static int store_in_sa_collection(void *data, GthSA *sa,
GT_UNUSED const char *outputfilename,
GT_UNUSED GtError *err)
{
SACollectionData *sa_collection_data = (SACollectionData*) data;
bool inserted;
inserted = gth_sa_collection_insert_sa(sa_collection_data->sa_collection, sa,
sa_collection_data->sa_filter,
sa_collection_data->stat);
if (!inserted) { /* unsuccessful insertion; discard sa */
gth_sa_delete(sa);
}
return 0;
}
/* the following function saves <sa> by inserting it into <sa_collection> and
sets <significantmatchfound> to true, if the insertion was successful */
static void save_sa(GthSACollection *sa_collection, GthSA *sa,
GthSAFilter *sa_filter, GthMatchInfo *match_info,
GthStat *stat)
{
if (!gth_sa_collection_insert_sa(sa_collection, sa, sa_filter, stat)) {
/* unsuccessful insertion; discard sa */
gth_sa_delete(sa);
match_info->call_number--;
}
else {
/* else successful insertion */
match_info->significant_match_found = true;
}
}
static int store_in_subset_file(void *data, GthSA *sa,
const char *outputfilename, GtError *err)
{
Store_in_subset_file_data *store_in_subset_file_data =
(Store_in_subset_file_data*) data;
double split_determing_percentage = 0.0;
unsigned long filenum;
char filenamesuffix[4];
int had_err = 0;
gt_error_check(err);
/* filter before we do any further processing */
if (gth_sa_filter_filter_sa(store_in_subset_file_data->sa_filter, sa)) {
/* and free it afterwards */
gth_sa_delete(sa);
/* discard */
return 0;
}
/* check whether we got a new output file to process */
if (!store_in_subset_file_data->current_outputfilename) {
store_in_subset_file_data->current_outputfilename =
gt_cstr_dup(outputfilename);
}
else if (strcmp(store_in_subset_file_data->current_outputfilename,
outputfilename)) {
/* close current output files */
close_output_files(store_in_subset_file_data);
gt_free(store_in_subset_file_data->current_outputfilename);
}
/* determine in which file the current sa needs to be put */
switch (store_in_subset_file_data->gthsplitinfo->splitmode) {
case ALIGNMENTSCORE_SPLIT:
split_determing_percentage = gth_sa_score(sa);
strcpy(filenamesuffix, "scr");
break;
case COVERAGE_SPLIT:
split_determing_percentage = gth_sa_coverage(sa);
strcpy(filenamesuffix, "cov");
break;
default: gt_assert(0);
}
gt_assert(split_determing_percentage >= 0.0);
/* XXX: change into an assertion when coverage problem is fixed */
if (split_determing_percentage > 1.0)
split_determing_percentage = 1.0;
if (split_determing_percentage == 1.0)
filenum = store_in_subset_file_data->num_of_subset_files - 1;
else {
filenum = floor(split_determing_percentage * 100.0 /
store_in_subset_file_data->gthsplitinfo->range);
}
gt_assert(filenum < store_in_subset_file_data->num_of_subset_files);
/* make sure the file exists and is open */
if (!store_in_subset_file_data->subset_files[filenum]) {
gt_assert(store_in_subset_file_data->subset_filenames[filenum] == NULL);
store_in_subset_file_data->subset_filenames[filenum] = gt_str_new();
gt_str_append_cstr_nt(store_in_subset_file_data->subset_filenames[filenum],
outputfilename,
gt_file_basename_length(outputfilename));
gt_str_append_char(store_in_subset_file_data->subset_filenames[filenum],
'.');
gt_str_append_cstr(store_in_subset_file_data->subset_filenames[filenum],
filenamesuffix);
gt_str_append_ulong(store_in_subset_file_data->subset_filenames[filenum],
filenum *
store_in_subset_file_data->gthsplitinfo->range);
gt_str_append_char(store_in_subset_file_data->subset_filenames[filenum],
'-');
gt_str_append_ulong(store_in_subset_file_data->subset_filenames[filenum],
(filenum + 1) *
store_in_subset_file_data->gthsplitinfo->range);
gt_str_append_cstr(store_in_subset_file_data->subset_filenames[filenum],
