/**
* @param read a read containing the variant
* @return the number of hard clipped and low qual bases at the read start (where start is the
* leftmost end w.r.t. the reference)
*/
public static int getNumClippedBasesAtStart(final SAMRecord read) {
// check for hard clips (never consider these bases):
final Cigar c = read.getCigar();
final CigarElement first = c.getCigarElement(0);
int numStartClippedBases = 0;
if (first.getOperator() == CigarOperator.H) {
numStartClippedBases = first.getLength();
}
final byte[] unclippedReadBases = read.getReadBases();
final byte[] unclippedReadQuals = read.getBaseQualities();
// Do a stricter base clipping than provided by CIGAR string, since this one may be too
// conservative,
// and may leave a string of Q2 bases still hanging off the reads.
// TODO: this code may not even get used because HaplotypeCaller already hard clips low quality
// tails
for (int i = numStartClippedBases; i < unclippedReadBases.length; i++) {
if (unclippedReadQuals[i] < PairHMMIndelErrorModel.BASE_QUAL_THRESHOLD)
numStartClippedBases++;
else break;
}
return numStartClippedBases;
}
@Override
public String process(File page, Map<String, String> query) {
loadContigs();
if (query.get("contigName").matches("^[ACGT]+$")) {
contigs.put("manual", query.get("contigName"));
query.put("contigName", "manual");
} else if (query.get("contigName").matches("^Pf3D7.+$")) {
String[] pieces = query.get("contigName").split("[:-]");
int start = Integer.valueOf(pieces[1].replaceAll(",", ""));
int end = Integer.valueOf(pieces[2].replaceAll(",", ""));
ReferenceSequence rseq = REF.getSubsequenceAt(pieces[0], start, end);
contigs.put("manual", new String(rseq.getBases()));
query.put("contigName", "manual");
}
if (query.containsKey("contigName")
&& contigs.containsKey(query.get("contigName"))
&& graphs.containsKey(query.get("graphName"))) {
boolean showLinks = query.get("showLinks").equals("links_on");
String contig = contigs.get(query.get("contigName"));
String originalContig = contigs.get(query.get("contigName"));
String refFormattedString = "";
String kmerOrigin = "";
if (metrics.containsKey(query.get("contigName"))) {
String[] loc = metrics.get(query.get("contigName")).get("canonicalLocus").split("[:-]");
if (!loc[0].equals("*")) {
boolean isRc = metrics.get(query.get("contigName")).get("isRcCanonical").equals("1");
if (isRc) {
contig = SequenceUtils.reverseComplement(contig);
originalContig = SequenceUtils.reverseComplement(originalContig);
}
int locStart = Integer.valueOf(loc[1]);
int locEnd = Integer.valueOf(loc[2]);
Cigar cigar =
cigarStringToCigar(metrics.get(query.get("contigName")).get("cigarCanonical"));
if (cigar.getCigarElement(0).getOperator().equals(CigarOperator.S)) {
locStart -= cigar.getCigarElement(0).getLength();
}
if (cigar
.getCigarElement(cigar.getCigarElements().size() - 1)
.getOperator()
.equals(CigarOperator.S)) {
locEnd += cigar.getCigarElement(cigar.getCigarElements().size() - 1).getLength();
}
String ref = new String(REF.getSubsequenceAt(loc[0], locStart, locEnd).getBases());
StringBuilder refFormatted = new StringBuilder();
int pos = 0;
for (CigarElement ce : cigar.getCigarElements()) {
CigarOperator co = ce.getOperator();
switch (co) {
case S:
refFormatted.append(ref.substring(pos, pos + ce.getLength()).toLowerCase());
break;
case M:
refFormatted.append(ref.substring(pos, pos + ce.getLength()));
break;
case I:
refFormatted.append(StringUtils.repeat("-", ce.getLength()));
break;
}
if (ce.getOperator().consumesReferenceBases()) {
pos += ce.getLength();
}
}
refFormattedString = refFormatted.toString();
kmerOrigin = metrics.get(query.get("contigName")).get("kmerOrigin");
}
}
CortexGraph cg = graphs.get(query.get("graphName"));
String sampleName = cg.getColor(0).getSampleName();
Set<CortexLinksMap> links = new HashSet<CortexLinksMap>();
if (LINKS != null && !LINKS.isEmpty()) {
for (CortexLinksMap link : LINKS) {
