static const char* get_node_name_or_id(GtFeatureNode *gn)
{
const char *ret;
if (!gn) return NULL;
if (!(ret = gt_feature_node_get_attribute(gn, GT_GFF_NAME))) {
if (!(ret = gt_feature_node_get_attribute(gn, GT_GFF_ID)))
ret = NULL;
}
return ret;
}
static void infer_cds_visitor_check_cds_multi(AgnInferCDSVisitor *v)
{
if(gt_array_size(v->cds) <= 1)
{
return;
}
GtFeatureNode **firstsegment = gt_array_get(v->cds, 0);
const char *id = gt_feature_node_get_attribute(*firstsegment, "ID");
if(id == NULL)
{
char newid[64];
sprintf(newid, "CDS%lu", v->cdscounter++);
gt_feature_node_add_attribute(*firstsegment, "ID", newid);
}
gt_feature_node_make_multi_representative(*firstsegment);
GtUword i;
for(i = 0; i < gt_array_size(v->cds); i++)
{
GtFeatureNode **segment = gt_array_get(v->cds, i);
if(!gt_feature_node_is_multi(*segment))
{
gt_feature_node_set_multi_representative(*segment, *firstsegment);
}
}
}
static GtFeatureNode* find_root(const GtFeatureInfo *fi, const char *id)
{
const char *delim, *parents;
GtFeatureNode *this_feature, *parent_pseudo_feature;
gt_assert(fi && id);
/* get feature */
delim = strchr(id, ';');
if (delim) {
char *first_parent = gt_cstr_dup_nt(id, delim - id);
this_feature = gt_hashmap_get(fi->id_to_genome_node, first_parent);
parent_pseudo_feature = gt_hashmap_get(fi->id_to_pseudo_parent,
first_parent);
gt_free(first_parent);
}
else {
this_feature = gt_hashmap_get(fi->id_to_genome_node, id);
parent_pseudo_feature = gt_hashmap_get(fi->id_to_pseudo_parent, id);
}
gt_assert(this_feature);
/* recursion */
parents = gt_feature_node_get_attribute(this_feature, GT_GFF_PARENT);
if (parents)
return find_root(fi, parents);
else if (parent_pseudo_feature)
return parent_pseudo_feature;
return this_feature;
}
static void split_cds_feature(GtFeatureNode *cds_feature, GtFeatureNode *fn)
{
GtArray *parents;
unsigned long i;
gt_assert(cds_feature && fn);
/* find parents */
parents = find_cds_parents(cds_feature, fn);
/* remove CDS feature */
gt_feature_node_remove_leaf(fn, cds_feature);
/* add CDS feature to all parents */
for (i = 0; i < gt_array_size(parents); i++) {
GtFeatureNode *parent = *(GtFeatureNode**) gt_array_get(parents, i);
const char *id = gt_feature_node_get_attribute(parent, GT_GFF_ID);
if (!i) {
gt_feature_node_set_attribute(cds_feature, GT_GFF_PARENT, id);
gt_feature_node_add_child(parent, cds_feature);
}
else {
GtFeatureNode *new_cds = gt_feature_node_clone(cds_feature);
gt_feature_node_set_attribute(new_cds, GT_GFF_PARENT, id);
gt_feature_node_add_child(parent, new_cds);
gt_genome_node_delete((GtGenomeNode*) cds_feature);
}
}
gt_array_delete(parents);
}
static int m2i_change_seqid(GtFeatureNode *fn, void *data, GtError *err)
{
const char *target;
M2IChangeSeqidInfo *info = (M2IChangeSeqidInfo*) data;
gt_error_check(err);
gt_assert(fn && info);
gt_genome_node_change_seqid((GtGenomeNode*) fn, info->new_seqid);
if ((target = gt_feature_node_get_attribute(fn, GT_GFF_TARGET)))
return m2i_change_target_seqids(fn, target, info->region_mapping, err);
return 0;
}
void gt_feature_info_replace_pseudo_parent(GtFeatureInfo *fi,
GtFeatureNode *child,
GtFeatureNode *new_pseudo_parent)
{
const char *id;
gt_assert(fi && child && new_pseudo_parent);
gt_assert(gt_feature_node_is_pseudo((GtFeatureNode*) new_pseudo_parent));
id = gt_feature_node_get_attribute(child, GT_GFF_ID);
gt_assert(id);
gt_hashmap_remove(fi->id_to_pseudo_parent, id);
gt_feature_info_add_pseudo_parent(fi, id, new_pseudo_parent);
}
static void infer_cds_visitor_check_stop(AgnInferCDSVisitor *v)
{
if(gt_array_size(v->cds) == 0)
return;
const char *mrnaid = gt_feature_node_get_attribute(v->mrna, "ID");
unsigned int ln = gt_genome_node_get_line_number((GtGenomeNode *)v->mrna);
GtStrand strand = gt_feature_node_get_strand(v->mrna);
GtRange stoprange;
GtUword threeprimeindex = gt_array_size(v->cds) - 1;
GtGenomeNode **threeprimesegment = gt_array_get(v->cds, threeprimeindex);
stoprange = gt_genome_node_get_range(*threeprimesegment);
stoprange.start = stoprange.end - 2;
if(strand == GT_STRAND_REVERSE)
{
threeprimesegment = gt_array_get(v->cds, 0);
stoprange = gt_genome_node_get_range(*threeprimesegment);
stoprange.end = stoprange.start + 2;
}
if(gt_array_size(v->stops) > 1)
{
gt_logger_log(v->logger, "mRNA '%s' (line %u) has %lu stop codons", mrnaid,
ln, gt_array_size(v->starts));
}
else if(gt_array_size(v->stops) == 1)
{
GtGenomeNode **codon = gt_array_get(v->stops, 0);
GtRange testrange = gt_genome_node_get_range(*codon);
if(gt_range_compare(&stoprange, &testrange) != 0)
{
gt_logger_log(v->logger, "stop codon inferred from CDS [%lu, %lu] does "
"not match explicitly provided stop codon [%lu, %lu] for "
"mRNA '%s'", stoprange.start, stoprange.end,
testrange.start, testrange.end, mrnaid);
}
}
else // agn_assert(gt_array_size(v->stops) == 0)
{
GtStr *seqid = gt_genome_node_get_seqid((GtGenomeNode *)v->mrna);
GtGenomeNode *codonfeature = gt_feature_node_new(seqid, "stop_codon",
stoprange.start,
stoprange.end,
strand);
if(v->source)
gt_feature_node_set_source((GtFeatureNode *)codonfeature, v->source);
GtFeatureNode *cf = (GtFeatureNode *)codonfeature;
gt_feature_node_add_child(v->mrna, cf);
gt_array_add(v->stops, cf);
}
}
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 int feature_node_lua_get_attribute(lua_State *L)
{
GtGenomeNode **gn = check_genome_node(L, 1);
const char *attr = NULL, *attrval = NULL;
attr = luaL_checkstring(L, 2);
GtFeatureNode *fn;
/* make sure we get a feature node */
fn = gt_feature_node_try_cast(*gn);
luaL_argcheck(L, fn, 1, "not a feature node");
attrval = gt_feature_node_get_attribute(fn, attr);
if (attrval)
lua_pushstring(L, attrval);
else
lua_pushnil(L);
return 1;
}
static int store_ids(GtFeatureNode *fn, void *data, GtError *err)
{
GtGFF3Visitor *gff3_visitor = (GtGFF3Visitor*) data;
AddIDInfo add_id_info;
int had_err = 0;
GtStr *id;
gt_error_check(err);
gt_assert(fn && gff3_visitor);
if (gt_feature_node_has_children(fn) || gt_feature_node_is_multi(fn) ||