gt_file_mode_suffix(store_in_subset_file_data
->gthsplitinfo->file_mode));
/* if not disabled by -force, check if file already exists */
if (!store_in_subset_file_data->gthsplitinfo->force) {
store_in_subset_file_data->subset_files[filenum] =
gt_file_open(store_in_subset_file_data->gthsplitinfo->file_mode,
gt_str_get(store_in_subset_file_data
->subset_filenames[filenum]), "r", NULL);
if (store_in_subset_file_data->subset_files[filenum]) {
gt_error_set(err, "file \"%s\" exists already. use option -%s to "
"overwrite", gt_str_get(store_in_subset_file_data
->subset_filenames[filenum]),
GT_FORCE_OPT_CSTR);
had_err = -1;
}
}
if (!had_err) {
/* open split file for writing */
store_in_subset_file_data->subset_files[filenum] =
gt_file_xopen_file_mode(store_in_subset_file_data->gthsplitinfo
->file_mode,
gt_str_get(store_in_subset_file_data
->subset_filenames[filenum]), "w");
/* store XML header in file */
gth_xml_show_leader(true,
store_in_subset_file_data->subset_files[filenum]);
}
}
/* put it there */
if (!had_err) {
gth_xml_inter_sa_visitor_set_outfp(store_in_subset_file_data->sa_visitor,
store_in_subset_file_data
->subset_files[filenum]);
gth_sa_visitor_visit_sa(store_in_subset_file_data->sa_visitor, sa);
}
/* adjust counter */
if (!had_err)
store_in_subset_file_data->subset_file_sa_counter[filenum]++;
/* and free it afterwards */
gth_sa_delete(sa);
return had_err;
}
static int calc_spliced_alignments(GthSACollection *sa_collection,
GthChainCollection *chain_collection,
GthCallInfo *call_info,
GthInput *input,
GthStat *stat,
GtUword gen_file_num,
GtUword ref_file_num,
bool directmatches,
GthMatchInfo *match_info,
GthDNACompletePathMatrixJT
dna_complete_path_matrix_jt,
GthProteinCompletePathMatrixJT
protein_complete_path_matrix_jt)
{
const unsigned char *ref_seq_tran, *ref_seq_orig, *ref_seq_tran_rc = NULL,
*ref_seq_orig_rc = NULL;
GtUword chainctr, gen_offset = GT_UNDEF_UWORD, gen_total_length,
ref_total_length;
GtFile *outfp = call_info->out->outfp;
GtRange gen_seq_bounds, gen_seq_bounds_rc;
bool refseqisdna;
GthChain *chain;
GtRange range;
GthSA *saA;
int rval;
gt_assert(sa_collection && chain_collection);
refseqisdna = gth_input_ref_file_is_dna(input, ref_file_num);
for (chainctr = 0;
chainctr < gth_chain_collection_size(chain_collection);
chainctr++) {
chain = gth_chain_collection_get(chain_collection, chainctr);
if (++match_info->call_number > call_info->firstalshown &&
call_info->firstalshown > 0) {
if (!(call_info->out->xmlout || call_info->out->gff3out))
gt_file_xfputc('\n', outfp);
else if (call_info->out->xmlout)
gt_file_xprintf(outfp, "<!--\n");
if (!call_info->out->gff3out) {
gt_file_xprintf(outfp, "Maximal matching %s count (%u) reached.\n",
refseqisdna ? "EST" : "protein",
call_info->firstalshown);
gt_file_xprintf(outfp, "Only the first %u matches will be "
"displayed.\n", call_info->firstalshown);
}
if (!(call_info->out->xmlout || call_info->out->gff3out))
gt_file_xfputc('\n', outfp);
else if (call_info->out->xmlout)
gt_file_xprintf(outfp, "-->\n");
match_info->max_call_number_reached = true;
break; /* break out of loop */
}
/* compute considered genomic regions if not set by -frompos */
if (!gth_input_use_substring_spec(input)) {
gen_seq_bounds = gth_input_get_genomic_range(input, chain->gen_file_num,
chain->gen_seq_num);
gen_total_length = gt_range_length(&gen_seq_bounds);
gen_offset = gen_seq_bounds.start;
gen_seq_bounds_rc = gen_seq_bounds;
}
else {
/* genomic multiseq contains exactly one sequence */
gt_assert(gth_input_num_of_gen_seqs(input, chain->gen_file_num) == 1);
gen_total_length = gth_input_genomic_file_total_length(input,
chain
->gen_file_num);
gen_seq_bounds.start = gth_input_genomic_substring_from(input);
