/** Ensure that splitting read groups still works when dealing with a sample-split pileup. */
@Test
public void testSplitBySample() {
SAMReadGroupRecord readGroupOne = new SAMReadGroupRecord("rg1");
readGroupOne.setSample("sample1");
SAMReadGroupRecord readGroupTwo = new SAMReadGroupRecord("rg2");
readGroupTwo.setSample("sample2");
SAMFileHeader header = ArtificialSAMUtils.createArtificialSamHeader(1, 1, 1000);
header.addReadGroup(readGroupOne);
header.addReadGroup(readGroupTwo);
GATKSAMRecord read1 = ArtificialSAMUtils.createArtificialRead(header, "read1", 0, 1, 10);
read1.setAttribute("RG", readGroupOne.getId());
GATKSAMRecord read2 = ArtificialSAMUtils.createArtificialRead(header, "read2", 0, 1, 10);
read2.setAttribute("RG", readGroupTwo.getId());
GATKSAMRecord read3 = ArtificialSAMUtils.createArtificialRead(header, "read3", 0, 1, 10);
read3.setAttribute("RG", readGroupOne.getId());
GATKSAMRecord read4 = ArtificialSAMUtils.createArtificialRead(header, "read4", 0, 1, 10);
read4.setAttribute("RG", readGroupTwo.getId());
ReadBackedPileupImpl sample1Pileup =
new ReadBackedPileupImpl(null, Arrays.asList(read1, read3), Arrays.asList(1, 1));
ReadBackedPileupImpl sample2Pileup =
new ReadBackedPileupImpl(null, Arrays.asList(read2, read4), Arrays.asList(1, 1));
Map<String, ReadBackedPileupImpl> sampleToPileupMap =
new HashMap<String, ReadBackedPileupImpl>();
sampleToPileupMap.put(readGroupOne.getSample(), sample1Pileup);
sampleToPileupMap.put(readGroupTwo.getSample(), sample2Pileup);
ReadBackedPileup compositePileup = new ReadBackedPileupImpl(null, sampleToPileupMap);
ReadBackedPileup rg1Pileup = compositePileup.getPileupForReadGroup("rg1");
List<GATKSAMRecord> rg1Reads = rg1Pileup.getReads();
Assert.assertEquals(rg1Reads.size(), 2, "Wrong number of reads in read group rg1");
Assert.assertEquals(
rg1Reads.get(0), read1, "Read " + read1.getReadName() + " should be in rg1 but isn't");
Assert.assertEquals(
rg1Reads.get(1), read3, "Read " + read3.getReadName() + " should be in rg1 but isn't");
ReadBackedPileup rg2Pileup = compositePileup.getPileupForReadGroup("rg2");
List<GATKSAMRecord> rg2Reads = rg2Pileup.getReads();
Assert.assertEquals(rg1Reads.size(), 2, "Wrong number of reads in read group rg2");
Assert.assertEquals(
rg2Reads.get(0), read2, "Read " + read2.getReadName() + " should be in rg2 but isn't");
Assert.assertEquals(
rg2Reads.get(1), read4, "Read " + read4.getReadName() + " should be in rg2 but isn't");
}
@Test
public void testGetPileupForSample() {
String sample1 = "sample1";
String sample2 = "sample2";
SAMReadGroupRecord readGroupOne = new SAMReadGroupRecord("rg1");
readGroupOne.setSample(sample1);
SAMReadGroupRecord readGroupTwo = new SAMReadGroupRecord("rg2");
readGroupTwo.setSample(sample2);
SAMFileHeader header = ArtificialSAMUtils.createArtificialSamHeader(1, 1, 1000);
header.addReadGroup(readGroupOne);
header.addReadGroup(readGroupTwo);
GATKSAMRecord read1 = ArtificialSAMUtils.createArtificialRead(header, "read1", 0, 1, 10);
read1.setAttribute("RG", readGroupOne.getId());
GATKSAMRecord read2 = ArtificialSAMUtils.createArtificialRead(header, "read2", 0, 1, 10);
read2.setAttribute("RG", readGroupTwo.getId());
Map<String, ReadBackedPileupImpl> sampleToPileupMap =
new HashMap<String, ReadBackedPileupImpl>();
sampleToPileupMap.put(
sample1, new ReadBackedPileupImpl(null, Collections.singletonList(read1), 0));
sampleToPileupMap.put(
sample2, new ReadBackedPileupImpl(null, Collections.singletonList(read2), 0));
ReadBackedPileup pileup = new ReadBackedPileupImpl(null, sampleToPileupMap);
ReadBackedPileup sample2Pileup = pileup.getPileupForSample(sample2);
Assert.assertEquals(
sample2Pileup.getNumberOfElements(), 1, "Sample 2 pileup has wrong number of elements");
Assert.assertEquals(
sample2Pileup.getReads().get(0), read2, "Sample 2 pileup has incorrect read");
ReadBackedPileup missingSamplePileup = pileup.getPileupForSample("missing");
Assert.assertNull(missingSamplePileup, "Pileup for sample 'missing' should be null but isn't");
missingSamplePileup = pileup.getPileupForSample("not here");
Assert.assertNull(missingSamplePileup, "Pileup for sample 'not here' should be null but isn't");
}
public void setupTest2(
final int ID,
final String readGroupId,
final SAMReadGroupRecord readGroupRecord,
final String sample,
final String library,
final SAMFileHeader header,
final SAMRecordSetBuilder setBuilder)
throws IOException {
final String separator = ":";
final int contig1 = 0;
final int contig2 = 1;
final int contig3 = 2;
readGroupRecord.setSample(sample);
readGroupRecord.setPlatform(platform);
readGroupRecord.setLibrary(library);
readGroupRecord.setPlatformUnit(readGroupId);
setBuilder.setReadGroup(readGroupRecord);
setBuilder.setUseNmFlag(true);
setBuilder.setHeader(header);
final int max = 800;
final int min = 1;
final Random rg = new Random(5);
// add records that align to all 3 chr in reference file
for (int i = 0; i < NUM_READS; i++) {
final int start = rg.nextInt(max) + min;
final String newReadName = READ_NAME + separator + ID + separator + i;
if (i <= NUM_READS / 3) {
setBuilder.addPair(newReadName, contig1, start + ID, start + ID + LENGTH);
} else if (i < (NUM_READS - (NUM_READS / 3))) {
setBuilder.addPair(newReadName, contig2, start + ID, start + ID + LENGTH);
} else {
setBuilder.addPair(newReadName, contig3, start + ID, start + ID + LENGTH);
}
}
}
@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();
}