static int create_block_features(GtBEDParser *bed_parser, GtFeatureNode *fn,
GtUword block_count,
GtSplitter *size_splitter,
GtSplitter *start_splitter, GtIO *bed_file,
GtError *err)
{
GtUword i;
int had_err = 0;
gt_assert(fn && block_count && size_splitter && start_splitter);
gt_assert(gt_splitter_size(size_splitter) == block_count);
gt_assert(gt_splitter_size(start_splitter) == block_count);
for (i = 0; !had_err && i < block_count; i++) {
GtUword block_size, block_start, start, end;
GtGenomeNode *block;
const char *name;
if (gt_parse_uword(&block_size, gt_splitter_get_token(size_splitter, i))) {
gt_error_set(err,
"file \"%s\": line "GT_WU": could not parse blockSize '%s'",
gt_io_get_filename(bed_file),
gt_io_get_line_number(bed_file),
gt_splitter_get_token(size_splitter, i));
had_err = -1;
}
if (!had_err && gt_parse_uword(&block_start,
gt_splitter_get_token(start_splitter, i))) {
gt_error_set(err, "file \"%s\": line "GT_WU": could not parse blockStart "
"'%s'", gt_io_get_filename(bed_file),
gt_io_get_line_number(bed_file),
gt_splitter_get_token(start_splitter, i));
had_err = -1;
}
if (!had_err) {
start = gt_genome_node_get_start((GtGenomeNode*) fn) + block_start;
end = start + block_size - 1;
block = gt_feature_node_new(gt_genome_node_get_seqid((GtGenomeNode*) fn),
bed_parser->block_type
? bed_parser->block_type
: BED_BLOCK_TYPE,
start, end, gt_feature_node_get_strand(fn));
if ((name = gt_feature_node_get_attribute(fn, GT_GFF_NAME))) {
gt_feature_node_add_attribute((GtFeatureNode*) block, GT_GFF_NAME,
name);
}
gt_feature_node_set_score((GtFeatureNode*) block,
gt_feature_node_get_score(fn));
gt_feature_node_set_strand((GtFeatureNode*) block,
gt_feature_node_get_strand(fn));
gt_feature_node_add_child(fn, (GtFeatureNode*) block);
}
}
return had_err;
}
static void pbs_attach_results_to_gff3(GtPBSResults *results,
GtLTRElement *element,
GtStrand *canonical_strand,
GtStr *tag)
{
GtRange pbs_range;
GtGenomeNode *gf;
unsigned long i = 0;
char buffer[BUFSIZ];
GtPBSHit* hit = gt_pbs_results_get_ranked_hit(results, i++);
if (*canonical_strand == GT_STRAND_UNKNOWN)
*canonical_strand = gt_pbs_hit_get_strand(hit);
else
{
/* do we have to satisfy a strand constraint?
* then find best-scoring PBS on the given canonical strand */
while (gt_pbs_hit_get_strand(hit) != *canonical_strand
&& i < gt_pbs_results_get_number_of_hits(results))
{
gt_log_log("dropping PBS because of nonconsistent strand: %s\n",
gt_feature_node_get_attribute(element->mainnode, "ID"));
hit = gt_pbs_results_get_ranked_hit(results, i++);
}
/* if there is none, do not report a PBS */
if (gt_pbs_hit_get_strand(hit) != *canonical_strand)
return;
}
pbs_range = gt_pbs_hit_get_coords(hit);
pbs_range.start++; pbs_range.end++; /* GFF3 is 1-based */
gf = gt_feature_node_new(gt_genome_node_get_seqid((GtGenomeNode*)
element->mainnode),
GT_PBS_TYPE,
pbs_range.start,
pbs_range.end,
gt_pbs_hit_get_strand(hit));
gt_feature_node_set_source((GtFeatureNode*) gf, tag);
gt_feature_node_set_score((GtFeatureNode*) gf,
(float) gt_pbs_hit_get_score(hit));
if (gt_pbs_hit_get_trna(hit) != NULL) {
gt_feature_node_add_attribute((GtFeatureNode*) gf, "trna",
gt_pbs_hit_get_trna(hit));
}
gt_feature_node_set_strand(element->mainnode, gt_pbs_hit_get_strand(hit));
(void) snprintf(buffer, BUFSIZ-1, "%lu", gt_pbs_hit_get_tstart(hit));
gt_feature_node_add_attribute((GtFeatureNode*) gf, "trnaoffset", buffer);
(void) snprintf(buffer, BUFSIZ-1, "%lu", gt_pbs_hit_get_offset(hit));
gt_feature_node_add_attribute((GtFeatureNode*) gf, "pbsoffset", buffer);
(void) snprintf(buffer, BUFSIZ-1, "%lu", gt_pbs_hit_get_edist(hit));
gt_feature_node_add_attribute((GtFeatureNode*) gf, "edist", buffer);
gt_feature_node_add_child(element->mainnode, (GtFeatureNode*) gf);
}
static void construct_thick_feature(GtBEDParser *bed_parser, GtFeatureNode *fn,
GtRange range)
{
GtGenomeNode *thick_feature;
const char *name;
gt_assert(fn);
thick_feature = gt_feature_node_new(gt_genome_node_get_seqid((GtGenomeNode*)
fn),