gen_seq_bounds.end = gth_input_genomic_substring_to(input);
gen_offset = 0;
gen_seq_bounds_rc.start = gen_total_length - 1 - gen_seq_bounds.end;
gen_seq_bounds_rc.end = gen_total_length - 1 - gen_seq_bounds.start;
}
/* "retrieving" the reference sequence */
range = gth_input_get_reference_range(input, chain->ref_file_num,
chain->ref_seq_num);
ref_seq_tran = gth_input_current_ref_seq_tran(input) + range.start;
ref_seq_orig = gth_input_current_ref_seq_orig(input) + range.start;
if (refseqisdna) {
ref_seq_tran_rc = gth_input_current_ref_seq_tran_rc(input) + range.start;
ref_seq_orig_rc = gth_input_current_ref_seq_orig_rc(input) + range.start;
}
ref_total_length = range.end - range.start + 1;
/* check if protein sequences have a stop amino acid */
if (!refseqisdna && !match_info->stop_amino_acid_warning &&
ref_seq_orig[ref_total_length - 1] != GT_STOP_AMINO) {
GtStr *ref_id = gt_str_new();
gth_input_save_ref_id(input, ref_id, chain->ref_file_num,
chain->ref_seq_num);
gt_warning("protein sequence '%s' (#" GT_WU " in file %s) does not end "
"with a stop amino acid ('%c'). If it is not a protein "
"fragment you should add a stop amino acid to improve the "
"prediction. For example with `gt seqtransform "
"-addstopaminos` (see http://genometools.org for details).",
gt_str_get(ref_id), chain->ref_seq_num,
gth_input_get_reference_filename(input, chain->ref_file_num),
GT_STOP_AMINO);
match_info->stop_amino_acid_warning = true;
gt_str_delete(ref_id);
}
/* allocating space for alignment */
saA = gth_sa_new_and_set(directmatches, true, input, chain->gen_file_num,
chain->gen_seq_num, chain->ref_file_num,
chain->ref_seq_num, match_info->call_number,
gen_total_length, gen_offset, ref_total_length);
/* extend the DP borders to the left and to the right */
gth_chain_extend_borders(chain, &gen_seq_bounds, &gen_seq_bounds_rc,
gen_total_length, gen_offset);
/* From here on the dp positions always refer to the forward strand of the
genomic DNA. */
/* call the Dynamic Programming */
if (refseqisdna) {
rval = call_dna_DP(directmatches, call_info, input, stat,
sa_collection, saA, gen_file_num, ref_file_num,
gen_total_length, gen_offset, &gen_seq_bounds,
&gen_seq_bounds_rc, ref_total_length, range.start,
chainctr, gth_chain_collection_size(chain_collection),
match_info, ref_seq_tran, ref_seq_orig,
ref_seq_tran_rc, ref_seq_orig_rc, chain,
dna_complete_path_matrix_jt,
protein_complete_path_matrix_jt);
}
else {
rval = call_protein_DP(directmatches, call_info, input,
stat, sa_collection, saA, gen_file_num,
ref_file_num, gen_total_length, gen_offset,
&gen_seq_bounds, &gen_seq_bounds_rc,
ref_total_length, range.start, chainctr,
gth_chain_collection_size(chain_collection),
match_info, ref_seq_tran, ref_seq_orig, chain,
dna_complete_path_matrix_jt,
protein_complete_path_matrix_jt);
}
/* check return value */
if (rval == GTH_ERROR_DP_PARAMETER_ALLOCATION_FAILED) {
/* statistics bookkeeping */
gth_stat_increment_numoffailedDPparameterallocations(stat);
gth_stat_increment_numofundeterminedSAs(stat);
/* free space */
gth_sa_delete(saA);
match_info->call_number--;
continue; /* continue with the next DP range */
}
else if (rval)
return -1;
}
if (!call_info->out->xmlout && !call_info->out->gff3out && !directmatches &&
!match_info->significant_match_found &&
match_info->call_number <= call_info->firstalshown) {
show_no_match_line(gth_input_get_alphatype(input, ref_file_num), outfp);
}
return 0;
}
static int call_protein_DP(bool directmatches,
GthCallInfo *call_info,
GthInput *input,
GthStat *stat,
GthSACollection *sa_collection,
GthSA *saA,
GtUword gen_file_num,
GtUword ref_file_num,
GtUword gen_total_length,
GtUword gen_offset,
const GtRange *gen_seq_bounds,
