public void find_coverage(SAMResource sres) {
int start_base = sres.region.range.start;
int end_base = sres.region.range.end;
int coverage_len = (end_base - start_base) + 1;
int i, end, ref_i, read_i, len;
int[] coverage = new int[coverage_len];
Arrays.fill(coverage, 0);
WorkingFile wf = null;
if (outfile != null) {
try {
wf = new WorkingFile(outfile);
ps = wf.getPrintStream();
} catch (Exception e) {
System.err.println("I/O error: " + e); // debug
e.printStackTrace();
System.exit(1);
}
}
try {
//
// gather coverage info:
//
CloseableIterator<SAMRecord> iterator = sres.get_iterator();
int read_count = 0;
int ref_min = -1;
int ref_max = -1;
while (iterator.hasNext()) {
SAMRecord sr = iterator.next();
read_count++;
// System.err.println(sr.getReadName() + ": " + sr.getAlignmentStart() + "-" +
// sr.getAlignmentEnd()); // debug
if (sr.getReadUnmappedFlag()) continue;
if (sr.getDuplicateReadFlag()) {
if (verbose_mode)
System.err.println(
sr.getReadName()
+ "."
+ (sr.getReadNegativeStrandFlag() ? "R" : "F")
+ " ignoring, duplicate");
continue;
}
byte[] read = sr.getReadBases();
byte[] quals = sr.getBaseQualities();
for (AlignmentBlock ab : sr.getAlignmentBlocks()) {
len = ab.getLength();
read_i = ab.getReadStart() - 1;
ref_i = ab.getReferenceStart() - start_base;
if (ref_min == -1 || ref_i < ref_min) ref_min = ref_i;
for (i = read_i, end = read_i + len; i < end; i++, ref_i++) {
if (ref_i >= 0 && ref_i < coverage_len) {
if (quals[i] >= MIN_QUALITY) {
if (verbose_mode)
System.err.println(
sr.getReadName()
+ "."
+ (sr.getReadNegativeStrandFlag() ? "R" : "F")
+ " hit at "
+ (ref_i + start_base)
+ " as="
+ sr.getAlignmentStart()
+ " ae="
+ sr.getAlignmentEnd());
coverage[ref_i]++;
} else if (verbose_mode) {
System.err.println(
sr.getReadName()
+ "."
+ (sr.getReadNegativeStrandFlag() ? "R" : "F")
+ " qual_reject at "
+ (ref_i + start_base)
+ " as="
+ sr.getAlignmentStart()
+ " ae="
+ sr.getAlignmentEnd());
}
}
}
if (ref_max == -1 || ref_i > ref_max) ref_max = ref_i;
}
}
sres.close();
System.err.println(
"records:"
+ read_count
+ " ref_min:"
+ (ref_min + start_base)
+ " ref_max:"
+ (ref_max + start_base)); // debug
//
// report coverage info:
//
for (i = 0; i < coverage.length; i++) {
if (name != null) ps.print(name + ",");
ps.println((i + start_base) + "," + coverage[i]); // debug
}
if (wf != null) wf.finish();
} catch (Exception e) {
System.err.println("ERROR: " + e); // debug
e.printStackTrace();
}
}
private void collectQualityData(final SAMRecord record, final ReferenceSequence reference) {
// If the read isnt an aligned PF read then look at the read for no-calls
if (record.getReadUnmappedFlag()
|| record.getReadFailsVendorQualityCheckFlag()
|| !doRefMetrics) {
final byte[] readBases = record.getReadBases();
for (int i = 0; i < readBases.length; i++) {
if (SequenceUtil.isNoCall(readBases[i])) {
badCycleHistogram.increment(
CoordMath.getCycle(record.getReadNegativeStrandFlag(), readBases.length, i));
}
}