(gff3_visitor->retain_ids && gt_feature_node_get_attribute(fn, "ID"))) {
if (gt_feature_node_is_multi(fn)) {
id = gt_hashmap_get(gff3_visitor->feature_node_to_unique_id_str,
gt_feature_node_get_multi_representative(fn));
if (!id) {
/* the representative does not have its own id yet -> create it */
if (gff3_visitor->retain_ids) {
id = make_id_unique(gff3_visitor,
gt_feature_node_get_multi_representative(fn));
}
else {
id = create_unique_id(gff3_visitor,
gt_feature_node_get_multi_representative(fn));
}
}
/* store id for feature, if the feature was not the representative */
if (gt_feature_node_get_multi_representative(fn) != fn) {
gt_hashmap_add(gff3_visitor->feature_node_to_unique_id_str, fn,
gt_str_ref(id));
}
}
else {
if (gff3_visitor->retain_ids)
id = make_id_unique(gff3_visitor, fn);
else
id = create_unique_id(gff3_visitor, fn);
}
/* for each child -> store the parent feature in the hash map */
add_id_info.gt_feature_node_to_id_array =
gff3_visitor->feature_node_to_id_array,
add_id_info.id = gt_str_get(id);
had_err = gt_feature_node_traverse_direct_children(fn, &add_id_info, add_id,
err);
}
return had_err;
}
static int gff3_visitor_feature_node(GtNodeVisitor *nv, GtFeatureNode *fn,
GtError *err)
{
GtGFF3Visitor *gff3_visitor;
int had_err;
gt_error_check(err);
gff3_visitor = gff3_visitor_cast(nv);
gff3_version_string(nv);
had_err = gt_feature_node_traverse_children(fn, gff3_visitor, store_ids, true,
err);
if (!had_err) {
if (gt_feature_node_is_tree(fn)) {
had_err = gt_feature_node_traverse_children(fn, gff3_visitor,
gff3_show_feature_node, true,
err);
}
else {
/* got a DAG -> traverse in topologically sorted depth first fashion to
make sure that the 'Parent' attributes are shown in correct order */
had_err =
gt_feature_node_traverse_children_top(fn, gff3_visitor,
gff3_show_feature_node, err);
}
}
/* reset hashmaps */
gt_hashmap_reset(gff3_visitor->feature_node_to_id_array);
gt_hashmap_reset(gff3_visitor->feature_node_to_unique_id_str);
/* show terminator, if the feature has children (otherwise it is clear that
the feature is complete, because no ID attribute has been shown) */
if (gt_feature_node_has_children(fn) ||
(gff3_visitor->retain_ids && gt_feature_node_get_attribute(fn, "ID"))) {
if (!gff3_visitor->outstr)
gt_file_xprintf(gff3_visitor->outfp, "%s\n", GT_GFF_TERMINATOR);
else {
gt_str_append_cstr(gff3_visitor->outstr, GT_GFF_TERMINATOR);
gt_str_append_char(gff3_visitor->outstr, '\n');
}
}
return had_err;
}
static int
gaeval_visitor_visit_feature_node(GtNodeVisitor *nv, GtFeatureNode *fn,
GtError *error)
{
AgnGaevalVisitor *v = gaeval_visitor_cast(nv);
gt_error_check(error);
GtFeatureNodeIterator *feats = gt_feature_node_iterator_new(fn);
GtFeatureNode *tempfeat;
for(tempfeat = gt_feature_node_iterator_next(feats);
tempfeat != NULL;
tempfeat = gt_feature_node_iterator_next(feats))
{
if(agn_typecheck_mrna(tempfeat) == false)
continue;
double coverage = gaeval_visitor_calculate_coverage(v, tempfeat, error);
char covstr[16];
sprintf(covstr, "%.3lf", coverage);
gt_feature_node_add_attribute(tempfeat, "gaeval_coverage", covstr);
double integrity_components[5];
double integrity = gaeval_visitor_calculate_integrity(
v, tempfeat, coverage, integrity_components, error
);
char intstr[16];
sprintf(intstr, "%.3lf", integrity);
gt_feature_node_add_attribute(tempfeat, "gaeval_integrity", intstr);
if(v->tsvout)
{
const char *mrnaid = gt_feature_node_get_attribute(tempfeat, "ID");
const char *mrnalabel = agn_feature_node_get_label(tempfeat);
GtUword num_introns = agn_typecheck_count(tempfeat, agn_typecheck_intron);
fprintf(v->tsvout, "%s\t%s\t%s\t%s\t%lu\t%.3lf\t%.3lf\t%.3lf\t%.3lf\n",
mrnaid, mrnalabel, intstr, covstr, num_introns,
integrity_components[0], integrity_components[1],
integrity_components[2], integrity_components[3]);
}
}
gt_feature_node_iterator_delete(feats);
return 0;
}
void gt_orphanage_add(GtOrphanage *o, GtGenomeNode *orphan,
const char *orphan_id, GtStrArray *missing_parents)
{
const char *missing_parent;
GtUword i;
gt_assert(o && orphan);
gt_assert(gt_feature_node_get_attribute((GtFeatureNode*) orphan,
GT_GFF_PARENT));
gt_queue_add(o->orphans, orphan);
if (orphan_id && !gt_cstr_table_get(o->orphan_ids, orphan_id))
gt_cstr_table_add(o->orphan_ids, orphan_id);
if (missing_parents) {
for (i = 0; i < gt_str_array_size(missing_parents); i++) {
missing_parent = gt_str_array_get(missing_parents, i);
if (!gt_cstr_table_get(o->missing_parents, missing_parent))
gt_cstr_table_add(o->missing_parents, missing_parent);
}
}
}
static bool filter_targetstrand(GtFeatureNode *fn, GtStrand targetstrand)
{
const char *target;
gt_assert(fn);
if (targetstrand != GT_NUM_OF_STRAND_TYPES &&
(target = gt_feature_node_get_attribute(fn, GT_GFF_TARGET))) {
unsigned long num_of_targets;
GtStrand parsed_strand;
GT_UNUSED int had_err;
had_err = gt_gff3_parser_parse_target_attributes(target, &num_of_targets,
NULL, NULL, &parsed_strand,
"", 0, NULL);
gt_assert(!had_err);
if (num_of_targets == 1 && parsed_strand != GT_NUM_OF_STRAND_TYPES &&
parsed_strand != targetstrand) {
return true;
}
}
return false;
}
static int extracttarget_from_node(GtGenomeNode *gn, GtStrArray *seqfiles,
GtError *err)
{
GtFeatureNodeIterator *fni;
int had_err = 0;
gt_error_check(err);
gt_assert(gn && seqfiles);
if (gt_genome_node_cast(gt_feature_node_class(), gn)) {
const char *target;
GtFeatureNode *child;
fni = gt_feature_node_iterator_new(gt_feature_node_cast(gn));
while (!had_err && /* XXX remove cast */
(child = (GtFeatureNode*) gt_feature_node_iterator_next(fni))) {
if ((target = gt_feature_node_get_attribute(child, "Target")))
had_err = extracttarget_from_seqfiles(target, seqfiles, err);
}
gt_feature_node_iterator_delete(fni);
}
return had_err;
}
static GtStr* make_id_unique(GtGFF3Visitor *gff3_visitor, GtFeatureNode *fn)
{
GtUword i = 1;
GtStr *id = gt_str_new_cstr(gt_feature_node_get_attribute(fn, "ID"));
if (gt_cstr_table_get(gff3_visitor->used_ids, gt_str_get(id))) {