const GtRange *gen_seq_bounds_rc,
GtUword ref_total_length,
GtUword ref_offset,
GtUword chainctr,
GtUword num_of_chains,
GthMatchInfo *match_info,
const unsigned char *ref_seq_tran,
const unsigned char *ref_seq_orig,
GthChain *chain,
GthDNACompletePathMatrixJT
dna_complete_path_matrix_jt,
GthProteinCompletePathMatrixJT
protein_complete_path_matrix_jt)
{
GtFile *outfp = call_info->out->outfp;
int rval;
#ifndef NDEBUG
/* strand is in searchmode */
if (directmatches)
gt_assert(gth_input_forward(input));
else
gt_assert(gth_input_reverse(input));
#endif
/* calculate alignment */
rval = callsahmt(false, saA, directmatches, gen_file_num, ref_file_num,
chain, gen_total_length, gen_offset, gen_seq_bounds,
gen_seq_bounds_rc, ref_seq_tran, ref_seq_orig,
ref_total_length, ref_offset, input,
&call_info->simfilterparam.introncutoutinfo, stat, chainctr,
num_of_chains, call_info->translationtable, directmatches,
call_info->proteinexonpenal, call_info->splice_site_model,
call_info->dp_options_core, call_info->dp_options_est,
call_info->dp_options_postpro, dna_complete_path_matrix_jt,
protein_complete_path_matrix_jt, call_info->out);
if (rval && rval != GTH_ERROR_SA_COULD_NOT_BE_DETERMINED) {
/* ^ this error is treated below */
return rval;
}
if (rval == GTH_ERROR_SA_COULD_NOT_BE_DETERMINED ||
isunsuccessfulalignment(saA, call_info->out->comments, outfp)) {
match_info->call_number--;
/* if the spliced alignment was unsuccessful, it is deleted and the
next hit is considered. */
gth_sa_delete(saA);
/* continue */
return 0;
}
/* we can save the alignment now */
save_sa(sa_collection, saA, call_info->sa_filter, match_info, stat);
return 0;
}
static int call_dna_DP(bool directmatches, GthCallInfo *call_info,
GthInput *input, GthStat *stat,
GthSACollection *sa_collection, GthSA *saA,
GtUword gen_file_num,
GtUword ref_file_num,
GtUword gen_total_length,
GtUword gen_offset,
const GtRange *gen_seq_bounds,
const GtRange *gen_seq_bounds_rc,
GtUword ref_total_length, GtUword ref_offset,
GtUword chainctr,
GtUword num_of_chains, GthMatchInfo *match_info,
const unsigned char *ref_seq_tran,
const unsigned char *ref_seq_orig,
const unsigned char *ref_seq_tran_rc,
const unsigned char *ref_seq_orig_rc,
GthChain *chain,
GthDNACompletePathMatrixJT dna_complete_path_matrix_jt,
GthProteinCompletePathMatrixJT
protein_complete_path_matrix_jt)
{
int rval;
bool bothstrandsanalyzed, firstdp = true,
GT_UNUSED gs2outdirectmatches = directmatches;
GthSA *saB = NULL;
GtFile *outfp = call_info->out->outfp;
if (directmatches ? gth_input_forward(input)
: gth_input_reverse(input)) {
/* calculate alignment */
rval = callsahmt(true, saA, directmatches, gen_file_num, ref_file_num,
chain, gen_total_length, gen_offset, gen_seq_bounds,
gen_seq_bounds_rc,
ref_seq_tran, ref_seq_orig, ref_total_length, ref_offset,
input, &call_info->simfilterparam.introncutoutinfo, stat,
chainctr, num_of_chains, call_info->translationtable,
directmatches, call_info->proteinexonpenal,
call_info->splice_site_model, call_info->dp_options_core,
call_info->dp_options_est, call_info->dp_options_postpro,
dna_complete_path_matrix_jt,
protein_complete_path_matrix_jt, call_info->out);
if (rval && rval != GTH_ERROR_SA_COULD_NOT_BE_DETERMINED) {
/* ^ this error is treated below */
return rval;
}
firstdp = false;
bothstrandsanalyzed = gth_input_both(input);
if (rval == GTH_ERROR_SA_COULD_NOT_BE_DETERMINED ||
isunsuccessfulalignment(saA, call_info->out->comments, outfp)) {
match_info->call_number--;
/* if the spliced alignment was unsuccessful, it is deleted and the
next hit is considered. */
gth_sa_delete(saA);
return 0; /* continue */
}
/* if not both strands are analyzed, we can save this alignment now.
