@Test(description = "Write SAM records with null SAMFileHeader")
public void samNullHeaderRoundTrip() throws Exception {
final File input = new File(TEST_DATA_DIR, "roundtrip.sam");
final SamReader reader = SamReaderFactory.makeDefault().open(input);
final File outputFile = File.createTempFile("nullheader-out", ".sam");
outputFile.delete();
outputFile.deleteOnExit();
FileOutputStream os = new FileOutputStream(outputFile);
final SAMFileWriterFactory factory = new SAMFileWriterFactory();
final SAMFileWriter writer = factory.makeSAMWriter(reader.getFileHeader(), false, os);
for (SAMRecord rec : reader) {
rec.setHeader(null);
writer.addAlignment(rec);
}
writer.close();
os.close();
InputStream is = new FileInputStream(input);
String originalsam = IOUtil.readFully(is);
is.close();
is = new FileInputStream(outputFile);
String writtensam = IOUtil.readFully(is);
is.close();
Assert.assertEquals(writtensam, originalsam);
}
@Test(
description =
"Read and then write SAM to verify header attribute ordering does not change depending on JVM version")
public void samRoundTrip() throws Exception {
final File input = new File(TEST_DATA_DIR, "roundtrip.sam");
final SamReader reader = SamReaderFactory.makeDefault().open(input);
final File outputFile = File.createTempFile("roundtrip-out", ".sam");
outputFile.delete();
outputFile.deleteOnExit();
FileOutputStream os = new FileOutputStream(outputFile);
final SAMFileWriterFactory factory = new SAMFileWriterFactory();
final SAMFileWriter writer = factory.makeSAMWriter(reader.getFileHeader(), false, os);
for (SAMRecord rec : reader) {
writer.addAlignment(rec);
}
writer.close();
os.close();
InputStream is = new FileInputStream(input);
String originalsam = IOUtil.readFully(is);
is.close();
is = new FileInputStream(outputFile);
String writtensam = IOUtil.readFully(is);
is.close();
Assert.assertEquals(writtensam, originalsam);
}
@Test(dataProvider = "data")
public void testProgramGroupAndReadGroupMerge(File inputFiles[], File expectedOutputFile)
throws IOException {
BufferedReader reader = new BufferedReader(new FileReader(expectedOutputFile));
String line;
String expected_output = "";
while ((line = reader.readLine()) != null) {
expected_output += line + "\n";
}
final List<SAMFileReader> readers = new ArrayList<SAMFileReader>();
final List<SAMFileHeader> headers = new ArrayList<SAMFileHeader>();
for (final File inFile : inputFiles) {
IOUtil.assertFileIsReadable(inFile);
final SAMFileReader in = new SAMFileReader(inFile);
// We are now checking for zero-length reads, so suppress complaint about that.