GtStr *buf = gt_str_new();
while (!id_string_is_unique(id, buf, gff3_visitor->used_ids, i++));
gt_warning("feature ID \"%s\" not unique: changing to %s", gt_str_get(id),
gt_str_get(buf));
gt_str_set(id, gt_str_get(buf));
gt_str_delete(buf);
}
/* update table with the new id */
gt_cstr_table_add(gff3_visitor->used_ids, gt_str_get(id));
/* store (unique) id */
gt_hashmap_add(gff3_visitor->feature_node_to_unique_id_str, fn, id);
return id;
}
static void filter_targetbest(GtFeatureNode *current_feature,
GtDlist *trees, GtHashmap *target_to_elem)
{
unsigned long num_of_targets;
GtDlistelem *previous_elem;
GtStr *first_target_id;
const char *target;
int had_err;
gt_assert(current_feature && trees);
target = gt_feature_node_get_attribute(current_feature, TARGET_STRING);
gt_assert(target);
first_target_id = gt_str_new();
had_err = gt_gff3_parser_parse_target_attributes(target, &num_of_targets,
first_target_id, NULL, NULL,
"", 0, NULL);
gt_assert(!had_err);
if (num_of_targets == 1) {
GtStr *key = gt_str_new();
build_key(key, current_feature, first_target_id);
if (!(previous_elem = gt_hashmap_get(target_to_elem, gt_str_get(key)))) {
/* element with this target_id not included yet -> include it */
include_feature(trees, target_to_elem, current_feature, key);
}
else {
GtFeatureNode *previous_feature = gt_dlistelem_get_data(previous_elem);
/* element with this target_id included already -> compare them */
if (gt_feature_node_get_score(current_feature) >
gt_feature_node_get_score(previous_feature)) {
/* current feature is better -> replace previous feature */
replace_previous_elem(previous_elem, current_feature, trees,
target_to_elem, key);
}
else /* current feature is not better -> remove it */
gt_genome_node_delete((GtGenomeNode*) current_feature);
}
gt_str_delete(key);
}
else
gt_dlist_add(trees, current_feature);
gt_str_delete(first_target_id);
}
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 int targetbest_filter_stream_next(GtNodeStream *gs, GtGenomeNode **gn,
GtError *err)
{
GtTargetbestFilterStream *tfs;
GtGenomeNode *node;
int had_err = 0;
gt_error_check(err);
tfs = targetbest_filter_stream_cast(gs);
if (!tfs->in_stream_processed) {
while (!(had_err = gt_node_stream_next(tfs->in_stream, &node, err)) &&
node) {
if (gt_feature_node_try_cast(node) &&
gt_feature_node_get_attribute((GtFeatureNode*) node, "Target")) {
filter_targetbest((GtFeatureNode*) node, tfs->trees,
tfs->target_to_elem);
}
else
gt_dlist_add(tfs->trees, node);
}
tfs->next = gt_dlist_first(tfs->trees);
tfs->in_stream_processed = true;
}
if (!had_err) {
gt_assert(tfs->in_stream_processed);
if (tfs->next) {
*gn = gt_dlistelem_get_data(tfs->next);
tfs->next = gt_dlistelem_next(tfs->next);
}
else
*gn = NULL;
return 0;
}
return had_err;
}
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 gt_extract_feature_sequence_generic(GtStr *sequence,
GtGenomeNode *gn,
const char *type, bool join, GtStr *seqid,
GtStrArray *target_ids,
unsigned int *out_phase_offset,
GtRegionMapping *region_mapping, GtError *err)
{
GtFeatureNode *fn;
GtRange range;
unsigned int phase_offset = 0;
char *outsequence;
const char *target;
int had_err = 0;
gt_error_check(err);
fn = gt_genome_node_cast(gt_feature_node_class(), gn);
gt_assert(fn);
if (seqid)
gt_str_append_str(seqid, gt_genome_node_get_seqid(gn));
if (target_ids &&
(target = gt_feature_node_get_attribute(fn, GT_GFF_TARGET))) {
had_err = gt_gff3_parser_parse_all_target_attributes(target, false,
target_ids, NULL,
NULL, "", 0, err);
}
if (!had_err) {
if (join) {
GtFeatureNodeIterator *fni;
GtFeatureNode *child;
bool reverse_strand = false,
first_child = true,
first_child_of_type_seen = false;
GtPhase phase = GT_PHASE_UNDEFINED;
/* in this case we have to traverse the children */
fni = gt_feature_node_iterator_new_direct(gt_feature_node_cast(gn));
while (!had_err && (child = gt_feature_node_iterator_next(fni))) {
if (first_child) {
if (target_ids &&
(target = gt_feature_node_get_attribute(child, GT_GFF_TARGET))) {
gt_str_array_reset(target_ids);
had_err = gt_gff3_parser_parse_all_target_attributes(target, false,
target_ids,
NULL,
NULL, "", 0,
err);
}
first_child = false;
}
if (!had_err) {
if (extract_join_feature((GtGenomeNode*) child, type, region_mapping,
sequence, &reverse_strand,
&first_child_of_type_seen,
&phase, err)) {
had_err = -1;
}
if (phase != GT_PHASE_UNDEFINED) {
phase_offset = (int) phase;
}
}
}
gt_feature_node_iterator_delete(fni);
gt_assert(phase_offset <= (unsigned int) GT_PHASE_UNDEFINED);
if (!had_err && gt_str_length(sequence)) {
if (reverse_strand) {
had_err = gt_reverse_complement(gt_str_get(sequence),
gt_str_length(sequence), err);
}
}
}
else if (gt_feature_node_get_type(fn) == type) {
GtPhase phase = gt_feature_node_get_phase(fn);
gt_assert(!had_err);
if (phase != GT_PHASE_UNDEFINED)
phase_offset = (unsigned int) phase;
/* otherwise we only have to look at this feature */
range = gt_genome_node_get_range(gn);
gt_assert(range.start); /* 1-based coordinates */
had_err = gt_region_mapping_get_sequence(region_mapping, &outsequence,
gt_genome_node_get_seqid(gn),
range.start, range.end, err);
if (!had_err) {
gt_str_append_cstr_nt(sequence, outsequence, gt_range_length(&range));
gt_free(outsequence);
if (gt_feature_node_get_strand(fn) == GT_STRAND_REVERSE) {
had_err = gt_reverse_complement(gt_str_get(sequence),
gt_str_length(sequence), err);
}
}
}
}
if (out_phase_offset && phase_offset != GT_PHASE_UNDEFINED) {
*out_phase_offset = phase_offset;
}
return had_err;
}
static int add_ids_visitor_feature_node(GtNodeVisitor *nv, GtFeatureNode *fn,
GtError *err)
{
AutomaticSequenceRegion *auto_sr;
GtAddIDsVisitor *aiv;
const char *seqid;
bool is_circular;
aiv = add_ids_visitor_cast(nv);
seqid = gt_str_get(gt_genome_node_get_seqid((GtGenomeNode*) fn));
if (aiv->ensure_sorting && !gt_cstr_table_get(aiv->defined_seqids, seqid)) {
gt_error_set(err, "the file %s is not sorted (seqid \"%s\" on line %u has "
"not been previously introduced with a \"%s\" line)",
