toupper(value[i]);
}
complement_seq(upperb, patlen);
/* Loop */
if (dir) {
start = xx->cursor_apos + (dir ? 1 : -1);
end = start + (WIN_WIDTH-1);
} else {
end = xx->cursor_apos + (dir ? 1 : -1);
start = end - (WIN_WIDTH-1);
}
fpos = xx->cursor_apos;
c = cache_search(xx->io, GT_Contig, xx->cnum);
cache_incr(xx->io, c);
do {
char *ind, *indt = NULL, *indb = NULL;
calculate_consensus_simple(xx->io, xx->cnum, start, end, cons, NULL);
cons[WIN_WIDTH] = 0;
if (dir) {
if (strand == '+' || strand == '=')
indt = pstrstr_inexact(cons, uppert, mismatches, NULL);
if (strand == '-' || strand == '=')
indb = pstrstr_inexact(cons, upperb, mismatches, NULL);
} else {
if (strand == '+' || strand == '=')
indt = prstrstr_inexact(cons, uppert, mismatches, NULL);
if (strand == '-' || strand == '=')
indb = prstrstr_inexact(cons, upperb, mismatches, NULL);
}
if (indt && indb)
ind = MIN(indt, indb);
else if (indt)
ind = indt;
else if (indb)
ind = indb;
else
ind = NULL;
if (ind != NULL) {
if (dir) {
if (fpos <= start + ind-cons) {
found = 1;
fpos = start + ind-cons;
fseq = xx->cnum;
}
} else {
if (fpos >= start + ind-cons) {
found = 1;
fpos = start + ind-cons;
fseq = xx->cnum;
}
}
break;
}
/* Next search region - overlapping by patlen+pads */
if (dir) {
for (i = WIN_WIDTH-1, j = patlen; j && i; i--) {
if (cons[i] != '*')
j--;
}
if (i == 0)
break;
start += i;
end += i;
if (start > c->end)
at_end = 1;
} else {
for (i = 0, j = patlen; j && i < WIN_WIDTH; i++) {
if (cons[i] != '*')
j--;
}
if (i == WIN_WIDTH)
break;
start -= WIN_WIDTH-i;
end -= WIN_WIDTH-i;
if (end < c->start)
at_end = 1;
}
} while (!at_end);
cache_decr(xx->io, c);
if (found) {
edSetCursorPos(xx, fseq == xx->cnum ? GT_Contig : GT_Seq,
fseq, fpos, 1);
}
free(uppert);
free(upperb);
return found ? 0 : -1;
}
int edview_search_sequence(edview *xx, int dir, int strand, char *value) {
int mismatches = 0; /* exact match */
int where = 2; /* consensus */
char *p;
int start, end;
int patlen;
char *uppert, *upperb;
int found = 0, at_end = 0;
tg_rec fseq;
int fpos, i, j;
contig_t *c;
contig_iterator *iter;
rangec_t *(*ifunc)(GapIO *io, contig_iterator *ci);
rangec_t *r;
int best_pos;
if (dir) {
start = xx->cursor_apos + 1;
end = CITER_CEND;
iter = contig_iter_new(xx->io, xx->cnum, 1,
CITER_FIRST | CITER_ISTART,
start, end);
ifunc = contig_iter_next;
best_pos = INT_MAX;
} else {
start = CITER_CSTART;
end = xx->cursor_apos -1;
iter = contig_iter_new(xx->io, xx->cnum, 1,
CITER_LAST | CITER_IEND,
start, end);
ifunc = contig_iter_prev;
best_pos = INT_MIN;
}
if (!iter)
return -1;
/*
* Parse value search string. It optionally includes two extra params
* separated by #. Ie:
* <string>#<N.mismatches>#<where>.
* <where> is 1 for readings, 2 for consensus, 3 for both.
