/** Find every path that overlaps with the specified path. */
static void findOverlaps(const Graph& g,
const Paths& paths, const SeedMap& seedMap,
const Vertex& v, Overlaps& overlaps)
{
ContigPath rc;
if (v.sense) {
rc = paths[v.id];
reverseComplement(rc.begin(), rc.end());
}
const ContigPath& path = v.sense ? rc : paths[v.id];
for (ContigPath::const_iterator it = path.begin();
it != path.end(); ++it) {
if (it->ambiguous())
continue;
pair<SeedMap::const_iterator, SeedMap::const_iterator>
range = seedMap.equal_range(*it);
for (SeedMap::const_iterator seed = range.first;
seed != range.second; ++seed) {
if (v == seed->second)
continue;
int distance = 0;
unsigned overlap = findOverlap(g, paths, it, path.end(),
seed->second, distance);
if (overlap > 0)
overlaps.push_back(Overlap(v, seed->second,
overlap, distance));
}
}
}
/** Record the trimmed contigs. */
static void recordTrimmedContigs(
ContigPath::const_iterator first,
ContigPath::const_iterator last)
{
for (ContigPath::const_iterator it = first; it != last; ++it)
if (!it->ambiguous())
s_trimmedContigs.push_back(it->contigIndex());
}
static void appendToMergeQ(deque<ContigNode>& mergeQ,
set<ContigNode>& seen, const ContigPath& path)
{
for (ContigPath::const_iterator it = path.begin();
it != path.end(); ++it)
if (!it->ambiguous() && seen.insert(*it).second)
mergeQ.push_back(*it);
}
/** Finds all contigs used in each path in paths, and
* marks them as seen in the vector seen. */
static void seenContigs(vector<bool>& seen, const ContigPaths& paths)
{
for (ContigPaths::const_iterator it = paths.begin();
it != paths.end(); ++it)
for (ContigPath::const_iterator itc = it->begin();
itc != it->end(); ++itc)
if (itc->id() < seen.size())
seen[itc->id()] = true;
}
/** Return a pivot suitable for aligning the two paths if one exists,
* otherwise return false.
*/
static pair<ContigNode, bool> findPivot(
const ContigPath& path1, const ContigPath& path2)
{
for (ContigPath::const_iterator it = path2.begin();
it != path2.end(); ++it) {
if (it->ambiguous())
continue;
if (count(path2.begin(), path2.end(), *it) == 1
&& count(path1.begin(), path1.end(), *it) == 1)
return make_pair(*it, true);
}
return make_pair(ContigNode(0), false);
}
/** Return the set of contigs that appear more than once in a single
* solution.
*/
static set<ContigID> findRepeats(ContigID seed,
const ContigPaths& solutions)
{
set<ContigID> repeats;
for (ContigPaths::const_iterator solIt = solutions.begin();
solIt != solutions.end(); ++solIt) {
map<ContigID, unsigned> count;
count[seed]++;
for (ContigPath::const_iterator it = solIt->begin();
it != solIt->end(); ++it)
count[it->contigIndex()]++;
for (map<ContigID, unsigned>::const_iterator
it = count.begin(); it != count.end(); ++it)
if (it->second > 1)
repeats.insert(it->first);
}
return repeats;
}
/** Return all contigs that are tandem repeats, identified as those
* contigs that appear more than once in a single path.
*/
static set<ContigID> findRepeats(const ContigPathMap& paths)
{
set<ContigID> repeats;
for (ContigPathMap::const_iterator pathIt = paths.begin();
pathIt != paths.end(); ++pathIt) {
const ContigPath& path = pathIt->second;
map<ContigID, unsigned> count;
for (ContigPath::const_iterator it = path.begin();
it != path.end(); ++it)
if (!it->ambiguous())
count[it->contigIndex()]++;
for (map<ContigID, unsigned>::const_iterator
it = count.begin(); it != count.end(); ++it)
if (it->second > 1)
repeats.insert(it->first);
}
return repeats;
}
/** Merge the specified path. */
static Contig mergePath(const Graph& g, const Contigs& contigs,
const ContigPath& path)
{
Sequence seq;
unsigned coverage = 0;
for (ContigPath::const_iterator it = path.begin();
it != path.end(); ++it) {
if (!it->ambiguous())
coverage += g[*it].coverage;
if (seq.empty()) {
seq = sequence(contigs, *it);
} else {
assert(it != path.begin());
mergeContigs(g, contigs, *(it-1), *it, seq, path);
}
}
ostringstream ss;
ss << seq.size() << ' ' << coverage << ' ';
pathToComment(ss, g, path);
return Contig(ss.str(), seq);
}