} else if (!record.getReadFailsVendorQualityCheckFlag()) {
final boolean highQualityMapping = isHighQualityMapping(record);
if (highQualityMapping) metrics.PF_HQ_ALIGNED_READS++;
final byte[] readBases = record.getReadBases();
final byte[] refBases = reference.getBases();
final byte[] qualities = record.getBaseQualities();
final int refLength = refBases.length;
long mismatchCount = 0;
long hqMismatchCount = 0;
for (final AlignmentBlock alignmentBlock : record.getAlignmentBlocks()) {
final int readIndex = alignmentBlock.getReadStart() - 1;
final int refIndex = alignmentBlock.getReferenceStart() - 1;
final int length = alignmentBlock.getLength();
for (int i = 0; i < length && refIndex + i < refLength; ++i) {
final int readBaseIndex = readIndex + i;
boolean mismatch =
!SequenceUtil.basesEqual(readBases[readBaseIndex], refBases[refIndex + i]);
boolean bisulfiteBase = false;
if (mismatch && isBisulfiteSequenced) {
if ((record.getReadNegativeStrandFlag()
&& (refBases[refIndex + i] == 'G' || refBases[refIndex + i] == 'g')
&& (readBases[readBaseIndex] == 'A' || readBases[readBaseIndex] == 'a'))
|| ((!record.getReadNegativeStrandFlag())
&& (refBases[refIndex + i] == 'C' || refBases[refIndex + i] == 'c')
&& (readBases[readBaseIndex] == 'T')
|| readBases[readBaseIndex] == 't')) {
bisulfiteBase = true;
mismatch = false;
}
}
if (mismatch) mismatchCount++;
metrics.PF_ALIGNED_BASES++;
if (!bisulfiteBase) nonBisulfiteAlignedBases++;
if (highQualityMapping) {
metrics.PF_HQ_ALIGNED_BASES++;
if (!bisulfiteBase) hqNonBisulfiteAlignedBases++;
if (qualities[readBaseIndex] >= BASE_QUALITY_THRESHOLD)
metrics.PF_HQ_ALIGNED_Q20_BASES++;
if (mismatch) hqMismatchCount++;
}
if (mismatch || SequenceUtil.isNoCall(readBases[readBaseIndex])) {
badCycleHistogram.increment(
CoordMath.getCycle(record.getReadNegativeStrandFlag(), readBases.length, i));
}
}
}
mismatchHistogram.increment(mismatchCount);
hqMismatchHistogram.increment(hqMismatchCount);
// Add any insertions and/or deletions to the global count
for (final CigarElement elem : record.getCigar().getCigarElements()) {
final CigarOperator op = elem.getOperator();
if (op == CigarOperator.INSERTION || op == CigarOperator.DELETION) ++this.indels;
}
}
}
@Override
public void execute() {
log.info("Initializing kmer code map...");
Map<Character, Integer> kmerCodeIndices = new HashMap<Character, Integer>();
kmerCodeIndices.put('0', 1);
kmerCodeIndices.put('A', 3);
kmerCodeIndices.put('B', 4);
kmerCodeIndices.put('C', 5);
kmerCodeIndices.put('_', 6);
kmerCodeIndices.put('.', 7);
kmerCodeIndices.put('1', 9);
Map<Character, String> kmerCodeNames = new LinkedHashMap<Character, String>();
kmerCodeNames.put('0', "ref0");
kmerCodeNames.put('A', "repetitive");
kmerCodeNames.put('B', "both");
kmerCodeNames.put('C', "lowcoverage");
kmerCodeNames.put('_', "lowconfidence");
kmerCodeNames.put('.', "novel");
kmerCodeNames.put('1', "ref1");
if (KMER_CODE_NAMES != null) {
for (Character c : kmerCodeNames.keySet()) {