gt_genome_node_get_filename((GtGenomeNode*) fn), seqid,
gt_genome_node_get_line_number((GtGenomeNode*) fn),
GT_GFF_SEQUENCE_REGION);
return -1;
}
if (!gt_cstr_table_get(aiv->defined_seqids, seqid)) {
GtFeatureNodeIterator *fni;
GtFeatureNode *node;
GtRange range = gt_genome_node_get_range((GtGenomeNode*) fn);
is_circular = gt_feature_node_get_attribute(fn, GT_GFF_IS_CIRCULAR)
? true : false;
if (!is_circular) {
fni = gt_feature_node_iterator_new(fn);
while ((node = gt_feature_node_iterator_next(fni))) {
GtRange node_range = gt_genome_node_get_range((GtGenomeNode*) node);
range = gt_range_join(&range, &node_range);
}
gt_feature_node_iterator_delete(fni);
}
/* sequence region has not been previously introduced -> check if one has
already been created automatically */
auto_sr = gt_hashmap_get(aiv->undefined_sequence_regions, seqid);
if (!auto_sr) {
GtStr *seqid_str;
/* sequence region has not been createad automatically -> do it now */
gt_warning("seqid \"%s\" on line %u in file \"%s\" has not been "
"previously introduced with a \"%s\" line, create such a line "
"automatically", seqid,
gt_genome_node_get_line_number((GtGenomeNode*) fn),
gt_genome_node_get_filename((GtGenomeNode*) fn),
GT_GFF_SEQUENCE_REGION);
auto_sr = automatic_sequence_region_new(is_circular);
seqid_str = gt_genome_node_get_seqid((GtGenomeNode*) fn);
auto_sr->sequence_region = gt_region_node_new(seqid_str, range.start,
range.end);
gt_hashmap_add(aiv->undefined_sequence_regions, gt_str_get(seqid_str),
auto_sr);
}
else {
if (auto_sr->is_circular) {
gt_assert(!is_circular); /* XXX */
}
else if (is_circular) {
gt_assert(!auto_sr->is_circular); /* XXX */
auto_sr->is_circular = true;
gt_genome_node_set_range(auto_sr->sequence_region, &range);
}
else {
GtRange joined_range,
sr_range = gt_genome_node_get_range(auto_sr->sequence_region);
/* update the range of the sequence region */
joined_range = gt_range_join(&range, &sr_range);
gt_genome_node_set_range(auto_sr->sequence_region, &joined_range);
}
}
gt_array_add(auto_sr->feature_nodes, fn);
}
else
gt_queue_add(aiv->node_buffer, fn);
return 0;
}
static int snp_annotator_classify_snp(GtSNPAnnotatorVisitor *sav,
GtFeatureNode *mRNA,
GtFeatureNode *snp,
GtUword variant_pos,
GtUword variant_idx,
char variant_char,
#ifndef NDEBUG
GT_UNUSED char reference_char,
#endif
GT_UNUSED GtError *err)
{
int had_err = 0;
char *mrnaseq;
const char *variant_effect = NULL;
gt_assert(mRNA && snp && sav);
gt_log_log("processing variant char %c for SNP %s\n",
variant_char, gt_feature_node_get_attribute(snp, "Dbxref"));
mrnaseq = gt_hashmap_get(sav->rnaseqs, mRNA);
gt_assert(mrnaseq);
if (mrnaseq) {
char codon[3],
variant_codon[3];
GtStr *effect_string;
char oldamino,
newamino;
GT_UNUSED GtUword mrnalen;
GtUword startpos = variant_pos / GT_CODON_LENGTH,
variantoffset = variant_pos % GT_CODON_LENGTH;
mrnalen = strlen(mrnaseq);
gt_assert(variant_pos < mrnalen);
variant_codon[0] = codon[0] = mrnaseq[3*startpos];
variant_codon[1] = codon[1] = mrnaseq[3*startpos+1];
variant_codon[2] = codon[2] = mrnaseq[3*startpos+2];
variant_codon[variantoffset] = variant_char;
#ifndef NDEBUG
gt_assert(toupper(codon[variantoffset]) == toupper(reference_char));
#endif
if (gt_trans_table_is_stop_codon(sav->tt, codon[0], codon[1], codon[2])) {
if (gt_trans_table_is_stop_codon(sav->tt, variant_codon[0],
variant_codon[1], variant_codon[2])) {
variant_effect = gt_symbol(GT_SNP_SYNONYMOUS_STOP_EFFECT);
} else {
variant_effect = gt_symbol(GT_SNP_STOP_LOST_EFFECT);
}
} else {
if (gt_trans_table_is_stop_codon(sav->tt, variant_codon[0],
variant_codon[1], variant_codon[2])) {
variant_effect = gt_symbol(GT_SNP_NONSENSE_EFFECT);
} else {
had_err = gt_trans_table_translate_codon(sav->tt, codon[0], codon[1],
codon[2], &oldamino, err);
if (!had_err) {
had_err = gt_trans_table_translate_codon(sav->tt, variant_codon[0],
variant_codon[1],
variant_codon[2],
&newamino, err);
}
if (!had_err) {
if (newamino == oldamino) {
variant_effect = gt_symbol(GT_SNP_SYNONYMOUS_AMINO_EFFECT);
} else {
variant_effect = gt_symbol(GT_SNP_MISSENSE_EFFECT);
}
}
}
}
if (!had_err) {
const char *var_attrib;
gt_assert(variant_effect != NULL);
if ((var_attrib = gt_feature_node_get_attribute(snp,
GT_GVF_VARIANT_EFFECT))) {
effect_string = gt_str_new_cstr(var_attrib);
gt_str_append_cstr(effect_string, ",");
gt_str_append_cstr(effect_string, variant_effect);
} else {
effect_string = gt_str_new_cstr(variant_effect);
}
gt_str_append_cstr(effect_string, " ");
gt_str_append_ulong(effect_string, variant_idx);
gt_str_append_cstr(effect_string, " ");
gt_str_append_cstr(effect_string, gt_feature_node_get_type(mRNA));
gt_str_append_cstr(effect_string, " ");
gt_str_append_cstr(effect_string,
gt_feature_node_get_attribute(mRNA, GT_GFF_ID));
gt_feature_node_set_attribute(snp, GT_GVF_VARIANT_EFFECT,
gt_str_get(effect_string));
gt_str_reset(effect_string);
gt_str_delete(effect_string);
}
}
return had_err;
}
static int snp_annotator_visitor_feature_node(GtNodeVisitor *nv,
GtFeatureNode *fn,
GtError *err)
{
int had_err = 0;
GtSNPAnnotatorVisitor *sav;
GtFeatureNodeIterator *fni,
*mrnafni;
GtFeatureNode *curnode,
*curnode2;
GtRange snp_rng;
gt_error_check(err);
sav = snp_annotator_visitor_cast(nv);
/* ignore non-nodes */
if (!fn) return 0;
/* only process SNPs */
if (!(gt_feature_node_get_type(fn) == sav->SNV_type ||
gt_feature_node_get_type(fn) == sav->SNP_type)) {
return 0;
}
fni = gt_feature_node_iterator_new_direct(sav->gene);
snp_rng = gt_genome_node_get_range((GtGenomeNode*) fn);
while (!had_err && (curnode = gt_feature_node_iterator_next(fni))) {
if (gt_feature_node_get_type(curnode) == sav->mRNA_type) {
GtStrand mrna_strand = gt_feature_node_get_strand(curnode);
#ifndef NDEBUG
const char *refstr;
#endif
GtUword mrnasnppos = 0;
mrnafni = gt_feature_node_iterator_new(curnode);
while (!had_err && (curnode2 = gt_feature_node_iterator_next(mrnafni))) {
if (gt_feature_node_get_type(curnode2) == sav->CDS_type) {
