private static void resetAttributes( SAMRecord rec, List<String> optFieldTags, List<Object> optFieldValues) { ListIterator<String> iterTags = optFieldTags.listIterator(); ListIterator<Object> iterValues = optFieldValues.listIterator(); while (iterTags.hasNext()) { rec.setAttribute(iterTags.next(), iterValues.next()); } }
private static void updateSAM( SAMRecord rec, ReferenceSequence sequence, SAMProgramRecord programRecord, AlignHeapNode bestAlignHeapNode, SRMAUtil.Space space, String read, String qualities, String softClipStartBases, String softClipStartQualities, String softClipEndBases, String softClipEndQualities, boolean strand, boolean correctBases) throws Exception { AlignHeapNode curAlignHeapNode = null; AlignHeapNode prevAlignHeapNode = null; int alignmentStart = 0; int readIndex = -1; byte readBases[] = null; byte baseQualities[] = null; byte colorErrors[] = null; int i; int numEdits = 0; List<String> optFieldTags = new LinkedList<String>(); List<Object> optFieldValues = new LinkedList<Object>(); Object attr; // Debugging stuff String readName = rec.getReadName(); if (null == bestAlignHeapNode) { // Do not modify the alignment return; } // To generate a new CIGAR List<CigarElement> cigarElements = null; CigarOperator prevCigarOperator = null, curCigarOperator = null; int prevCigarOperatorLength = 0; // TODO // setInferredInsertSize (invalidates paired end reads) // setMappingQuality (?) // setFlag // update base qualities for color space reads // clear attributes, but save some Align.clearAttributes(rec, optFieldTags, optFieldValues); readBases = new byte[read.length()]; baseQualities = new byte[qualities.length()]; for (i = 0; i < qualities.length(); i++) { // Must subtract 33 for PHRED scaling baseQualities[i] = (byte) (qualities.charAt(i) - 33); } if (strand) { readIndex = 0; } else { readIndex = read.length() - 1; } cigarElements = new LinkedList<CigarElement>(); if (strand) { // reverse strand is the current position alignmentStart = bestAlignHeapNode.node.position; } else { alignmentStart = bestAlignHeapNode.startPosition; } assert null != bestAlignHeapNode; curAlignHeapNode = bestAlignHeapNode; while (null != curAlignHeapNode) { // Get the current cigar operator if (null != prevAlignHeapNode && CigarOperator.DELETION != prevCigarOperator && 1 < Math.abs(curAlignHeapNode.node.position - prevAlignHeapNode.node.position)) { curCigarOperator = CigarOperator.DELETION; } else { switch (curAlignHeapNode.node.type) { case Node.MISMATCH: // Fall through case Node.MATCH: curCigarOperator = CigarOperator.MATCH_OR_MISMATCH; break; case Node.INSERTION: // System.out.println("INS"); curCigarOperator = CigarOperator.INSERTION; break; default: throw new Exception("Unknown node type"); } if (space == SRMAUtil.Space.COLORSPACE || correctBases) { readBases[readIndex] = (byte) curAlignHeapNode.node.base; if (strand) { readIndex++; } else { readIndex--; } // count the number of mismatches switch (curAlignHeapNode.node.type) { case Node.MISMATCH: case Node.INSERTION: numEdits++; break; default: break; } } else { // count the number of mismatches switch (curAlignHeapNode.node.type) { case Node.MATCH: if (read.charAt(curAlignHeapNode.readOffset) != curAlignHeapNode.node.base) { numEdits++; } break; case Node.MISMATCH: // Fall through if (read.charAt(curAlignHeapNode.readOffset) != sequence.getBases()[curAlignHeapNode.node.position - 1]) { numEdits++; } break; case Node.INSERTION: numEdits++; break; default: break; } } } if (prevCigarOperator != curCigarOperator) { // different cigar operator // add the previous cigar operator if (null != prevCigarOperator) { if (strand) { // reverse // append cigarElements.add(new CigarElement(prevCigarOperatorLength, prevCigarOperator)); } else { // prepend cigarElements.add(0, new CigarElement(prevCigarOperatorLength, prevCigarOperator)); } } // update prevCigarOperator prevCigarOperator = curCigarOperator; if (curCigarOperator == CigarOperator.DELETION) { // length of deletion prevCigarOperatorLength = Math.abs(curAlignHeapNode.node.position - prevAlignHeapNode.node.position) - 1; numEdits += prevCigarOperatorLength; // deletions } else { prevCigarOperatorLength = 1; } } else { // same cigar operator prevCigarOperatorLength++; } // Update if (CigarOperator.DELETION != curCigarOperator) { prevAlignHeapNode = curAlignHeapNode; curAlignHeapNode = curAlignHeapNode.prev; } } if (0 < prevCigarOperatorLength) { if (null == prevCigarOperator || CigarOperator.DELETION == prevCigarOperator) { throw new Exception("Ended with a null cigar operator or a deletion cigar operator"); } if (strand) { // reverse // append cigarElements.add(new CigarElement(prevCigarOperatorLength, prevCigarOperator)); } else { // prepend cigarElements.add(0, new CigarElement(prevCigarOperatorLength, prevCigarOperator)); } } if (space == SRMAUtil.Space.COLORSPACE) { // color space, read bases already inferred // Get color error string colorErrors = new