String cStr = String.valueOf(c);
if (KMER_CODE_NAMES.containsKey(cStr)) {
kmerCodeNames.put(c, KMER_CODE_NAMES.get(cStr));
}
}
}
for (Character c : kmerCodeNames.keySet()) {
log.info(" {} {}: {}", c, kmerCodeIndices.get(c), kmerCodeNames.get(c));
}
log.info("Loading annotated contigs...");
Map<String, Map<String, String>> annotatedContigs = new HashMap<String, Map<String, String>>();
int kmerSize = 0;
if (ANN.length() > 0) {
TableReader tr = new TableReader(ANN);
for (Map<String, String> te : tr) {
String contigName = te.get("contigName");
if (kmerSize == 0) {
kmerSize = te.get("seq").length() - te.get("kmerOrigin").length() + 1;
}
annotatedContigs.put(contigName, te);
String[] ref0ToCanonicalExact =
(te.get("ref0ToCanonicalExact").equals("NA")
|| te.get("ref0ToCanonicalExact").equals("*:0-0")
? "NA:0-0"
: te.get("ref0ToCanonicalExact"))
.split("[:-]");
String[] ref1ToCanonicalExact =
(te.get("ref1ToCanonicalExact").equals("NA")
|| te.get("ref1ToCanonicalExact").equals("*:0-0")
? "NA:0-0"
: te.get("ref1ToCanonicalExact"))
.split("[:-]");
cout.println(
te.get("sampleName")
+ "_"
+ te.get("accession")
+ "_"
+ contigName
+ " "
+ ref0ToCanonicalExact[0]
+ " "
+ ref0ToCanonicalExact[1]
+ " "
+ ref0ToCanonicalExact[2]
+ " radius1=0.8r");
cout.println(
te.get("sampleName")
+ "_"
+ te.get("accession")
+ "_"
+ contigName
+ " "
+ ref1ToCanonicalExact[0]
+ " "
+ ref1ToCanonicalExact[1]
+ " "
+ ref1ToCanonicalExact[2]
+ " radius2=0.6r");
}
}
log.info(" contigs: {}", annotatedContigs.size());
log.info(" kmer size: {}", kmerSize);
log.info("Computing kmer inheritance information...");
SAMFileHeader sfh = CONTIGS.getFileHeader();
for (Character c : kmerCodeNames.keySet()) {
SAMReadGroupRecord rgr = new SAMReadGroupRecord(kmerCodeNames.get(c));
rgr.setSample(kmerCodeNames.get(c));
sfh.addReadGroup(rgr);
}
SAMFileWriterFactory sfwf = new SAMFileWriterFactory();
sfwf.setCreateIndex(true);
SAMFileWriter sfw = sfwf.makeBAMWriter(sfh, false, bout);
TableWriter tw = new TableWriter(sout);
Set<IGVEntry> igvEntries = new TreeSet<IGVEntry>();
int numContigs = 0;
for (SAMRecord contig : CONTIGS) {
if (CONTIG_NAMES == null
|| CONTIG_NAMES.isEmpty()
|| CONTIG_NAMES.contains(contig.getReadName())) {
Map<String, String> te = annotatedContigs.get(contig.getReadName());
if (annotatedContigs.containsKey(contig.getReadName())) {
String seq = contig.getReadString();
// log.debug(" te: {}", te);
String annSeq = te.get("seq");
String kmerOrigin = te.get("kmerOrigin");
Map<CortexKmer, Character> kmerCodes = new HashMap<CortexKmer, Character>();
for (int i = 0; i < kmerOrigin.length(); i++) {
CortexKmer kmer = new CortexKmer(annSeq.substring(i, i + kmerSize));
Character code = kmerOrigin.charAt(i);
kmerCodes.put(kmer, code);
}
Map<Character, Integer> kmerStats = new HashMap<Character, Integer>();
for (Character c : kmerCodeNames.keySet()) {
kmerStats.put(c, 0);
}
boolean changed = false;
// We want to be able to examine soft-clipped regions as well.