GtRange cds_rng = gt_genome_node_get_range((GtGenomeNode*) curnode2);
if (gt_range_overlap(&snp_rng, &cds_rng)) {
char *mRNA,
origchar;
char *variantchars, *variantptr = NULL;
GT_UNUSED char *refchars, *refptr = NULL;
mRNA = (char*) gt_hashmap_get(sav->rnaseqs, curnode);
gt_assert(mRNA);
gt_assert(snp_rng.start >= cds_rng.start);
mrnasnppos += (snp_rng.start - cds_rng.start);
if (mrna_strand == GT_STRAND_REVERSE)
mrnasnppos = strlen(mRNA) - mrnasnppos - 1;
gt_assert(mrnasnppos < strlen(mRNA));
origchar = mRNA[mrnasnppos];
#ifndef NDEBUG
refstr = refptr = gt_cstr_dup(gt_feature_node_get_attribute(fn,
GT_GVF_REFERENCE_SEQ));
if (!had_err && refstr) {
if (gt_feature_node_get_strand(curnode) == GT_STRAND_REVERSE) {
int rval = gt_complement(&origchar, origchar, err);
gt_assert(rval == 0);
}
gt_assert(toupper(origchar) == toupper(refstr[0]));
}
#endif
variantchars = variantptr = gt_cstr_dup(
gt_feature_node_get_attribute(fn, GT_GVF_VARIANT_SEQ));
if (!had_err && variantchars) {
GtUword i = 0;
while (!had_err &&
(*variantchars != ';' && *variantchars != '\0')) {
if (*variantchars != ',' && *variantchars != origchar) {
char variantchar = *variantchars;
#ifndef NDEBUG
char refchar = refstr ? refstr[0] : '-'; /* XXX */
if (!had_err && mrna_strand == GT_STRAND_REVERSE)
had_err = gt_complement(&refchar, refchar, err);
#endif
if (!had_err && mrna_strand == GT_STRAND_REVERSE)
had_err = gt_complement(&variantchar, variantchar, err);
if (!had_err) {
had_err = snp_annotator_classify_snp(sav, curnode, fn,
mrnasnppos,
i++,
variantchar,
#ifndef NDEBUG
refchar,
#endif
err);
}
} else if (*variantchars == origchar) {
i++;
}
variantchars++;
}
gt_free(variantptr);
gt_free(refptr);
}
} else {
mrnasnppos += gt_range_length(&cds_rng);
}
}
}
gt_feature_node_iterator_delete(mrnafni);
}
}
gt_feature_node_iterator_delete(fni);
return had_err;
}
int gt_ltrfileout_stream_next(GtNodeStream *ns, GtGenomeNode **gn, GtError *err)
{
GtLTRdigestFileOutStream *ls;
GtFeatureNode *fn;
GtRange lltr_rng = {GT_UNDEF_UWORD, GT_UNDEF_UWORD},
rltr_rng = {GT_UNDEF_UWORD, GT_UNDEF_UWORD},
ppt_rng = {GT_UNDEF_UWORD, GT_UNDEF_UWORD},
pbs_rng = {GT_UNDEF_UWORD, GT_UNDEF_UWORD};
int had_err;
GtUword i=0;
gt_error_check(err);
ls = gt_ltrdigest_file_out_stream_cast(ns);
/* initialize this element */
memset(&ls->element, 0, sizeof (GtLTRElement));
/* get annotations from parser */
had_err = gt_node_stream_next(ls->in_stream, gn, err);
if (!had_err && *gn)
{
GtFeatureNodeIterator* gni;
GtFeatureNode *mygn;
/* only process feature nodes */
if (!(fn = gt_feature_node_try_cast(*gn)))
return 0;
ls->element.pdomorder = gt_array_new(sizeof (const char*));
/* fill LTRElement structure from GFF3 subgraph */
gni = gt_feature_node_iterator_new(fn);
for (mygn = fn; mygn != NULL; mygn = gt_feature_node_iterator_next(gni))
(void) gt_genome_node_accept((GtGenomeNode*) mygn,
(GtNodeVisitor*) ls->lv,
err);
gt_feature_node_iterator_delete(gni);
}
if (!had_err && ls->element.mainnode != NULL)
{
char desc[GT_MAXFASTAHEADER];
GtFeatureNode *ltr3, *ltr5;
GtStr *sdesc, *sreg, *seq;
/* find sequence in GtEncseq */
sreg = gt_genome_node_get_seqid((GtGenomeNode*) ls->element.mainnode);
sdesc = gt_str_new();
had_err = gt_region_mapping_get_description(ls->rmap, sdesc, sreg, err);
if (!had_err) {
GtRange rng;
ls->element.seqid = gt_calloc((size_t) ls->seqnamelen+1, sizeof (char));
(void) snprintf(ls->element.seqid,
MIN((size_t) gt_str_length(sdesc),
(size_t) ls->seqnamelen)+1,
"%s", gt_str_get(sdesc));
gt_cstr_rep(ls->element.seqid, ' ', '_');
if (gt_str_length(sdesc) > (GtUword) ls->seqnamelen)
ls->element.seqid[ls->seqnamelen] = '\0';
(void) gt_ltrelement_format_description(&ls->element,
ls->seqnamelen,
desc,
(size_t) (GT_MAXFASTAHEADER-1));
gt_str_delete(sdesc);
/* output basic retrotransposon data */
lltr_rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.leftLTR);
rltr_rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.rightLTR);
rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.mainnode);
gt_file_xprintf(ls->tabout_file,
GT_WU"\t"GT_WU"\t"GT_WU"\t%s\t"GT_WU"\t"GT_WU"\t"GT_WU"\t"
GT_WU"\t"GT_WU"\t"GT_WU"\t",
rng.start, rng.end, gt_ltrelement_length(&ls->element),
ls->element.seqid, lltr_rng.start, lltr_rng.end,
gt_ltrelement_leftltrlen(&ls->element), rltr_rng.start,
rltr_rng.end, gt_ltrelement_rightltrlen(&ls->element));
}
seq = gt_str_new();
/* output TSDs */
if (!had_err && ls->element.leftTSD != NULL)
{
GtRange tsd_rng;
tsd_rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.leftTSD);
had_err = gt_extract_feature_sequence(seq,
(GtGenomeNode*) ls->element.leftTSD,
gt_symbol(gt_ft_target_site_duplication),
false,
NULL, NULL, ls->rmap, err);
if (!had_err) {
gt_file_xprintf(ls->tabout_file,
""GT_WU"\t"GT_WU"\t%s\t",
tsd_rng.start,
tsd_rng.end,
gt_str_get(seq));
}
gt_str_reset(seq);
} else gt_file_xprintf(ls->tabout_file, "\t\t\t");
if (!had_err && ls->element.rightTSD != NULL)
{
GtRange tsd_rng;
tsd_rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.rightTSD);
had_err = gt_extract_feature_sequence(seq,
(GtGenomeNode*) ls->element.rightTSD,
gt_symbol(gt_ft_target_site_duplication),
false,
NULL, NULL, ls->rmap, err);
if (!had_err) {
gt_file_xprintf(ls->tabout_file,
""GT_WU"\t"GT_WU"\t%s\t",
tsd_rng.start,
tsd_rng.end,
gt_str_get(seq));
}
gt_str_reset(seq);
} else gt_file_xprintf(ls->tabout_file, "\t\t\t");
/* output PPT */
if (!had_err && ls->element.ppt != NULL)
{
GtStrand ppt_strand = gt_feature_node_get_strand(ls->element.ppt);
ppt_rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.ppt);
had_err = gt_extract_feature_sequence(seq,
(GtGenomeNode*) ls->element.ppt,
gt_symbol(gt_ft_RR_tract), false,
NULL, NULL, ls->rmap, err);
if (!had_err) {
gt_fasta_show_entry(desc, gt_str_get(seq), gt_range_length(&ppt_rng),
GT_FSWIDTH, ls->pptout_file);
gt_file_xprintf(ls->tabout_file,
""GT_WU"\t"GT_WU"\t%s\t%c\t%d\t",
ppt_rng.start,
ppt_rng.end,
gt_str_get(seq),
GT_STRAND_CHARS[ppt_strand],
(ppt_strand == GT_STRAND_FORWARD ?
