/**
* \brief TODO
*/
bool
Exon::sameRegion(const GenomicRegion &r) const {
return ((r.get_chrom() == this->region.get_chrom()) &&
(r.get_start() == this->region.get_start()) &&
(r.get_end() == this->region.get_end()) &&
(r.get_strand() == this->region.get_strand()));
}
void
merge(istream &cpg_stream, ostream &dmr_stream, double cutoff) {
bool skipped_last_cpg, sig_raw;
GenomicRegion dmr;
dmr.set_name("dmr");
size_t dmr_test_cov = 0; size_t dmr_test_meth = 0;
size_t dmr_rest_cov = 0; size_t dmr_rest_meth = 0;
size_t test_cov = 0; size_t test_meth = 0;
size_t rest_cov = 0; size_t rest_meth = 0;
// Find the first significant CpG, or terminate the function if none exist.
if (!read_next_significant_cpg(cpg_stream, dmr, cutoff, skipped_last_cpg,
sig_raw, test_cov, test_meth, rest_cov, rest_meth))
return;
dmr.set_score(sig_raw);
dmr_test_cov += test_cov;
dmr_test_meth += test_meth;
dmr_rest_cov += rest_cov;
dmr_rest_meth += rest_meth;
GenomicRegion cpg;
cpg.set_name("dmr");
while(read_next_significant_cpg(cpg_stream, cpg, cutoff, skipped_last_cpg,
sig_raw, test_cov, test_meth, rest_cov, rest_meth)) {
if (skipped_last_cpg || cpg.get_chrom() != dmr.get_chrom()) {
if (dmr.get_score() != 0)
dmr_stream << dmr.get_chrom() << '\t'
<< dmr.get_start() << '\t'
<< dmr.get_end() << '\t'
<< dmr.get_name() << '\t'
<< dmr.get_score() << '\t'
<< double(dmr_test_meth)/dmr_test_cov -
double(dmr_rest_meth)/dmr_rest_cov << std::endl;
dmr = cpg;
dmr.set_score(sig_raw);
dmr_test_cov = test_cov;
dmr_test_meth = test_meth;
dmr_rest_cov = rest_cov;
dmr_rest_meth = rest_meth;
} else {
dmr.set_end(cpg.get_end());
dmr.set_score(dmr.get_score() + sig_raw);
dmr_test_cov += test_cov;
dmr_test_meth += test_meth;
dmr_rest_cov += rest_cov;
dmr_rest_meth += rest_meth;
}
}
dmr_stream << dmr << std::endl;
}
/**
* \brief Does this exon partially overlap the given genomic region? This
* function doesn't consider an exact region match (i.e. same chrom,
* start, end, and strand) to be a 'partial overlap'.
*/
bool
Exon::partialOverlap(const GenomicRegion &r) const {
if (r.get_chrom() != this->region.get_chrom()) return false;
if (r.get_strand() != this->region.get_strand()) return false;
// by now they must be one the same chromosome and strand... first see if
// they match exactly
if ((r.get_start() == this->region.get_start()) &&
(r.get_end() == this->region.get_end())) return false;
// not an exact match
return (r.overlaps(this->region));
}
static void
convert_coordinates(const unordered_map<size_t, size_t> &cpgs,
GenomicRegion ®ion) {
const unordered_map<size_t, size_t>::const_iterator
start_itr(cpgs.find(region.get_start()));
const unordered_map<size_t, size_t>::const_iterator
end_itr(cpgs.find(region.get_end()));
if (start_itr == cpgs.end() || end_itr == cpgs.end())
throw runtime_error("could not convert:\n" + region.tostring());
region.set_start(start_itr->second);
region.set_end(end_itr->second);
}
static bool
precedes(const string &chrom, const size_t position,
const GenomicRegion &r) {
return chrom < r.get_chrom() ||
(chrom == r.get_chrom() && position < r.get_start());
}
static inline bool
start_leq(const GenomicRegion &a, const GenomicRegion &b) {
return a.get_start() <= b.get_start();
}
