static int difference_clip_contig(GapIO *io, int contig, int start, int end,
int *old_start, int *old_end, int add_tags)
{
int rnum;
int win_len = 31;
double disfract = 0.25;
char *seq;
int win_len2;
int start2, end2, length2;
int i, i_end, j;
char *con;
int count, worst_count;
int lowest_dis, lowest_pos;
int score, highest_score, highest_pos;
int new_left, new_right;
GReadings r;
int nbases = 0;
win_len2 = win_len / 2;
/* Get consensus sequence */
if (NULL == (con = (char *)xmalloc(io_clength(io, contig)+1)))
return -1;
calc_consensus(contig, 1, io_clength(io, contig), CON_SUM,
con, NULL, NULL, NULL,
consensus_cutoff, quality_cutoff,
database_info, (void *)io);
/* Find start reading */
for (rnum = io_clnbr(io, contig);
io_relpos(io, rnum) < start;
rnum = io_rnbr(io, rnum))
;
/* Loop through readings up to 'end' position */
for (; rnum && io_relpos(io, rnum) <= end; rnum = io_rnbr(io, rnum)) {
/*
* Find segment with the lowest difference count, averaged
* over 'win_len' bases.
*/
io_aread_seq(io, rnum, &length2, &start2, &end2, &seq, NULL, NULL, 0);
if (length2 < win_len) {
if (seq) xfree(seq);
continue;
}
count = 0;
worst_count = 0;
if (win_len > end2 - start2) {
if (seq) xfree(seq);
continue; /* too short to compare over win_len */
}
for (j = io_relpos(io, rnum) - 1, i = start2;
i < start2 + win_len && i < end2 - 1; i++, j++) {
if (!same_char(seq[i], con[j]))
count++;
}
i_end = end2 - 2 - win_len2;
lowest_dis = count;
lowest_pos = start2 + win_len2;
for (i = start2 + win_len2, j = io_relpos(io, rnum) - 1 + win_len2;
i < i_end; i++, j++) {
if (count < lowest_dis) {
lowest_dis = count;
lowest_pos = i;
}
if (count > worst_count)
worst_count = count;
count -= !same_char(seq[i-win_len2], con[j-win_len2]);
count += !same_char(seq[i+win_len2+1], con[j+win_len2+1]);
}
if (worst_count < disfract * win_len) {
if (seq) xfree(seq);
continue;
}
gel_read(io, rnum, r);
/*
* Work outwards from lowest_pos in both directions finding the
* location where the alignment score is highest. We used a fixed
* score/penalty system.
*/
score = highest_score = 0;
highest_pos = lowest_pos;
for (j = io_relpos(io, rnum) - 1 + lowest_pos - start2, i = lowest_pos;
i < r.end - 1; i++, j++) {
if (same_char(seq[i], con[j]))
score += SCORE_MATCH;
else
score += SCORE_MISMATCH;
if (score > highest_score) {
highest_score = score;
highest_pos = i;
}
}
/* printf("Read %d: Best match at %d, extend R to %d (score %d)\n",
rnum, lowest_pos, highest_pos, highest_score); */
new_right = highest_pos + 2;
score = highest_score = 0;
highest_pos = lowest_pos;
for (j = io_relpos(io, rnum) - 1 + lowest_pos - start2, i = lowest_pos;
i >= r.start; i--, j--) {
if (same_char(seq[i], con[j]))
score += SCORE_MATCH;
else
score += SCORE_MISMATCH;
if (score > highest_score) {
highest_score = score;
highest_pos = i;
}
}
/* printf("Read %d: Best match at %d, extend L to %d (score %d)\n",
rnum, lowest_pos, highest_pos, highest_score); */
new_left = highest_pos;
if (new_right < new_left + 2)
new_right = new_left + 2;
/* Tag the change */
if (add_tags && new_left != r.start) {
char buf[100];
int tag_st;
if (r.sense) {
tag_st = r.length - new_left + 1;
} else {
tag_st = r.start + 1;
}
sprintf(buf, "Difference clipped from old start at %d\n", r.start);
insert_NEW_tag(io, rnum, tag_st, new_left - r.start, "DIFF",
buf, 2);
}
if (add_tags && new_right != r.end) {
char buf[100];
int tag_st;
if (r.sense) {
tag_st = r.length - r.end + 2;
} else {
tag_st = new_right;
}
sprintf(buf, "Difference clipped from old end at %d\n", r.end);
insert_NEW_tag(io, rnum, tag_st, r.end - new_right, "DIFF",
buf, 2);
}
/* Update reading positions/length */
gel_read(io, rnum, r);
r.position += new_left - r.start;
old_start[rnum] = r.start;
old_end[rnum] = r.end;
r.start = new_left;
r.end = new_right;
nbases += r.sequence_length - (r.end - r.start - 1);
r.sequence_length = r.end - r.start - 1;
gel_write(io, rnum, r);
io_relpos(io, rnum) = r.position;
io_length(io, rnum) = r.sense ? -r.sequence_length : r.sequence_length;
if (seq) xfree(seq);
}
xfree(con);
return nbases;
}
static void seg_test(
__segment seg_var,
char __far* seg_beg,
char __far* elements,
unsigned line_no )
{
char __based( seg_var ) *pseg;
pseg = (unsigned)(elements - seg_beg);
same_char( 'a', *(pseg+0), line_no );
same_char( 'b', *(pseg+1), line_no );
same_char( 'c', *(pseg+2), line_no );
same_char( 'd', *(pseg+3), line_no );
same_char( 'e', *(pseg+4), line_no );
}
static void addrs_of()
{
char *p01 = &char_file_scope; // regular
same_char( *p01, 'q', __LINE__ );
char __far16 *p02 = &char_file_scope; // far16
same_char( *p02, 'q', __LINE__ );
char __based( base_ptr ) *p03 = &array_file_scope[2];
same_char( *p03, 't', __LINE__ );
char _far *p04;
p04 = p01; same_char( *p04, 'q', __LINE__ );
p04 = p02; same_char( *p04, 'q', __LINE__ );
p04 = p03; same_char( *p04, 't', __LINE__ );
same_char( 'j', char_area, __LINE__ );
same_char( 'c', array_area[2], __LINE__ );
same_char( 'j', char_data, __LINE__ );
same_char( 'c', array_data[2], __LINE__ );
same_char( 'j', char_code, __LINE__ );
same_char( 'c', array_code[2], __LINE__ );
same_char( 'j', char_text, __LINE__ );
same_char( 'c', array_text[2], __LINE__ );
}