if (sampleName.equals(link.getCortexLinks().getColor(0).getSampleName())) {
links.add(link);
}
}
}
Set<String> contigKmers = new HashSet<String>();
for (int i = 0; i <= contig.length() - cg.getKmerSize(); i++) {
String curKmer = contig.substring(i, i + cg.getKmerSize());
contigKmers.add(curKmer);
}
StringBuilder firstFlank = new StringBuilder();
String firstKmer = contig.substring(0, cg.getKmerSize());
Set<String> pks = CortexUtils.getPrevKmers(cg, firstKmer, 0);
Set<String> usedPrevKmers = new HashSet<String>();
usedPrevKmers.add(firstKmer);
while (pks.size() == 1 && usedPrevKmers.size() <= 100) {
String kmer = pks.iterator().next();
firstFlank.insert(0, kmer.charAt(0));
if (usedPrevKmers.contains(kmer)) {
break;
}
usedPrevKmers.add(kmer);
pks = CortexUtils.getPrevKmers(cg, kmer, 0);
}
StringBuilder lastFlank = new StringBuilder();
String lastKmer = contig.substring(contig.length() - cg.getKmerSize(), contig.length());
Set<String> nks = CortexUtils.getNextKmers(cg, lastKmer, 0);
Set<String> usedNextKmers = new HashSet<String>();
usedNextKmers.add(lastKmer);
while (nks.size() == 1 && usedNextKmers.size() <= 100) {
String kmer = nks.iterator().next();
lastFlank.append(kmer.charAt(kmer.length() - 1));
if (usedNextKmers.contains(kmer)) {
break;
}
usedNextKmers.add(kmer);
nks = CortexUtils.getNextKmers(cg, kmer, 0);
}
contig = firstFlank.toString() + contig + lastFlank.toString();
DirectedGraph<CtxVertex, MultiEdge> g =
new DefaultDirectedGraph<CtxVertex, MultiEdge>(MultiEdge.class);
for (int i = 0; i <= contig.length() - cg.getKmerSize(); i++) {
String curKmer = contig.substring(i, i + cg.getKmerSize());
CortexKmer ck = new CortexKmer(curKmer);
CtxVertex curVer =
new CtxVertex(
curKmer,
i,
contigKmers.contains(curKmer) ? VertexType.CONTIG : VertexType.CLIPPED,
cg.findRecord(ck));
g.addVertex(curVer);
String expectedPrevKmer =
(i > 0) ? contig.substring(i - 1, i - 1 + cg.getKmerSize()) : "";
String expectedNextKmer =
(i < contig.length() - cg.getKmerSize())
? contig.substring(i + 1, i + 1 + cg.getKmerSize())
: "";
Set<String> prevKmers = CortexUtils.getPrevKmers(cg, curKmer, 0);
for (String prevKmer : prevKmers) {
if (!expectedPrevKmer.equals(prevKmer)) {
CortexKmer pk = new CortexKmer(prevKmer);
CtxVertex prevVer = new CtxVertex(prevKmer, i - 1, VertexType.IN, cg.findRecord(pk));
MultiEdge me =
g.containsEdge(prevVer, curVer) ? g.getEdge(prevVer, curVer) : new MultiEdge();
me.addGraphName(cg.getCortexFile().getName());
g.addVertex(prevVer);
g.addEdge(prevVer, curVer, me);
}
}
Set<String> nextKmers = CortexUtils.getNextKmers(cg, curKmer, 0);
for (String nextKmer : nextKmers) {
if (!expectedNextKmer.equals(nextKmer)) {
CortexKmer nk = new CortexKmer(nextKmer);
CtxVertex nextVer = new CtxVertex(nextKmer, i + 1, VertexType.OUT, cg.findRecord(nk));
MultiEdge me =
g.containsEdge(curVer, nextVer) ? g.getEdge(curVer, nextVer) : new MultiEdge();
me.addGraphName(cg.getCortexFile().getName());
g.addVertex(nextVer);
g.addEdge(curVer, nextVer, me);
}
}
}
Set<Map<String, Object>> verticesWithLinks = new HashSet<Map<String, Object>>();
DataFrame<String, String, Integer> hv = new DataFrame<String, String, Integer>(0);
for (int q = 0; q <= contig.length() - cg.getKmerSize(); q++) {
// String sk = cv.getBinaryKmer();
String sk = contig.substring(q, q + cg.getKmerSize());
CortexKmer ck = new CortexKmer(sk);
for (CortexLinksMap link : links) {
if (link.containsKey(ck)) {
CortexLinksRecord clr = link.get(ck);
Map<String, Integer> lc =
(!showLinks)
? new HashMap<String, Integer>()
: CortexUtils.getKmersAndCoverageInLink(cg, sk, clr);
Map<String, Object> entry = new HashMap<String, Object>();
entry.put("kmer", sk);
entry.put("lc", lc);
verticesWithLinks.add(entry);
if (showLinks) {
for (CortexJunctionsRecord cjr : clr.getJunctions()) {
List<String> lk = CortexUtils.getKmersInLink(cg, sk, cjr);
for (int i = 0; i < lk.size(); i++) {
String kili = lk.get(i);
for (int j = 0; j < lk.size(); j++) {