*/
if (p = strchr(value, '#')) {
mismatches = atoi(p+1);
*p = 0;
if (p = strchr(p+1, '#'))
where = atoi(p+1);
}
/* uppercase search string, remove pads, and store fwd/rev copies */
patlen = strlen(value);
depad_seq(value, &patlen, NULL);
if (NULL == (uppert = (char *)xmalloc(patlen + 1)))
return 0;
if (NULL == (upperb = (char *)xmalloc(patlen + 1)))
return 0;
uppert[patlen] = upperb[patlen] = 0;
for (i = patlen-1; i >= 0; i--) {
upperb[i] = uppert[i] = toupper(value[i]);
}
complement_seq(upperb, patlen);
while ((r = ifunc(xx->io, iter))) {
seq_t *s, *sorig;
char *ind, *indt = NULL, *indb = NULL, *seq;
int seq_len, comp, off = 0;
if (found && dir && r->start > best_pos)
break;
if (found && !dir && r->end < best_pos)
break;
if (NULL == (s = sorig = cache_search(xx->io, GT_Seq, r->rec)))
break;
if (r->comp ^ (s->len < 0)) {
s = dup_seq(s);
complement_seq_t(s);
}
seq = s->seq;
seq_len = ABS(s->len);
if (r->start < start) {
off = start - r->start;
seq += off;
seq_len -= off;
}
if (r->end - (patlen-1) > end)
seq_len -= r->end - (patlen-1) - end;
if (dir) {
if (strand == '+' || strand == '=')
indt = pstrnstr_inexact(seq, seq_len, uppert, patlen,
mismatches, NULL);
if (strand == '-' || strand == '=')
indb = pstrnstr_inexact(seq, seq_len, upperb, patlen,
mismatches, NULL);
} else {
if (strand == '+' || strand == '=')
indt = prstrnstr_inexact(seq, seq_len, uppert, patlen,
mismatches, NULL);
if (strand == '-' || strand == '=')
indb = prstrnstr_inexact(seq, seq_len, upperb, patlen,
mismatches, NULL);
}
if (indt && indb)
ind = MIN(indt, indb);
else if (indt)
ind = indt;
else if (indb)
ind = indb;
else
ind = NULL;
if (ind) {
int pos = r->start + ind - seq + off;
if (dir) {
if (best_pos > pos) {
found = 1;
best_pos = pos;
fpos = ind - s->seq;
fseq = r->rec;
}
} else {
if (best_pos < pos) {
found = 1;
best_pos = pos;
fpos = ind - s->seq;
fseq = r->rec;
}
}
//printf("Matches #%"PRIrec": at abs pos %d\n", r->rec, pos);
}
if (s != sorig)
free(s);
}
if (found) {
edSetCursorPos(xx, fseq == xx->cnum ? GT_Contig : GT_Seq,
fseq, fpos, 1);
}
free(uppert);
free(upperb);
contig_iter_del(iter);
return found ? 0 : -1;
}
int repeat_search (
int mode, /* 1=f, 2=r, 3=b */
int min_match, /* the minimum match length */
int **seq1_match, /* positions of matches in seq1 */
int **seq2_match, /* positions of matches in seq2 */
int **len_match, /* length of matches */
int max_mat, /* maximum number of matches */
char *seq1, /* seq1 */
int seq1_len, /* size of seq1 and its hash array */
int *num_f_matches,
int *num_r_matches
) {
int n_matches,seq2_len,max_matches,nres;
char *seq2,sense;
Hash *h;
char *depadded_seq;
int depadded_len;
int *depad_to_pad;
int i;
/* Depad sequence */
if (NULL == (depad_to_pad = (int *)xmalloc(sizeof(int) * seq1_len)))
return -1;
if (NULL == (depadded_seq = (char *)xmalloc(seq1_len+1))) {
xfree(depad_to_pad);
return -1;
}
copy_seq(depadded_seq, seq1, seq1_len);
depadded_len = seq1_len;
depad_seq(depadded_seq, &depadded_len, depad_to_pad);
seq1 = depadded_seq;
seq1_len = depadded_len;
max_matches = max_mat;
seq2_len = seq1_len;
seq2 = NULL;
if ( init_hash8n ( seq1_len, seq2_len,
8, max_mat, min_match, 1, &h )) {
free_hash8n(h);