byte[read.length()]; char prevBase = SRMAUtil.COLORSPACE_ADAPTOR; if (strand) { // reverse for (i = 0; i < read.length(); i++) { char nextBase = SRMAUtil.colorSpaceNextBase(prevBase, read.charAt(i)); if (nextBase == SRMAUtil.getCompliment((char) readBases[read.length() - i - 1])) { colorErrors[i] = (byte) Alignment.GAP; } else { colorErrors[i] = (byte) read.charAt(i); } if (0 < i) { // qualities are assumed to be always in the same direction as the color errors baseQualities[read.length() - i] = getColorQuality( colorErrors[i - 1], colorErrors[i], (byte) (qualities.charAt(i - 1) - 33), (byte) (qualities.charAt(i) - 33)); } prevBase = SRMAUtil.getCompliment((char) readBases[read.length() - i - 1]); } // last color baseQualities[0] = (byte) (qualities.charAt(read.length() - 1) - 33); } else { for (i = 0; i < read.length(); i++) { char nextBase = SRMAUtil.colorSpaceNextBase(prevBase, read.charAt(i)); if (nextBase == readBases[i]) { colorErrors[i] = (byte) Alignment.GAP; } else { colorErrors[i] = (byte) read.charAt(i); } if (0 < i) { baseQualities[i - 1] = getColorQuality( colorErrors[i - 1], colorErrors[i], (byte) (qualities.charAt(i - 1) - 33), (byte) (qualities.charAt(i) - 33)); } prevBase = (char) readBases[i]; } // last color baseQualities[read.length() - 1] = (byte) (qualities.charAt(read.length() - 1) - 33); } } else if (correctBases) { // bases were corrected if (strand) { for (i = 0; i < read.length(); i++) { if (readBases[i] == (byte) read.charAt(read.length() - i - 1)) { baseQualities[i] = (byte) (qualities.charAt(read.length() - i - 1) - 33); } else { // TODO: how much to down-weight ? baseQualities[i] = (byte) (SRMAUtil.QUAL2CHAR( SRMAUtil.CHAR2QUAL(qualities.charAt(read.length() - i - 1)) - CORRECT_BASE_QUALITY_PENALTY) - 33); if (baseQualities[i] <= 0) { baseQualities[i] = 1; } } } } else { for (i = 0; i < read.length(); i++) { if (readBases[i] == (byte) read.charAt(i)) { baseQualities[i] = (byte) (qualities.charAt(i) - 33); } else { // TODO: how much to down-weight ? baseQualities[i] = (byte) (SRMAUtil.QUAL2CHAR( SRMAUtil.CHAR2QUAL(qualities.charAt(i)) - CORRECT_BASE_QUALITY_PENALTY) - 33); if (baseQualities[i] <= 0) { baseQualities[i] = 1; } } } } rec.setAttribute("XO", read); rec.setAttribute("XQ", qualities); } else { // bases not corrected readBases = new byte[read.length()]; baseQualities = new byte[qualities.length()]; // qualities.length() == read.length() if (strand) { // reverse for (i = 0; i < read.length(); i++) { readBases[i] = (byte) read.charAt(read.length() - i - 1); baseQualities[i] = (byte) (qualities.charAt(read.length() - i - 1) - 33); } } else { for (i = 0; i < read.length(); i++) { readBases[i] = (byte) read.charAt(i); baseQualities[i] = (byte) (qualities.charAt(i) - 33); } } } // Add in soft-clipping if (null != softClipStartBases) { // prepend cigarElements.add(0, new CigarElement(softClipStartBases.length(), CigarOperator.S)); byte tmpBases[] = new byte[readBases.length + softClipStartBases.length()]; System.arraycopy(readBases, 0, tmpBases, softClipStartBases.length(), readBases.length); readBases = tmpBases; for (i = 0; i < softClipStartBases.length(); i++) { readBases[i] = (byte) softClipStartBases.charAt(i); } byte tmpQualities[] = new byte[baseQualities.length + softClipStartQualities.length()]; System.arraycopy( baseQualities, 0, tmpQualities, softClipStartQualities.length(), baseQualities.length); baseQualities = tmpQualities; for (i = 0; i < softClipStartQualities.length(); i++) { baseQualities[i] = (byte) softClipStartQualities.charAt(i); } } if (null != softClipEndBases) { // append cigarElements.add(new CigarElement(softClipEndBases.length(), CigarOperator.S)); byte tmpBases[] = new byte[readBases.length + softClipEndBases.length()]; System.arraycopy(readBases, 0, tmpBases, 0, readBases.length); for (i = 0; i < softClipEndBases.length(); i++) { tmpBases[i + readBases.length] = (byte) softClipEndBases.charAt(i); } readBases = tmpBases; byte tmpQualities[] = new byte[baseQualities.length + softClipEndQualities.length()]; System.arraycopy(baseQualities, 0, tmpQualities, 0, baseQualities.length); for (i = 0; i < softClipEndQualities.length(); i++) { tmpQualities[i + baseQualities.length] = (byte) softClipEndQualities.charAt(i); } baseQualities = tmpQualities; } // Update SAM record rec.setCigar(new Cigar(cigarElements)); rec.setAlignmentStart(alignmentStart); rec.setReadBases(readBases); rec.setBaseQualities(baseQualities); // Reset saved attributes Align.resetAttributes(rec, optFieldTags, optFieldValues); // Set new attributes if (space == SRMAUtil.Space.COLORSPACE) { // set the XE attribute for colorError string rec.setAttribute("XE", new String(colorErrors)); } rec.setAttribute("AS", bestAlignHeapNode.score); rec.setAttribute("XC", bestAlignHeapNode.alleleCoverageSum); rec.setAttribute("PG", programRecord.getId()); rec.setAttribute("NM", numEdits); }
@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(); }