abs((int) (rltr_rng.start - ppt_rng.end)) :
abs((int) (lltr_rng.end - ppt_rng.start))));
}
gt_str_reset(seq);
} else gt_file_xprintf(ls->tabout_file, "\t\t\t\t\t");
/* output PBS */
if (!had_err && ls->element.pbs != NULL)
{
GtStrand pbs_strand;
pbs_strand = gt_feature_node_get_strand(ls->element.pbs);
pbs_rng = gt_genome_node_get_range((GtGenomeNode*) ls->element.pbs);
had_err = gt_extract_feature_sequence(seq,
(GtGenomeNode*) ls->element.pbs,
gt_symbol(gt_ft_primer_binding_site),
false, NULL, NULL, ls->rmap, err);
if (!had_err) {
gt_fasta_show_entry(desc, gt_str_get(seq), gt_range_length(&pbs_rng),
GT_FSWIDTH, ls->pbsout_file);
gt_file_xprintf(ls->tabout_file,
""GT_WU"\t"GT_WU"\t%c\t%s\t%s\t%s\t%s\t%s\t",
pbs_rng.start,
pbs_rng.end,
GT_STRAND_CHARS[pbs_strand],
gt_feature_node_get_attribute(ls->element.pbs, "trna"),
gt_str_get(seq),
gt_feature_node_get_attribute(ls->element.pbs,
"pbsoffset"),
gt_feature_node_get_attribute(ls->element.pbs,
"trnaoffset"),
gt_feature_node_get_attribute(ls->element.pbs,
"edist"));
}
gt_str_reset(seq);
} else gt_file_xprintf(ls->tabout_file, "\t\t\t\t\t\t\t\t");
/* output protein domains */
if (!had_err && ls->element.pdoms != NULL)
{
GtStr *pdomorderstr = gt_str_new();
for (i=0; !had_err && i<gt_array_size(ls->element.pdomorder); i++)
{
const char* key = *(const char**) gt_array_get(ls->element.pdomorder,
i);
GtArray *entry = (GtArray*) gt_hashmap_get(ls->element.pdoms, key);
had_err = write_pdom(ls, entry, key, ls->rmap, desc, err);
}
if (GT_STRAND_REVERSE == gt_feature_node_get_strand(ls->element.mainnode))
gt_array_reverse(ls->element.pdomorder);
for (i=0 ;!had_err && i<gt_array_size(ls->element.pdomorder); i++)
{
const char* name = *(const char**) gt_array_get(ls->element.pdomorder,
i);
gt_str_append_cstr(pdomorderstr, name);
if (i != gt_array_size(ls->element.pdomorder)-1)
gt_str_append_cstr(pdomorderstr, "/");
}
gt_file_xprintf(ls->tabout_file, "%s", gt_str_get(pdomorderstr));
gt_str_delete(pdomorderstr);
}
/* output LTRs (we just expect them to exist) */
switch (gt_feature_node_get_strand(ls->element.mainnode))
{
case GT_STRAND_REVERSE:
ltr5 = ls->element.rightLTR;
ltr3 = ls->element.leftLTR;
break;
case GT_STRAND_FORWARD:
default:
ltr5 = ls->element.leftLTR;
ltr3 = ls->element.rightLTR;
break;
}
if (!had_err) {
had_err = gt_extract_feature_sequence(seq, (GtGenomeNode*) ltr5,
gt_symbol(gt_ft_long_terminal_repeat),
false,
NULL, NULL, ls->rmap, err);
}
if (!had_err) {
gt_fasta_show_entry(desc, gt_str_get(seq), gt_str_length(seq),
GT_FSWIDTH, ls->ltr5out_file);
gt_str_reset(seq);
}
if (!had_err) {
had_err = gt_extract_feature_sequence(seq, (GtGenomeNode*) ltr3,
gt_symbol(gt_ft_long_terminal_repeat),
false,
NULL, NULL, ls->rmap, err);
}
if (!had_err) {
gt_fasta_show_entry(desc, gt_str_get(seq), gt_str_length(seq),
GT_FSWIDTH, ls->ltr3out_file);
gt_str_reset(seq);
}
/* output complete oriented element */
if (!had_err) {
had_err = gt_extract_feature_sequence(seq,
(GtGenomeNode*) ls->element.mainnode,
gt_symbol(gt_ft_LTR_retrotransposon),
false,
NULL, NULL, ls->rmap, err);
}
if (!had_err) {
gt_fasta_show_entry(desc,gt_str_get(seq), gt_str_length(seq),
GT_FSWIDTH, ls->elemout_file);
gt_str_reset(seq);
}
gt_file_xprintf(ls->tabout_file, "\n");
gt_str_delete(seq);
}
gt_hashmap_delete(ls->element.pdoms);
gt_array_delete(ls->element.pdomorder);
gt_free(ls->element.seqid);
return had_err;
}
static int gt_ltr_visitor_feature_node(GtNodeVisitor *nv, GtFeatureNode *fn,
GT_UNUSED GtError *err)
{
GtLTRVisitor *lv;
GtRange node_range;
GtArray *pdomarr = NULL;
const char *pfamname;
const char *fnt;
lv = gt_ltr_visitor_cast(nv);
gt_assert(lv);
gt_error_check(err);
fnt = gt_feature_node_get_type(fn);
if (strcmp(fnt, gt_ft_LTR_retrotransposon) == 0)
{
lv->element->mainnode = fn;
} else if (strcmp(fnt, gt_ft_long_terminal_repeat) == 0)
{
if (lv->element->leftLTR == NULL)
{
node_range = gt_genome_node_get_range((GtGenomeNode*) fn);
lv->element->leftLTR = fn;
/* compensate for 1-based node coords */
lv->element->leftLTR_5 = node_range.start - 1;
lv->element->leftLTR_3 = node_range.end - 1;
}
else
{
node_range = gt_genome_node_get_range((GtGenomeNode*) fn);
lv->element->rightLTR = fn;
/* compensate for 1-based node coords */
lv->element->rightLTR_5 = node_range.start - 1;
lv->element->rightLTR_3 = node_range.end - 1;
}
} else if (strcmp(fnt, gt_ft_target_site_duplication) == 0)
{
if (lv->element->leftTSD == NULL)
{
lv->element->leftTSD = fn;
}
else
{
lv->element->rightTSD = fn;
}
} else if (strcmp(fnt, gt_ft_RR_tract) == 0)
{
if (lv->element->ppt == NULL)
{
lv->element->ppt = fn;
}
} else if (strcmp(fnt, gt_ft_primer_binding_site) == 0)
{
if (lv->element->pbs == NULL)
{
lv->element->pbs = fn;
}
} else if (strcmp(fnt, gt_ft_protein_match) == 0)
{
char buf[BUFSIZ];
if (!lv->element->pdoms)
{
lv->element->pdoms = gt_hashmap_new(GT_HASH_STRING, gt_free_func,
(GtFree) gt_array_delete);
}
pfamname = gt_feature_node_get_attribute(fn, "name");
(void) snprintf(buf, BUFSIZ-1, "%s", pfamname);
gt_cstr_rep(buf, '/', '_');
if (!(pdomarr = (GtArray*) gt_hashmap_get(lv->element->pdoms, buf)))
{
char *pfamcpy = gt_cstr_dup(buf);
pdomarr = gt_array_new(sizeof (GtFeatureNode*));
gt_hashmap_add(lv->element->pdoms, pfamcpy, pdomarr);
if (lv->element->pdomorder != NULL)
gt_array_add(lv->element->pdomorder, pfamcpy);
}
gt_array_add(pdomarr, fn);
}
return 0;
}
int gt_gtf_parser_parse(GtGTFParser *parser, GtQueue *genome_nodes,
GtStr *filenamestr, GtFile *fpin, bool be_tolerant,
GtError *err)
{
GtStr *seqid_str, *source_str, *line_buffer;
char *line;
size_t line_length;
GtUword i, line_number = 0;
GtGenomeNode *gn;
GtRange range;
GtPhase phase_value;
GtStrand gt_strand_value;
GtSplitter *splitter, *attribute_splitter;
float score_value;
char *seqname,
*source,
*feature,
*start,
*end,
*score,
*strand,
*frame,
*attributes,
*token,
*gene_id,
*gene_name = NULL,
*transcript_id,
*transcript_name = NULL,
**tokens;
GtHashmap *transcript_id_hash; /* map from transcript id to array of genome
nodes */
GtArray *gt_genome_node_array;
ConstructionInfo cinfo;
GTF_feature_type gtf_feature_type;
GT_UNUSED bool gff_type_is_valid = false;
const char *type = NULL;
const char *filename;
bool score_is_defined;
int had_err = 0;
gt_assert(parser && genome_nodes);
gt_error_check(err);
filename = gt_str_get(filenamestr);
/* alloc */
line_buffer = gt_str_new();
splitter = gt_splitter_new(),
attribute_splitter = gt_splitter_new();
#define HANDLE_ERROR \
if (had_err) { \
if (be_tolerant) { \
fprintf(stderr, "skipping line: %s\n", gt_error_get(err)); \
gt_error_unset(err); \
gt_str_reset(line_buffer); \