xfree(depadded_seq);
xfree(depad_to_pad);
return -2;
}
h->seq1 = seq1;
h->seq1_len = seq1_len;
if ( hash_seqn ( h, 1)) {
verror(ERR_WARN, "hash_seqn", "sequence too short");
xfree(depadded_seq);
xfree(depad_to_pad);
return -1;
}
(void) store_hashn ( h );
if ( ! (seq2 = (char *) xmalloc ( sizeof(char)*(seq1_len) ))) {
free_hash8n ( h );
xfree(depadded_seq);
xfree(depad_to_pad);
return -1;
}
(void) copy_seq ( seq2, seq1, seq1_len );
h->seq2 = seq2;
h->seq2_len = seq2_len;
*num_f_matches = 0;
nres = 0;
n_matches = 0;
if ( mode & 1 ) {
if ( hash_seqn ( h, 2)) {
verror(ERR_WARN, "hash_seqn", "sequence too short");
free_hash8n ( h );
if (seq2) xfree(seq2);
xfree(depadded_seq);
xfree(depad_to_pad);
return -1;
}
sense = 'f';
n_matches = reps ( h, seq1_match, seq2_match, len_match, 0, sense);
*num_f_matches = n_matches;
nres += n_matches;
}
*num_r_matches = 0;
if ( mode & 2 ) {
(void) complement_seq(seq2, seq2_len);
if ( hash_seqn ( h, 2)) {
verror(ERR_WARN, "hash_seqn", "sequence too short");
free_hash8n ( h );
if (seq2) xfree(seq2);
xfree(depadded_seq);
xfree(depad_to_pad);
return -1;
}
sense = 'r';
n_matches = reps ( h, seq1_match, seq2_match, len_match, nres, sense);
*num_r_matches = n_matches;
n_matches += nres;
}
/* Remap depadded hits to padded positions */
for (i = 0; i < n_matches; i++) {
int p1, p2, p1_end;
p1 = depad_to_pad[(*seq1_match)[i]];
p2 = depad_to_pad[(*seq2_match)[i]];
p1_end = depad_to_pad[(*seq1_match)[i]+(*len_match)[i]-1];
(*seq1_match)[i] = p1;
(*seq2_match)[i] = p2;
(*len_match) [i] = p1_end - p1 + 1;
}
free_hash8n ( h );
if (seq2) xfree(seq2);
xfree(depadded_seq);
xfree(depad_to_pad);
return n_matches;
}
static int align_old(EdStruct *xx0, int pos0, int len0,
EdStruct *xx1, int pos1, int len1)
{
char *ol0,*ol1;
int *depad_to_pad0_m, *depad_to_pad1_m;
int *depad_to_pad0, *depad_to_pad1;
align_int *res, *S;
int old_def_conf0 = xx0->default_conf_n;
int old_def_conf1 = xx1->default_conf_n;
int off0 = 0, off1 = 0;
int left0 = 0, left1 = 0;
vfuncheader("Align contigs (join editor)");
/* Memory allocation */
ol0 = (char *) xmalloc(len0+1);
ol1 = (char *) xmalloc(len1+1);
depad_to_pad0 = depad_to_pad0_m = (int *)xmalloc((len0+1) * sizeof(int));
depad_to_pad1 = depad_to_pad1_m = (int *)xmalloc((len1+1) * sizeof(int));
S = res = (align_int *) xmalloc((len0+len1+1)*sizeof(align_int));
/* Compute the consensus */
DBcalcConsensus(xx0,pos0,len0,ol0,NULL,BOTH_STRANDS);
DBcalcConsensus(xx1,pos1,len1,ol1,NULL,BOTH_STRANDS);
/* Strip the pads from the consensus */
depad_seq(ol0, &len0, depad_to_pad0);
depad_seq(ol1, &len1, depad_to_pad1);
/* Do the actual alignment */
(void)calign(ol0, ol1, len0, len1,
NULL, NULL, NULL, NULL,
0, 0, gopenval, gextendval, 3, 0, res);
/* Clip left end */
if (*S != 0) {
/* Pad at start, so shift contigs */
if (*S < 0) {
left0 = -*S; /* used for display only */
depad_to_pad0 += -*S;
off0 = depad_to_pad0[0];
xx1->displayPos -= off0;
pos0 += off0;
len0 -= off0;
} else {
left1 = *S; /* used for display only */
depad_to_pad1 += *S;
off1 = depad_to_pad1[0];
xx0->displayPos -= off1;
pos1 += off1;
len1 -= off1;
}
S++;
xx0->link->lockOffset = xx1->displayPos - xx0->displayPos;
}