Otherwise we have to calculate the alignment to the other strand
first and then save the better one. */
if (!bothstrandsanalyzed)
save_sa(sa_collection, saA, call_info->sa_filter, match_info, stat);
}
if (directmatches ? gth_input_reverse(input)
: gth_input_forward(input)) {
if ((firstdp || gth_sa_is_poor(saA, call_info->minaveragessp)) &&
!call_info->cdnaforwardonly) {
if (firstdp) {
/* space for first alignment is already allocated, bu we have to
change the direction of the genomic and the reference strand */
gth_sa_set_gen_strand(saA, !directmatches);
gth_sa_set_ref_strand(saA, false);
}
else {
/* allocating space for second alignment */
saB = gth_sa_new_and_set(!directmatches, false, input,
chain->gen_file_num, chain->gen_seq_num,
chain->ref_file_num, chain->ref_seq_num,
match_info->call_number, gen_total_length,
gen_offset, ref_total_length);
}
/* setting gs2outdirectmatches (for compatibility) */
gs2outdirectmatches = (bool) !directmatches;
/* calculate alignment */
rval = callsahmt(true, firstdp ? saA : saB, !directmatches,
gen_file_num, ref_file_num, chain, gen_total_length,
gen_offset, gen_seq_bounds, gen_seq_bounds_rc,
ref_seq_tran_rc, ref_seq_orig_rc, ref_total_length,
ref_offset, input,
&call_info->simfilterparam.introncutoutinfo, stat,
chainctr, num_of_chains, call_info->translationtable,
directmatches, call_info->proteinexonpenal,
call_info->splice_site_model, call_info->dp_options_core,
call_info->dp_options_est, call_info->dp_options_postpro,
dna_complete_path_matrix_jt,
protein_complete_path_matrix_jt, call_info->out);
if (rval && rval != GTH_ERROR_SA_COULD_NOT_BE_DETERMINED) {
/* ^ this error is treated below */
return rval;
}
if (firstdp) {
if (rval == GTH_ERROR_SA_COULD_NOT_BE_DETERMINED ||
isunsuccessfulalignment(saA, call_info->out->comments, outfp)) {
/* for compatibility with GS2 */
/* XXX: makes no sense. Possibly only if -gs2out is used. */
match_info->significant_match_found= true;
/* if the spliced alignment was unsuccessful, it is deleted and
the next hit is considered. */
gth_sa_delete(saA);
return 0; /* continue */
}
save_sa(sa_collection, saA, call_info->sa_filter, match_info, stat);
}
else /* !firstdp */
{
if (rval == GTH_ERROR_SA_COULD_NOT_BE_DETERMINED ||
isunsuccessfulalignment(saB, call_info->out->comments, outfp) ||
!gth_sa_B_is_better_than_A(saA, saB)) {
/* insert first SA */
save_sa(sa_collection, saA, call_info->sa_filter, match_info, stat);
/* discard second SA */
gth_sa_delete(saB);
}
else {
/* insert second SA */
save_sa(sa_collection, saB, call_info->sa_filter, match_info, stat);
/* free first SA */
gth_sa_delete(saA);
}
}
}
else
save_sa(sa_collection, saA, call_info->sa_filter, match_info, stat);
}
return 0;
}