bed_parser->thick_feature_type
? bed_parser->thick_feature_type
: BED_THICK_FEATURE_TYPE,
range.start, range.end,
gt_feature_node_get_strand(fn));
if ((name = gt_feature_node_get_attribute(fn, "Name")))
gt_feature_node_add_attribute((GtFeatureNode*) thick_feature, "Name", name);
gt_feature_node_set_score((GtFeatureNode*) thick_feature,
gt_feature_node_get_score(fn));
gt_feature_node_set_strand((GtFeatureNode*) thick_feature,
gt_feature_node_get_strand(fn));
gt_feature_node_add_child(fn, (GtFeatureNode*) thick_feature);
}
static void ppt_attach_results_to_gff3(GtPPTResults *results,
GtLTRElement *element,
GtStrand *canonical_strand,
GtStr *tag)
{
GtRange ppt_range;
unsigned long i = 0;
GtGenomeNode *gf;
GtPPTHit* hit = gt_ppt_results_get_ranked_hit(results, i++);
if (*canonical_strand == GT_STRAND_UNKNOWN)
*canonical_strand = gt_ppt_hit_get_strand(hit);
else
{
/* find best-scoring PPT on the given canonical strand */
while (gt_ppt_hit_get_strand(hit) != *canonical_strand
&& i < gt_ppt_results_get_number_of_hits(results))
{
gt_log_log("dropping PPT because of nonconsistent strand: %s\n",
gt_feature_node_get_attribute(element->mainnode, "ID"));
hit = gt_ppt_results_get_ranked_hit(results, i++);
}
/* if there is none, do not report a PPT */
if (gt_ppt_hit_get_strand(hit) != *canonical_strand)
return;
}
ppt_range = gt_ppt_hit_get_coords(hit);
ppt_range.start++; ppt_range.end++; /* GFF3 is 1-based */
gf = gt_feature_node_new(gt_genome_node_get_seqid((GtGenomeNode*)
element->mainnode),
GT_PPT_TYPE,
ppt_range.start,
ppt_range.end,
gt_ppt_hit_get_strand(hit));
gt_feature_node_set_source((GtFeatureNode*) gf, tag);
gt_feature_node_set_strand(element->mainnode, gt_ppt_hit_get_strand(hit));
gt_feature_node_add_child(element->mainnode, (GtFeatureNode*) gf);
}
static int bed_rest(GtBEDParser *bed_parser, GtIO *bed_file, GtError *err)
{
GtUword block_count = 0;
GtGenomeNode *gn = NULL;
GtRange range;
GtStr *seqid;
int had_err;
gt_error_check(err);
/* column 1.: chrom */
seqid = get_seqid(bed_parser);
had_err = skip_blanks(bed_file, err);
/* column 2.: chromStart */
if (!had_err) {
word(bed_parser->word, bed_file);
had_err = skip_blanks(bed_file, err);
}
/* column 3.: chromEnd */
if (!had_err) {
word(bed_parser->another_word, bed_file);
had_err = parse_bed_range(&range, bed_parser->word,
bed_parser->another_word, bed_parser->offset,
bed_file, false, err);
}
if (!had_err) {
/* add region */
gt_region_node_builder_add_region(bed_parser->region_node_builder,
gt_str_get(seqid), range);
/* create feature */
gn = gt_feature_node_new(seqid,
bed_parser->feature_type
? bed_parser->feature_type
: BED_FEATURE_TYPE,
range.start, range.end, GT_STRAND_BOTH);
gt_queue_add(bed_parser->feature_nodes, gn);
if (bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
}
/* optional column 4.: name */
if (!had_err) {
word(bed_parser->word, bed_file);
if (gt_str_length(bed_parser->word)) {
gt_feature_node_add_attribute((GtFeatureNode*) gn, GT_GFF_NAME,
gt_str_get(bed_parser->word));
}
if (bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
}
/* optional column 5.: score */
if (!had_err) {
word(bed_parser->word, bed_file);
if (gt_str_length(bed_parser->word)) {
bool score_is_defined;
float score_value;
had_err = gt_parse_score(&score_is_defined, &score_value,
gt_str_get(bed_parser->word),
gt_io_get_line_number(bed_file),
gt_io_get_filename(bed_file), err);
if (!had_err && score_is_defined)
gt_feature_node_set_score((GtFeatureNode*) gn, score_value);
}
}
if (!had_err && bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
/* optional column 6.: strand */
if (!had_err) {
word(bed_parser->word, bed_file);
if (gt_str_length(bed_parser->word)) {
GtStrand strand;
had_err = gt_parse_strand(&strand, gt_str_get(bed_parser->word),
gt_io_get_line_number(bed_file),
gt_io_get_filename(bed_file), err);
if (!had_err)
gt_feature_node_set_strand((GtFeatureNode*) gn, strand);
}
}
if (!had_err && bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