/* Clip right end */
{
int i = 0, j = 0, op;
align_int *S2 = S;
while (i < len0 && j < len1) {
if ((op = *S2++) == 0)
i++, j++;
else if (op > 0)
j += op;
else
i -= op;
}
len0 = i;
len1 = j;
}
/* Display the alignment. */
{
char *exp0, *exp1;
int exp_len0, exp_len1;
char name0[100];
char name1[100];
exp0 = (char *) xmalloc(len0+len1+1);
exp1 = (char *) xmalloc(len0+len1+1);
sprintf(name0, "%d", xx0->DBi->DB_contigNum);
sprintf(name1, "%d", xx1->DBi->DB_contigNum);
cexpand(ol0+left0, ol1+left1, len0, len1,
exp0, exp1, &exp_len0, &exp_len1,
ALIGN_J_SSH | ALIGN_J_PADS, S);
list_alignment(exp0, exp1, name0, name1, pos0, pos1, "");
xfree(exp0);
xfree(exp1);
}
/*************************************************************************/
/* Now actually make the edits, keeping track of old and new pads. */
openUndo(DBI(xx0));
openUndo(DBI(xx1));
xx0->default_conf_n = -1;
xx1->default_conf_n = -1;
{
int depad_pos0 = 0, depad_pos1 = 0;
int curr_pad0; /* Current padded position in seq 0 */
int curr_pad1; /* Current padded position in seq 1 */
int extra_pads; /* Difference between padded positions */
int last_pad0 = -1;
int last_pad1 = -1;
int inserted_bases0 = 0;
int inserted_bases1 = 0;
while (depad_pos0 < len0 || depad_pos1 < len1) {
if (*S < 0) {
depad_pos0 -= *S;
} else if (*S > 0) {
depad_pos1 += *S;
}
curr_pad0 = depad_to_pad0[depad_pos0]-off0;
curr_pad1 = depad_to_pad1[depad_pos1]-off1;
extra_pads = (curr_pad1 - last_pad1) - (curr_pad0 - last_pad0);
if (extra_pads < 0) { /* Add to seq 0 */
add_pads(xx1, pos1 + curr_pad1 + inserted_bases1, -extra_pads);
inserted_bases1 -= extra_pads;
} else if (extra_pads > 0) { /* Add to seq 1 */
add_pads(xx0, pos0 + curr_pad0 + inserted_bases0, extra_pads);
inserted_bases0 += extra_pads;
}
last_pad0 = curr_pad0;
last_pad1 = curr_pad1;
if (*S == 0) {
depad_pos0++;
depad_pos1++;
}
S++;
}
}
xx0->default_conf_n = old_def_conf0;
xx1->default_conf_n = old_def_conf1;
/*************************************************************************/
closeUndo(xx1, DBI(xx1));
closeUndo(xx0, DBI(xx0));
xfree(res);
xfree(ol0);
xfree(ol1);
xfree(depad_to_pad0_m);
xfree(depad_to_pad1_m);
return(0);
}
static int align(EdStruct *xx0, int pos0, int len0,
EdStruct *xx1, int pos1, int len1)
{
char *ol0,*ol1, *cons0, *cons1;
int old_def_conf0 = xx0->default_conf_n;
int old_def_conf1 = xx1->default_conf_n;
OVERLAP *overlap;
int ierr;
char PAD_SYM = '.';
int *depad_to_pad0, *dp0, *depad_to_pad1, *dp1;
int *S, *res;
int off0 = 0, off1 = 0;
int left0 = 0, left1 = 0;
vfuncheader("Align contigs (join editor)");
/* Memory allocation */
ol0 = (char *) xmalloc(len0+1);
ol1 = (char *) xmalloc(len1+1);
cons0 = (char *) xmalloc(len0+1);
cons1 = (char *) xmalloc(len1+1);
dp0 = depad_to_pad0 = (int *)xmalloc((len0+1) * sizeof(int));
dp1 = depad_to_pad1 = (int *)xmalloc((len1+1) * sizeof(int));
/* Compute the consensus */
DBcalcConsensus(xx0,pos0,len0,ol0,NULL,BOTH_STRANDS);
DBcalcConsensus(xx1,pos1,len1,ol1,NULL,BOTH_STRANDS);
memcpy(cons0, ol0, len0+1);
memcpy(cons1, ol1, len1+1);
/* Strip the pads from the consensus */
depad_seq(ol0, &len0, depad_to_pad0);
depad_seq(ol1, &len1, depad_to_pad1);
