void gthcalcsplicesitescore(GthDbl *splicesitescore,
Traversealignmentstate *oldstate,
const unsigned char *gen_seq_tran,
const unsigned char *ref_seq_tran,
GtAlphabet *gen_alphabet,
GthDPOptionsEST *dp_options_est,
bool acceptorsite)
{
Traversealignmentfunctions travfunctions;
Traversealignmentstate newstate;
Calcsplicesitescoredata data;
gt_assert(dp_options_est);
travfunctions.processmismatch = calcsplicesitescoreprocmismatchordeletion;
travfunctions.processdeletion = calcsplicesitescoreprocmismatchordeletion;
travfunctions.processinsertion = calcsplicesitescoreprocinsertion;
travfunctions.processmatch = calcsplicesitescoreprocmatch;
travfunctions.processintron = calcsplicesitescoreprocintron;
travfunctions.breakcondition = calcsplicesitescorebreakcondition;
/* to prevent manipulation of oldstate we copy it to newstate */
newstate = *oldstate;
if (!acceptorsite) { /* i.e. we want to process a donorsite */
/* to go to the eopptr before the oldstate->eopptr */
newstate.eopptr++;
/* adjusting the sequence pointers to be able to correctly go backwards in
the alignment */
newstate.genomicptr--;
newstate.referenceptr--;
}
data.breaktraversealignment = false;
data.gen_seq_tran = gen_seq_tran;
data.ref_seq_tran = ref_seq_tran;
data.gen_alphabet = gen_alphabet;
data.dp_options_est = dp_options_est;
data.splicesiteweight = (GthFlt) 0.0;
data.maxsplicesiteweight = (GthFlt) 0.0;
data.processedalignmentpositions = 0;
/* for acceptorsites going forward, for donorsites going backward */
gthtraversealignment(acceptorsite, &newstate, false, &data, &travfunctions);
if ((data.processedalignmentpositions >= (SSSWINDOW_MINSIZE_FACTOR *
SPLICE_SITE_SCORE_WINDOW)) &&
(data.splicesiteweight > 0.0) && /* the weights must be positive */
(data.maxsplicesiteweight > 0.0)) {
*splicesitescore = (GthDbl) (data.splicesiteweight /
data.maxsplicesiteweight);
}
else
*splicesitescore = 0.0;
}
void gth_compute_scores(GthSA *sa,
bool proteineop,
GthDPParam *dp_param,
void *dp_options_est,
const unsigned char *gen_seq_tran,
const unsigned char *ref_seq_tran,
const unsigned char *ref_seq_orig,
const GtTransTable *transtable,
unsigned long gen_dp_start,
unsigned long scoreminexonlen,
bool introncutout,
bool gs2out,
GthSplicedSeq *spliced_seq,
unsigned long ref_dp_length,
GtAlphabet *gen_alphabet,
GtAlphabet *ref_alphabet,
GthDPScoresProtein *dp_scores_protein)
{
Traversealignmentfunctions travfunctions;
Traversealignmentstate travstate;
Computebordersandscoresdata data;
GthFlt score, coverageofgenomicsegment, coverageofreferencesegment;
gt_assert(!gth_sa_num_of_exons(sa));
gt_assert(!gth_sa_num_of_introns(sa));
travfunctions.processmismatch = computescoresprocmismatch;
travfunctions.processdeletion = computescoresprocdeletion;
travfunctions.processinsertion = computebordersandscoresprocinsertion;
travfunctions.processmatch = computebordersandscoresprocmatch;
travfunctions.processintron = computebordersandscoresprocintron;
travfunctions.breakcondition = NULL;
/* additional functions for protein edit operations */
travfunctions.processintron_with_1_base_left =
computebordersandscoresprocintron;
travfunctions.processintron_with_2_bases_left =
computebordersandscoresprocintron;
travfunctions.processmismatch_with_1_gap =
computescoresprocmismatchordeletionwithgap;
travfunctions.processmismatch_with_2_gaps =
computescoresprocmismatchordeletionwithgap;
travfunctions.processdeletion_with_1_gap =
computescoresprocmismatchordeletionwithgap;
travfunctions.processdeletion_with_2_gaps =
computescoresprocmismatchordeletionwithgap;
travstate.proteineop = proteineop;
travstate.processing_intron_with_1_base_left = false;
travstate.processing_intron_with_2_bases_left = false;