String kilj = lk.get(j);
if (i != j) {
hv.set(kili, kilj, hv.get(kili, kilj) + cjr.getCoverage(0));
}
}
}
}
}
}
}
}
/*
int hvMax = 0;
Map<String, Integer> hvlin = new HashMap<String, Integer>();
if (showLinks) {
for (String kili : hv.getRowNames()) {
for (String kilj : hv.getColNames()) {
int cov = hv.get(kili, kilj);
String id = kili + "_" + kilj;
hvlin.put(id, cov);
if (cov > hvMax) {
hvMax = cov;
}
}
}
}
*/
JSONObject jo = new JSONObject();
jo.put("contig", contig);
jo.put("originalContig", originalContig);
jo.put("ref", refFormattedString);
jo.put("kmerOrigin", kmerOrigin);
jo.put("kmerSize", cg.getKmerSize());
jo.put("clipStart", firstFlank.length());
jo.put("clipEnd", contig.length() - lastFlank.length());
List<Map<String, Object>> va = new ArrayList<Map<String, Object>>();
for (CtxVertex v : g.vertexSet()) {
Map<String, Object> vm = new HashMap<String, Object>();
vm.put("base", v.getBase());
vm.put("kmer", v.getKmer());
vm.put("pos", v.getPos());
vm.put("type", v.getVertexType().name());
vm.put("missing", v.isMissingFromGraph());
vm.put("cov", v.getCoverage());
va.add(vm);
}
jo.put("vertices", va);
jo.put("verticesWithLinks", verticesWithLinks);
// jo.put("hvlin", hvlin);
// jo.put("hvmax", hvMax);
return jo.toString();
}
return null;
}
public void printAlignment(final byte[] ref, final byte[] read, final int width) {
final StringBuilder bread = new StringBuilder();
final StringBuilder bref = new StringBuilder();
final StringBuilder match = new StringBuilder();
int i = 0;
int j = 0;
final int offset = getAlignmentStart2wrt1();
Cigar cigar = getCigar();
if (overhangStrategy != OverhangStrategy.SOFTCLIP) {
// we need to go through all the hassle below only if we do not do softclipping;
// otherwise offset is never negative
if (offset < 0) {
for (; j < (-offset); j++) {
bread.append((char) read[j]);
bref.append(' ');
match.append(' ');
}
// at negative offsets, our cigar's first element carries overhanging bases
// that we have just printed above. Tweak the first element to
// exclude those bases. Here we create a new list of cigar elements, so the original
// list/original cigar are unchanged (they are unmodifiable anyway!)
final List<CigarElement> tweaked = new ArrayList<>();
tweaked.addAll(cigar.getCigarElements());
tweaked.set(
0,
new CigarElement(
cigar.getCigarElement(0).getLength() + offset,
cigar.getCigarElement(0).getOperator()));
cigar = new Cigar(tweaked);
}
}
if (offset
> 0) { // note: the way this implementation works, cigar will ever start from S *only* if
// read starts before the ref, i.e. offset = 0
for (; i < getAlignmentStart2wrt1(); i++) {
bref.append((char) ref[i]);
bread.append(' ');
match.append(' ');
}
}
for (final CigarElement e : cigar.getCigarElements()) {
switch (e.getOperator()) {
case M:
for (int z = 0; z < e.getLength(); z++, i++, j++) {
bref.append((i < ref.length) ? (char) ref[i] : ' ');
bread.append((j < read.length) ? (char) read[j] : ' ');
match.append(
(i < ref.length && j < read.length) ? (ref[i] == read[j] ? '.' : '*') : ' ');
}
break;
case I:
for (int z = 0; z < e.getLength(); z++, j++) {
bref.append('-');
bread.append((char) read[j]);
match.append('I');
}
break;
case S:
for (int z = 0; z < e.getLength(); z++, j++) {
bref.append(' ');
bread.append((char) read[j]);
match.append('S');
}
break;
case D:
for (int z = 0; z < e.getLength(); z++, i++) {
bref.append((char) ref[i]);
bread.append('-');
match.append('D');
}
break;
default:
throw new GATKException("Unexpected Cigar element:" + e.getOperator());
}
}
for (; i < ref.length; i++) bref.append((char) ref[i]);
for (; j < read.length; j++) bread.append((char) read[j]);
int pos = 0;
final int maxlength = Math.max(match.length(), Math.max(bread.length(), bref.length()));
while (pos < maxlength) {
print_cautiously(match, pos, width);
print_cautiously(bread, pos, width);
print_cautiously(bref, pos, width);
System.out.println();
pos += width;
}
}