had_err = 0; \
continue; \
} \
else { \
had_err = -1; \
break; \
} \
}
while (gt_str_read_next_line_generic(line_buffer, fpin) != EOF) {
line = gt_str_get(line_buffer);
line_length = gt_str_length(line_buffer);
line_number++;
gene_name = gene_id = transcript_id = transcript_name = NULL;
had_err = 0;
if (line_length == 0) {
gt_warning("skipping blank line " GT_WU " in file \"%s\"", line_number,
filename);
}
else if (line[0] == '#') {
/* storing comment */
if (line_length >= 2 && line[1] == '#')
gn = gt_comment_node_new(line+2); /* store '##' line as '#' line */
else
gn = gt_comment_node_new(line+1);
gt_genome_node_set_origin(gn, filenamestr, line_number);
gt_queue_add(genome_nodes, gn);
}
else {
bool stop_codon = false;
char *tokendup, *attrkey;
GtStrArray *attrkeys, *attrvals;
/* process tab delimited GTF line */
gt_splitter_reset(splitter);
gt_splitter_split(splitter, line, line_length, '\t');
if (gt_splitter_size(splitter) != 9UL) {
gt_error_set(err, "line " GT_WU " in file \"%s\" contains " GT_WU
" tab (\\t) " "separated fields instead of 9", line_number,
filename,
gt_splitter_size(splitter));
had_err = -1;
break;
}
tokens = gt_splitter_get_tokens(splitter);
seqname = tokens[0];
source = tokens[1];
feature = tokens[2];
start = tokens[3];
end = tokens[4];
score = tokens[5];
strand = tokens[6];
frame = tokens[7];
attributes = tokens[8];
/* parse feature */
if (GTF_feature_type_get(>f_feature_type, feature) == -1) {
/* we skip unknown features */
fprintf(stderr, "skipping line " GT_WU " in file \"%s\": unknown "
"feature: \"%s\"\n", line_number, filename, feature);
gt_str_reset(line_buffer);
continue;
}
/* translate into GFF3 feature type */
switch (gtf_feature_type) {
case GTF_stop_codon:
stop_codon = true;
case GTF_CDS:
gff_type_is_valid = gt_type_checker_is_valid(parser->type_checker,
gt_ft_CDS);
type = gt_ft_CDS;
break;
case GTF_exon:
gff_type_is_valid = gt_type_checker_is_valid(parser->type_checker,
gt_ft_exon);
type = gt_ft_exon;
break;
case GTF_start_codon:
/* we can skip the start codons, they are part of the CDS anyway */
gt_str_reset(line_buffer);
continue;
}
gt_assert(gff_type_is_valid);
/* parse the range */
had_err = gt_parse_range(&range, start, end, line_number, filename, err);
HANDLE_ERROR;
/* process seqname (we have to do it here because we need the range) */
gt_region_node_builder_add_region(parser->region_node_builder, seqname,
range);
/* parse the score */
had_err = gt_parse_score(&score_is_defined, &score_value, score,
line_number, filename, err);
HANDLE_ERROR;
/* parse the strand */
had_err = gt_parse_strand(>_strand_value, strand, line_number, filename,
err);
HANDLE_ERROR;
/* parse the frame */
had_err = gt_parse_phase(&phase_value, frame, line_number, filename, err);
HANDLE_ERROR;
/* parse the attributes */
attrkeys = gt_str_array_new();
attrvals = gt_str_array_new();
gt_splitter_reset(attribute_splitter);
gene_id = NULL;
transcript_id = NULL;
gt_splitter_split(attribute_splitter, attributes, strlen(attributes),
';');
for (i = 0; i < gt_splitter_size(attribute_splitter); i++) {
token = gt_splitter_get_token(attribute_splitter, i);
/* skip leading blanks */
while (*token == ' ')
token++;
tokendup = gt_cstr_dup(token);
attrkey = strtok(tokendup, " ");
if (attrkey) {
char *attrval = strtok(NULL, " ");
if (attrval == NULL || strcmp(attrval, "") == 0 ||
strcmp(attrval, "\"\"") == 0)
{
gt_error_set(err, "missing value to attribute \"%s\" on line "
GT_WU " in file \"%s\"", attrkey,line_number,filename);
had_err = -1;
}
HANDLE_ERROR;
if (*attrval == '"')
attrval++;
if (attrval[strlen(attrval)-1] == '"')
attrval[strlen(attrval)-1] = '\0';
gt_assert(attrkey && strlen(attrkey) > 0);
gt_assert(attrval && strlen(attrval) > 0);
gt_str_array_add_cstr(attrkeys, attrkey);
gt_str_array_add_cstr(attrvals, attrval);
}
gt_free(tokendup);
/* look for the two mandatory attributes */
if (strncmp(token, GENE_ID_ATTRIBUTE, strlen(GENE_ID_ATTRIBUTE)) == 0) {
if (strlen(token) + 2 < strlen(GENE_ID_ATTRIBUTE)) {
gt_error_set(err, "missing value to attribute \"%s\" on line "
GT_WU "in file \"%s\"", GENE_ID_ATTRIBUTE, line_number,
filename);
had_err = -1;
}
HANDLE_ERROR;
gene_id = token + strlen(GENE_ID_ATTRIBUTE) + 1;
if (*gene_id == '"')
gene_id++;
if (gene_id[strlen(gene_id)-1] == '"')
gene_id[strlen(gene_id)-1] = '\0';
}
else if (strncmp(token, TRANSCRIPT_ID_ATTRIBUTE,
strlen(TRANSCRIPT_ID_ATTRIBUTE)) == 0) {
if (strlen(token) + 2 < strlen(TRANSCRIPT_ID_ATTRIBUTE)) {
gt_error_set(err, "missing value to attribute \"%s\" on line "
GT_WU "in file \"%s\"", TRANSCRIPT_ID_ATTRIBUTE,
line_number, filename);
had_err = -1;
}
HANDLE_ERROR;
transcript_id = token + strlen(TRANSCRIPT_ID_ATTRIBUTE) + 1;
if (*transcript_id == '"')
transcript_id++;
if (transcript_id[strlen(transcript_id)-1] == '"')
transcript_id[strlen(transcript_id)-1] = '\0';
}
else if (strncmp(token, GENE_NAME_ATTRIBUTE,
strlen(GENE_NAME_ATTRIBUTE)) == 0) {
if (strlen(token) + 2 < strlen(GENE_NAME_ATTRIBUTE)) {
gt_error_set(err, "missing value to attribute \"%s\" on line "
GT_WU "in file \"%s\"", GENE_NAME_ATTRIBUTE,
line_number, filename);
had_err = -1;
}
HANDLE_ERROR;
gene_name = token + strlen(GENE_NAME_ATTRIBUTE) + 1;
/* for output we want to strip quotes */
if (*gene_name == '"')
gene_name++;
if (gene_name[strlen(gene_name)-1] == '"')
gene_name[strlen(gene_name)-1] = '\0';
}
else if (strncmp(token, TRANSCRIPT_NAME_ATTRIBUTE,
strlen(TRANSCRIPT_NAME_ATTRIBUTE)) == 0) {
if (strlen(token) + 2 < strlen(TRANSCRIPT_NAME_ATTRIBUTE)) {
gt_error_set(err, "missing value to attribute \"%s\" on line "
GT_WU "in file \"%s\"", TRANSCRIPT_NAME_ATTRIBUTE,
line_number, filename);
had_err = -1;
}
HANDLE_ERROR;
transcript_name = token + strlen(TRANSCRIPT_NAME_ATTRIBUTE) + 1;
/* for output we want to strip quotes */
if (*transcript_name == '"')
transcript_name++;
if (transcript_name[strlen(transcript_name)-1] == '"')
transcript_name[strlen(transcript_name)-1] = '\0';
}
}
/* check for the mandatory attributes */
if (!gene_id) {
gt_error_set(err, "missing attribute \"%s\" on line " GT_WU
" in file \"%s\"", GENE_ID_ATTRIBUTE, line_number,
filename);
had_err = -1;
}
HANDLE_ERROR;
if (!transcript_id) {
gt_error_set(err, "missing attribute \"%s\" on line " GT_WU
" in file \"%s\"", TRANSCRIPT_ID_ATTRIBUTE, line_number,
filename);
had_err = -1;
}
HANDLE_ERROR;
/* process the mandatory attributes */
if (!(transcript_id_hash = gt_hashmap_get(parser->gene_id_hash,
gene_id))) {
transcript_id_hash = gt_hashmap_new(GT_HASH_STRING, gt_free_func,
(GtFree) gt_array_delete);
gt_hashmap_add(parser->gene_id_hash, gt_cstr_dup(gene_id),