travstate.alignment = gth_sa_get_editoperations(sa);
travstate.alignmentlength = gth_sa_get_editoperations_length(sa);
travstate.eopptr = travstate.alignment + travstate.alignmentlength - 1;
travstate.genomicptr = gth_sa_genomiccutoff_start(sa);
travstate.referenceptr = gth_sa_referencecutoff_start(sa);
if (travstate.alignmentlength <= 0) {
/* in this case the alignmentscore is set to 0, which leads to discarding
this alignment later */
gth_sa_set_score(sa, 0.0);
return;
}
/* editoperations contain no zero base exons */
gt_assert(gth_sa_contains_no_zero_base_exons(sa));
/* editoperations contain no leading or terminal introns or insertions */
gt_assert(containsnoleadingorterminalintronsorinsertions(travstate.alignment,
travstate
.alignmentlength,
proteineop));
/* sum of edit operations equals referencelength */
gt_assert(gt_eops_equal_referencelength(travstate.alignment,
travstate.alignmentlength,
ref_dp_length
- gth_sa_referencecutoff_start(sa)
- gth_sa_referencecutoff_end(sa),
proteineop));
data.proteineop = proteineop;
data.newexon = true;
data.newintron = true;
data.firstexon = true;
data.introncutout = introncutout;
data.gs2out = gs2out;
data.spliced_seq = spliced_seq;
data.singleexonweight = (GthFlt) 0.0;
data.maxsingleexonweight = (GthFlt) 0.0;
data.overallexonweight = (GthFlt) 0.0;
data.maxoverallexonweight = (GthFlt) 0.0;
data.cumulativelengthofscoredexons = 0;
data.exon.leftgenomicexonborder = GT_UNDEF_ULONG;
data.exon.rightgenomicexonborder = GT_UNDEF_ULONG;
data.exon.leftreferenceexonborder = GT_UNDEF_ULONG;
data.exon.rightreferenceexonborder = GT_UNDEF_ULONG;
data.exon.exonscore = GTH_UNDEF_GTHDBL;
data.intron.donorsiteprobability = GTH_UNDEF_GTHFLT;
data.intron.acceptorsiteprobability = GTH_UNDEF_GTHFLT;
data.intron.donorsitescore = GTH_UNDEF_GTHDBL;
data.intron.acceptorsitescore = GTH_UNDEF_GTHDBL;
data.sa = sa;
data.dp_param = dp_param;
data.dp_options_est = dp_options_est;
data.gen_seq_tran = gen_seq_tran;
data.ref_seq_tran = ref_seq_tran;
data.ref_seq_orig = ref_seq_orig;
data.transtable = transtable;
data.gen_dp_start = gen_dp_start;
data.scoreminexonlen = scoreminexonlen;
data.ref_dp_length = ref_dp_length;
data.gen_alphabet = gen_alphabet;
data.gen_alphabet_characters = gen_alphabet
? gt_alphabet_characters(gen_alphabet)
: NULL;
data.dp_scores_protein = dp_scores_protein;
gthtraversealignment(true, &travstate, proteineop, &data, &travfunctions);
/* this is for saving the last exon */
evalnewintronifpossible(proteineop, &data.newexon, &data.newintron, true,
data.introncutout, data.gs2out, data.spliced_seq,
&data.exon, &data.intron, &data.singleexonweight,
&data.maxsingleexonweight, &data.overallexonweight,
&data.maxoverallexonweight,
&data.cumulativelengthofscoredexons, sa, &travstate,
gen_alphabet, data.dp_param, data.dp_options_est,
data.gen_seq_tran, data.ref_seq_tran,
data.gen_dp_start, data.scoreminexonlen);
/* saving the scores for the whole alignment */
if (data.maxoverallexonweight > 0.0) {
score = data.overallexonweight / data.maxoverallexonweight;
/* XXX: the way the alignmentscore is computed, it is possible to get a
score > 1.0. Since we don't want this, we cap it */
if (score > 1.0)
score = 1.0;
}
else
score = 0.0;
gth_sa_set_score(sa, score);
gth_sa_set_cumlen_scored_exons(sa, data.cumulativelengthofscoredexons);
/* fraction of the gen_dp_length which is scored/weighted */
coverageofgenomicsegment = (GthFlt) data.cumulativelengthofscoredexons /
(GthFlt) gth_sa_gen_dp_length(sa);
/* coverage of genomic segment is valid value */
gt_assert(coverageofgenomicsegment >= 0.0 && coverageofgenomicsegment <= 1.0);
/* fraction of the referencelength which is scored/weighted */