List<CigarElement> ces = new ArrayList<CigarElement>();
for (CigarElement ce : contig.getCigar().getCigarElements()) {
if (ce.getOperator().equals(CigarOperator.S)) {
ces.add(new CigarElement(ce.getLength(), CigarOperator.M));
changed = true;
} else {
ces.add(ce);
}
}
if (changed) {
CigarElement firstCe = contig.getCigar().getCigarElements().get(0);
if (firstCe.getOperator().equals(CigarOperator.S)) {
contig.setAlignmentStart(contig.getAlignmentStart() - firstCe.getLength());
}
contig.setCigar(new Cigar(ces));
}
for (AlignmentBlock ab : contig.getAlignmentBlocks()) {
for (int i = ab.getReadStart() - 1; i < ab.getReadStart() + ab.getLength(); i++) {
if (i + kmerSize < seq.length()) {
CortexKmer kmer = new CortexKmer(seq.substring(i, i + kmerSize));
SAMRecord skmer = new SAMRecord(CONTIGS.getFileHeader());
skmer.setReadBases(seq.substring(i, i + kmerSize).getBytes());
List<CigarElement> cigarElements = new ArrayList<CigarElement>();
cigarElements.add(new CigarElement(kmerSize, CigarOperator.M));
Cigar cigar = new Cigar(cigarElements);
skmer.setReadName(contig.getReadName() + "." + kmer.getKmerAsString());
skmer.setReferenceName(contig.getReferenceName());
skmer.setCigar(cigar);
skmer.setReadPairedFlag(false);
skmer.setDuplicateReadFlag(false);
skmer.setMateNegativeStrandFlag(false);
skmer.setAlignmentStart(ab.getReferenceStart() - ab.getReadStart() + 1 + i);
skmer.setAttribute("RG", "none");
skmer.setMappingQuality(0);
Character c = kmerCodes.get(kmer);
String codeName = kmerCodeNames.get(c);
String parentReadGroupId = null;
String sampleReadGroupId = null;
for (SAMReadGroupRecord rgr : sfh.getReadGroups()) {
if (rgr.getSample().equals(codeName)) {
parentReadGroupId = rgr.getReadGroupId();
}
if (rgr.getSample().equals(contig.getReadGroup().getSample())) {
sampleReadGroupId = rgr.getReadGroupId();
}
}
skmer.setAttribute(
"RG", parentReadGroupId != null ? parentReadGroupId : sampleReadGroupId);
skmer.setMappingQuality(99);
sfw.addAlignment(skmer);
kmerStats.put(c, kmerStats.get(c) + 1);
IGVEntry igvEntry = new IGVEntry();
igvEntry.chromosome = contig.getReferenceName();
igvEntry.start = ab.getReferenceStart() - ab.getReadStart() + i;
igvEntry.parentageName = kmerCodeNames.get(c);
igvEntry.parentage = kmerCodeIndices.get(c);
igvEntries.add(igvEntry);
}
}
}
if (!contig.isSecondaryOrSupplementary()) {
beout.println(
contig.getReferenceName()
+ "\t"
+ contig.getAlignmentStart()
+ "\t"
+ contig.getAlignmentEnd()
+ "\t"
+ contig.getReadName()
+ "."
+ contig.getReadGroup().getSample());
if (annotatedContigs.size() > 10 && numContigs % (annotatedContigs.size() / 10) == 0) {
log.info(" processed {}/{} contigs", numContigs, annotatedContigs.size());
}
numContigs++;
}
Map<String, String> stats = new LinkedHashMap<String, String>();
stats.put("contigName", contig.getReadName());
stats.put("sampleName", contig.getReadGroup().getSample());
for (Character c : kmerCodeNames.keySet()) {
stats.put(kmerCodeNames.get(c), String.valueOf(kmerStats.get(c)));
}
tw.addEntry(stats);
}
}
}
log.info("Writing kmer inheritance information...");
out.printf("%s\t%s\t%s\t%s\t%s\n", "Chromosome", "Start", "End", "Feature", "Parentage");
for (IGVEntry igvEntry : igvEntries) {
out.printf(
"%s\t%d\t%d\t%s\t%d\n",
igvEntry.chromosome,
igvEntry.start,
igvEntry.start + 1,
igvEntry.parentageName,
igvEntry.parentage);
}
sfw.close();
}