transcript_id_hash);
}
gt_assert(transcript_id_hash);
if (!(gt_genome_node_array = gt_hashmap_get(transcript_id_hash,
transcript_id))) {
gt_genome_node_array = gt_array_new(sizeof (GtGenomeNode*));
gt_hashmap_add(transcript_id_hash, gt_cstr_dup(transcript_id),
gt_genome_node_array);
}
gt_assert(gt_genome_node_array);
/* save optional gene_name and transcript_name attributes */
if (transcript_name && strlen(transcript_name) > 0
&& !gt_hashmap_get(parser->transcript_id_to_name_mapping,
transcript_id)) {
gt_hashmap_add(parser->transcript_id_to_name_mapping,
gt_cstr_dup(transcript_id),
gt_cstr_dup(transcript_name));
}
if (gene_name && strlen(gene_name) > 0
&& !gt_hashmap_get(parser->gene_id_to_name_mapping,
gene_id)) {
gt_hashmap_add(parser->gene_id_to_name_mapping,
gt_cstr_dup(gene_id),
gt_cstr_dup(gene_name));
}
/* get seqid */
seqid_str = gt_hashmap_get(parser->seqid_to_str_mapping, seqname);
if (!seqid_str) {
seqid_str = gt_str_new_cstr(seqname);
gt_hashmap_add(parser->seqid_to_str_mapping, gt_str_get(seqid_str),
seqid_str);
}
gt_assert(seqid_str);
/* construct the new feature */
gn = gt_feature_node_new(seqid_str, type, range.start, range.end,
gt_strand_value);
gt_genome_node_set_origin(gn, filenamestr, line_number);
if (stop_codon) {
gt_feature_node_add_attribute((GtFeatureNode*) gn,
GTF_PARSER_STOP_CODON_FLAG, "true");
}
for (i = 0; i < gt_str_array_size(attrkeys); i++) {
GtFeatureNode *fn = (GtFeatureNode *)gn;
const char *key = gt_str_array_get(attrkeys, i);
const char *val = gt_str_array_get(attrvals, i);
/* Not a comprehensive solution to ensure correct encoding, just bare
minimum required to get Cufflinks output parsed */
if (strcmp(val, "=") == 0)
val = "%26";
if (gt_feature_node_get_attribute(fn, key) != NULL) {
const char *oldval = gt_feature_node_get_attribute(fn, key);
GtStr *newval = gt_str_new_cstr(oldval);
gt_str_append_char(newval, ',');
gt_str_append_cstr(newval, val);
gt_feature_node_set_attribute(fn, key, gt_str_get(newval));
gt_str_delete(newval);
}
else
gt_feature_node_add_attribute(fn, key, val);
}
gt_str_array_delete(attrkeys);
gt_str_array_delete(attrvals);
/* set source */
source_str = gt_hashmap_get(parser->source_to_str_mapping, source);
if (!source_str) {
source_str = gt_str_new_cstr(source);
gt_hashmap_add(parser->source_to_str_mapping, gt_str_get(source_str),
source_str);
}
gt_assert(source_str);
gt_feature_node_set_source((GtFeatureNode*) gn, source_str);
if (score_is_defined)
gt_feature_node_set_score((GtFeatureNode*) gn, score_value);
if (phase_value != GT_PHASE_UNDEFINED)
gt_feature_node_set_phase((GtFeatureNode*) gn, phase_value);
gt_array_add(gt_genome_node_array, gn);
}
gt_str_reset(line_buffer);
}
/* process all region nodes */
if (!had_err)
gt_region_node_builder_build(parser->region_node_builder, genome_nodes);
/* process all feature nodes */
cinfo.genome_nodes = genome_nodes;
cinfo.tidy = be_tolerant;
cinfo.gene_id_to_name_mapping = parser->gene_id_to_name_mapping;
cinfo.transcript_id_to_name_mapping = parser->transcript_id_to_name_mapping;
if (!had_err) {
had_err = gt_hashmap_foreach(parser->gene_id_hash, construct_genes,
&cinfo, err);
}
gt_hashmap_foreach(parser->gene_id_hash, delete_genes, NULL, err);
/* free */
gt_splitter_delete(splitter);
gt_splitter_delete(attribute_splitter);
gt_str_delete(line_buffer);
return had_err;
}
static int CpGI_score_stream_next(GtNodeStream * ns,
GtGenomeNode ** gn,
GtError * err)
{
GtGenomeNode * cur_node;
int err_num = 0;
*gn = NULL;
CpGI_score_stream * score_stream;
unsigned long island_start;
unsigned long island_end;
float island_score;
int chromosome_num;
GtStr * seqID_gtstr;
char * seqID_str;
char * num_cg_str;
unsigned long num_cg = 0;
score_stream = CpGI_score_stream_cast(ns);
// find the CpGI's, process methylome score
if(!gt_node_stream_next(score_stream->in_stream,
&cur_node,
err
) && cur_node != NULL
)
{
*gn = cur_node;
// try casting as a feature node so we can test type
if(!gt_genome_node_try_cast(gt_feature_node_class(), cur_node))
{
return 0;
}
else // we found a feature node
{
if(!gt_feature_node_has_type(cur_node, feature_type_CpGI))
return 0;
#if DEBUG_SCORE
printf("found CpGI\n");
#endif
island_start = gt_genome_node_get_start(cur_node);
island_end = gt_genome_node_get_end(cur_node);
seqID_gtstr = gt_genome_node_get_seqid(cur_node);
seqID_str = gt_str_get(seqID_gtstr);
sscanf(seqID_str, "Chr%d", &chromosome_num);
num_cg_str = gt_feature_node_get_attribute(cur_node, "sumcg");
if (!num_cg_str)
return 0;
sscanf(num_cg_str, "%d", &num_cg);
// now figure out the score
island_score = CpGI_score_stream_score_island(score_stream ,
chromosome_num,
num_cg,
island_start,
island_end);
// gt_str_delete(seqID_gtstr);
// save the score into the node
gt_feature_node_set_score(cur_node, island_score);
return 0;
}
}
return err_num;
}
static int CpGIOverlap_stream_next(GtNodeStream * ns,
GtGenomeNode ** gn,
GtError * err)
{
GtGenomeNode * cur_node, * next_node;
GtFeatureNodeIterator * iter;
int err_num = 0;
*gn = NULL;
CpGIOverlap_stream * context;
const char * gene_name = NULL;
const char * overlap_name = NULL;
char chr_str[255];
int chr_num;
unsigned int TSS;
float CpGIOverlap;
context = CpGIOverlap_stream_cast(ns);
// find the genes, determine expression level
if(!gt_node_stream_next(context->in_stream,
&cur_node,
err
) && cur_node != NULL
)
{
*gn = cur_node;
// try casting as a feature node so we can test type
if(!gt_genome_node_try_cast(gt_feature_node_class(), cur_node))
{
return 0;
}
else // we found a feature node
{
// first check if it is a pseudo node, if so find the gene in it if available
if (gt_feature_node_is_pseudo(cur_node))
{
iter = gt_feature_node_iterator_new(cur_node);
if (iter == NULL)
return;
while ((next_node = gt_feature_node_iterator_next(iter)) && !gt_feature_node_has_type(next_node, feature_type_gene));
gt_feature_node_iterator_delete(iter);
if (NULL == (cur_node = next_node))
return 0;
}
if(!gt_feature_node_has_type(cur_node, feature_type_gene))
return 0;
// find name of gene
gene_name = gt_feature_node_get_attribute(cur_node, "Name");
if (gene_name == NULL)
return;
if ( 1 != sscanf(gt_str_get(gt_genome_node_get_seqid(cur_node)), "Chr%d", &chr_num))
return 0;
TSS = (gt_feature_node_get_strand(cur_node) == GT_STRAND_FORWARD) ? gt_genome_node_get_start(cur_node) : gt_genome_node_get_end(cur_node);
// now figure out the overlapping gene
if (! (overlap_name = CpGIOverlap_stream_find_gene_overlap( context, TSS, chr_num)))
return 0;
// save the score into the node
gt_feature_node_set_attribute(cur_node, "cpgi_at_tss", overlap_name);
return 0;
}
}
return err_num;
}