if (NULL == (overlap = create_overlap())) return -1;
init_overlap (overlap, ol0, ol1, len0, len1);
if(-1 == (ierr = align_contigs (overlap))) {
xfree(ol0);
xfree(ol1);
destroy_overlap(overlap);
return -1;
}
/*
overlap->seq1_out[overlap->right+1] = 0;
overlap->seq2_out[overlap->right+1] = 0;
*/
S = res = rsalign2myers(overlap->seq1_out, strlen(overlap->seq1_out),
overlap->seq2_out, strlen(overlap->seq2_out),
PAD_SYM);
/* Clip left end */
if (*S != 0) {
/* Pad at start, so shift contigs */
if (*S < 0) {
left0 = -*S; /* used for display only */
depad_to_pad0 += -*S;
off0 = depad_to_pad0[0];
xx1->displayPos -= off0;
pos0 += off0;
len0 -= -*S;
} else {
left1 = *S; /* used for display only */
depad_to_pad1 += *S;
off1 = depad_to_pad1[0];
xx0->displayPos -= off1;
pos1 += off1;
len1 -= *S;
}
S++;
xx0->link->lockOffset = xx1->displayPos - xx0->displayPos;
}
/* Clip right end */
{
int pos0 = 0, pos1 = 0;
int *s = S;
while (pos0 < len0 && pos1 < len1) {
if (*s < 0) {
pos0 -= *s;
} else if (*s > 0) {
pos0 += *s;
} else {
pos0++;
pos1++;
}
s++;
}
if (*s < 0)
len0 += *s;
else if (*s > 0)
len1 -= *s;
}
/* Display the alignment. */
{
char *exp0, *exp1;
int exp_len0, exp_len1;
char name0[100];
char name1[100];
exp0 = (char *) xmalloc(len0+len1+1);
exp1 = (char *) xmalloc(len0+len1+1);
sprintf(name0, "%d", xx0->DBi->DB_contigNum);
sprintf(name1, "%d", xx1->DBi->DB_contigNum);
cexpand(ol0+left0, ol1+left1, len0, len1,
exp0, exp1, &exp_len0, &exp_len1,
ALIGN_J_SSH | ALIGN_J_PADS, S);
list_alignment(exp0, exp1, name0, name1, pos0, pos1, "");
xfree(exp0);
xfree(exp1);
}
/*************************************************************************/
/* Now actually make the edits, keeping track of old and new pads. */
openUndo(DBI(xx0));
openUndo(DBI(xx1));
xx0->default_conf_n = -1;
xx1->default_conf_n = -1;
{
int depad_pos0 = 0, depad_pos1 = 0;
int curr_pad0; /* Current padded position in seq 0 */
int curr_pad1; /* Current padded position in seq 1 */
int extra_pads; /* Difference between padded positions */
int last_pad0 = -1;
int last_pad1 = -1;
int inserted_bases0 = 0;
int inserted_bases1 = 0;
while (depad_pos0 < len0 && depad_pos1 < len1) {
if (*S < 0) {
depad_pos0 -= *S;
} else if (*S > 0) {
depad_pos1 += *S;
}
if (depad_pos0 >= len0 || depad_pos1 >= len1)
break;
curr_pad0 = depad_to_pad0[depad_pos0]-off0;
curr_pad1 = depad_to_pad1[depad_pos1]-off1;
extra_pads = (curr_pad1 - last_pad1) - (curr_pad0 - last_pad0);
if (extra_pads < 0) { /* Add to seq 0 */
add_pads(xx1, pos1 + curr_pad1 + inserted_bases1, -extra_pads);
inserted_bases1 -= extra_pads;
} else if (extra_pads > 0) { /* Add to seq 1 */
add_pads(xx0, pos0 + curr_pad0 + inserted_bases0, extra_pads);
inserted_bases0 += extra_pads;
}
last_pad0 = curr_pad0;
last_pad1 = curr_pad1;
if (*S == 0) {
depad_pos0++;
depad_pos1++;
}
S++;
}
}
xx0->default_conf_n = old_def_conf0;
xx1->default_conf_n = old_def_conf1;
/*************************************************************************/
closeUndo(xx1, DBI(xx1));
closeUndo(xx0, DBI(xx0));
xfree(res);
xx0->default_conf_n = old_def_conf0;
xx1->default_conf_n = old_def_conf1;
xfree(ol0);
xfree(ol1);
xfree(dp0);
xfree(dp1);
destroy_overlap(overlap);
return(0);
}