coverageofreferencesegment = (GthFlt) data.cumulativelengthofscoredexons /
(GthFlt) ((proteineop ? GT_CODON_LENGTH : 1) *
gth_sa_ref_total_length(sa));
if (coverageofgenomicsegment > coverageofreferencesegment) {
gth_sa_set_coverage(sa, coverageofgenomicsegment);
gth_sa_set_highest_cov(sa, true);
}
else {
gth_sa_set_coverage(sa, coverageofreferencesegment);
gth_sa_set_highest_cov(sa, false);
}
/* test the assumption that the coverage is never larger then the default */
gt_assert(gth_sa_coverage(sa) <= GTH_DEFAULT_MAX_COVERAGE);
/* compute poly(A) tail position */
gth_sa_calc_polyAtailpos(sa, ref_seq_tran, ref_alphabet);
/* determined exons are forward and consecutive */
gt_assert(gth_sa_exons_are_forward_and_consecutive(sa));
}
static void determine_cutoffs(GthBacktracePath *bp,
GthCutoffmode leadcutoffsmode,
GthCutoffmode termcutoffsmode,
unsigned long cutoffsminexonlen)
{
Traversealignmentfunctions travfunctions;
Traversealignmentstate travstate;
Relaxedcutoffsdata relaxedcutoffsdata;
Strictcutoffsdata strictcutoffsdata;
Minimalcutoffsdata minimalcutoffsdata;
bool proteineop = bp->alphatype == PROTEIN_ALPHA;
/* sum of edit operations equals referencelength (before cutoffs)", */
gt_assert(gth_backtrace_path_is_valid(bp));
/* setting the traverse alignment state */
travstate.proteineop = proteineop;
travstate.processing_intron_with_1_base_left = false;
travstate.processing_intron_with_2_bases_left = false;
travstate.alignment = gth_backtrace_path_get(bp);
travstate.alignmentlength = gth_backtrace_path_length(bp);
travstate.eopptr = travstate.alignment +
travstate.alignmentlength - 1;
travstate.genomicptr = 0;
travstate.referenceptr = 0;
/* cutting of leading indels in the sequences */
switch (leadcutoffsmode) {
case RELAXED:
gt_initRelaxedcutoffsTravfunctions(&travfunctions);
gt_initRelaxedcutoffsdata(&relaxedcutoffsdata, &bp->cutoffs.start);
gthtraversealignment(true, &travstate, proteineop, &relaxedcutoffsdata,
&travfunctions);
break;
case STRICT:
gt_initStrictcutoffsTravfunctions(&travfunctions);
gt_initStrictcutoffsdata(&strictcutoffsdata, &bp->cutoffs.start,
cutoffsminexonlen);
gthtraversealignment(true , &travstate , proteineop, &strictcutoffsdata,
&travfunctions);
break;
case MINIMAL:
gt_initMinimalcutoffsTravfunctions(&travfunctions);
gt_initMinimalcutoffsdata(&minimalcutoffsdata, &bp->cutoffs.start);
gthtraversealignment(true, &travstate, proteineop, &minimalcutoffsdata,
&travfunctions);
break;
default: gt_assert(0);
}
/* resetting the traverse alignment state */
travstate.processing_intron_with_1_base_left = false;
travstate.processing_intron_with_2_bases_left = false;
travstate.eopptr = gth_backtrace_path_get(bp);
travstate.genomicptr = 0;
travstate.referenceptr = 0;
/* cutting of terminal indels in the sequences */
switch (termcutoffsmode) {
case RELAXED:
gt_initRelaxedcutoffsTravfunctions(&travfunctions);
gt_initRelaxedcutoffsdata(&relaxedcutoffsdata, &bp->cutoffs.end);
gthtraversealignment(false, &travstate, proteineop, &relaxedcutoffsdata,
&travfunctions);
break;
case STRICT:
gt_initStrictcutoffsTravfunctions(&travfunctions);
gt_initStrictcutoffsdata(&strictcutoffsdata, &bp->cutoffs.end,
cutoffsminexonlen);
gthtraversealignment(false, &travstate, proteineop, &strictcutoffsdata,
&travfunctions);
break;
case MINIMAL:
gt_initMinimalcutoffsTravfunctions(&travfunctions);
gt_initMinimalcutoffsdata(&minimalcutoffsdata, &bp->cutoffs.end);
gthtraversealignment(false, &travstate, proteineop, &minimalcutoffsdata,
&travfunctions);
break;
default: gt_assert(0);
}
/* sum of edit operations equals referencelength (after cutoffs) */
gt_assert(gth_backtrace_path_is_valid(bp));
}