/* optional column 7.: thickStart */
if (!had_err) {
word(bed_parser->word, bed_file);
if (bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
}
/* optional column 8.: thickEnd */
if (!had_err) {
word(bed_parser->another_word, bed_file);
if (gt_str_length(bed_parser->another_word)) {
gt_assert(gt_str_length(bed_parser->word));
/* got a thickStart and a thickEnd -> construct corresponding feature */
had_err = parse_bed_range(&range, bed_parser->word,
bed_parser->another_word, bed_parser->offset,
bed_file, true, err);
if (!had_err && range.start <= range.end)
construct_thick_feature(bed_parser, (GtFeatureNode*) gn, range);
}
}
if (!had_err && bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
/* optional column 9.: itemRgb */
if (!had_err) {
word(bed_parser->word, bed_file);
/* we do not use the RGB values */
if (bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
}
/* optional column 10.: blockCount */
if (!had_err) {
word(bed_parser->word, bed_file);
if (gt_str_length(bed_parser->word)) {
if (gt_parse_uword(&block_count, gt_str_get(bed_parser->word))) {
gt_error_set(err,
"file \"%s\": line "GT_WU": could not parse blockCount",
gt_io_get_filename(bed_file),
gt_io_get_line_number(bed_file));
had_err = -1;
}
else {
/* reset to parse/process blockSizes and blockStarts properly */
gt_str_reset(bed_parser->word);
gt_str_reset(bed_parser->another_word);
}
}
}
if (!had_err && bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
/* optional column 11.: blockSizes */
if (!had_err) {
word(bed_parser->word, bed_file);
if (bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
}
/* optional column 12.: blockStarts */
if (!had_err) {
word(bed_parser->another_word, bed_file);
if (bed_separator(bed_file))
had_err = skip_blanks(bed_file, err);
}
/* process blocks if necessary */
if (!had_err && block_count) {
had_err = process_blocks(bed_parser, (GtFeatureNode*) gn, block_count,
bed_parser->word, bed_parser->another_word,
bed_file, err);
}
/* the end of the line should now be reached */
if (!had_err)
had_err = gt_io_expect(bed_file, GT_END_OF_LINE, err);
return had_err;
}
static int gt_ltrdigest_pdom_visitor_choose_strand(GtLTRdigestPdomVisitor *lv)
{
int had_err = 0;
double log_eval_fwd = 0.0,
log_eval_rev = 0.0;
GtFeatureNodeIterator *fni;
GtStrand strand;
double score;
bool seen_fwd = false,
seen_rev = false;
GtFeatureNode *curnode = NULL;
GtUword i;
GtArray *to_delete;
fni = gt_feature_node_iterator_new(lv->ltr_retrotrans);
while (!had_err && (curnode = gt_feature_node_iterator_next(fni))) {
if (strcmp(gt_feature_node_get_type(curnode),
gt_ft_protein_match) == 0) {
strand = gt_feature_node_get_strand(curnode);
score = (double) gt_feature_node_get_score(curnode);
if (strand == GT_STRAND_FORWARD) {
log_eval_fwd += log(score);
seen_fwd = true;
} else if (strand == GT_STRAND_REVERSE) {
log_eval_rev += log(score);
seen_rev = true;
}
}
}
gt_feature_node_iterator_delete(fni);
if (seen_rev && !seen_fwd)
gt_feature_node_set_strand(lv->ltr_retrotrans, GT_STRAND_REVERSE);
else if (!seen_rev && seen_fwd)
gt_feature_node_set_strand(lv->ltr_retrotrans, GT_STRAND_FORWARD);
else if (!seen_rev && !seen_fwd)
return had_err;
else {
gt_assert(seen_rev && seen_fwd);
if (gt_double_compare(log_eval_fwd, log_eval_rev) < 0)
strand = GT_STRAND_FORWARD;
else
strand = GT_STRAND_REVERSE;
gt_feature_node_set_strand(lv->ltr_retrotrans, strand);
to_delete = gt_array_new(sizeof (GtFeatureNode*));
fni = gt_feature_node_iterator_new(lv->ltr_retrotrans);
while (!had_err && (curnode = gt_feature_node_iterator_next(fni))) {
if (strcmp(gt_feature_node_get_type(curnode),
gt_ft_protein_match) == 0) {
if (strand != gt_feature_node_get_strand(curnode)) {
gt_array_add(to_delete, curnode);
}
}
}
gt_feature_node_iterator_delete(fni);
gt_assert(gt_array_size(to_delete) > 0);
for (i = 0; i < gt_array_size(to_delete); i++) {
gt_feature_node_remove_leaf(lv->ltr_retrotrans,
*(GtFeatureNode**) gt_array_get(to_delete,
i));
}
gt_array_delete(to_delete);
}
return had_err;
}