in.setValidationStringency(ValidationStringency.SILENT);
readers.add(in);
headers.add(in.getFileHeader());
}
final MergingSamRecordIterator iterator;
final SamFileHeaderMerger headerMerger =
new SamFileHeaderMerger(SAMFileHeader.SortOrder.coordinate, headers, true);
iterator = new MergingSamRecordIterator(headerMerger, readers, false);
ByteArrayOutputStream baos = new ByteArrayOutputStream();
SAMFileWriter writer =
new SAMFileWriterFactory().makeSAMWriter(headerMerger.getMergedHeader(), true, baos);
while (iterator.hasNext()) {
writer.addAlignment(iterator.next());
}
writer.close();
String actual_output = StringUtil.bytesToString(baos.toByteArray());
List<String> actual = Arrays.asList(actual_output.split("\\n"));
List<String> expected = Arrays.asList(expected_output.split("\\n"));
for (int i = 0; i < expected.size(); i++) {
if (expected.get(i).startsWith("@")) {
Assert.assertTrue(headersEquivalent(actual.get(i), expected.get(i)));
} else {
List<String> expectedSamParts = Arrays.asList(expected.get(i).split("\\s*"));
List<String> actualSamParts = Arrays.asList(actual.get(i).split("\\s*"));
for (String exp : expectedSamParts) {
Assert.assertTrue(actualSamParts.contains(exp));
}
for (String act : actualSamParts) {
Assert.assertTrue(expectedSamParts.contains(act));
}
}
}
}
private void createSmallBam(final File outputFile) {
final SAMFileWriterFactory factory = new SAMFileWriterFactory();
factory.setCreateIndex(true);
factory.setCreateMd5File(true);
final SAMFileHeader header = new SAMFileHeader();
// index only created if coordinate sorted
header.setSortOrder(SAMFileHeader.SortOrder.coordinate);
header.addSequence(new SAMSequenceRecord("chr1", 123));
final SAMFileWriter writer = factory.makeBAMWriter(header, false, outputFile);
fillSmallBam(writer);
writer.close();
}
private void createSmallBamToOutputStream(final OutputStream outputStream, boolean binary) {
final SAMFileWriterFactory factory = new SAMFileWriterFactory();
factory.setCreateIndex(false);
factory.setCreateMd5File(false);
final SAMFileHeader header = new SAMFileHeader();
// index only created if coordinate sorted
header.setSortOrder(SAMFileHeader.SortOrder.coordinate);
header.addSequence(new SAMSequenceRecord("chr1", 123));
final SAMFileWriter writer =
(binary
? factory.makeBAMWriter(header, false, outputStream)
: factory.makeSAMWriter(header, false, outputStream));
fillSmallBam(writer);
writer.close();
}
private void createBAM(final List<GATKSAMRecord> reads) throws IOException {
testBAM = File.createTempFile("TraverseActiveRegionsUnitTest", ".bam");
testBAM.deleteOnExit();
SAMFileWriter out =
new SAMFileWriterFactory()
.setCreateIndex(true)
.makeBAMWriter(reads.get(0).getHeader(), true, testBAM);
for (GATKSAMRecord read : reads) {
out.addAlignment(read);
}
out.close();
new File(testBAM.getAbsolutePath().replace(".bam", ".bai")).deleteOnExit();
new File(testBAM.getAbsolutePath() + ".bai").deleteOnExit();
}
protected int doWork() {
IoUtil.assertFileIsReadable(INPUT);
IoUtil.assertFileIsWritable(OUTPUT);
final SAMFileReader in = new SAMFileReader(INPUT);
// create the read group we'll be using
final SAMReadGroupRecord rg = new SAMReadGroupRecord(RGID);
rg.setLibrary(RGLB);
rg.setPlatform(RGPL);
rg.setSample(RGSM);
rg.setPlatformUnit(RGPU);
if (RGCN != null) rg.setSequencingCenter(RGCN);
if (RGDS != null) rg.setDescription(RGDS);
if (RGDT != null) rg.setRunDate(RGDT);
log.info(
String.format(
"Created read group ID=%s PL=%s LB=%s SM=%s%n",
rg.getId(), rg.getPlatform(), rg.getLibrary(), rg.getSample()));
// create the new header and output file
final SAMFileHeader inHeader = in.getFileHeader();
final SAMFileHeader outHeader = inHeader.clone();
outHeader.setReadGroups(Arrays.asList(rg));
if (SORT_ORDER != null) outHeader.setSortOrder(SORT_ORDER);
final SAMFileWriter outWriter =
new SAMFileWriterFactory()
.makeSAMOrBAMWriter(
outHeader, outHeader.getSortOrder() == inHeader.getSortOrder(), OUTPUT);
final ProgressLogger progress = new ProgressLogger(log);
for (final SAMRecord read : in) {
read.setAttribute(SAMTag.RG.name(), RGID);
outWriter.addAlignment(read);
progress.record(read);
}
// cleanup
in.close();
outWriter.close();
return 0;
}
@Test
public void testBamIntegers() throws IOException {
final List<String> errorMessages = new ArrayList<String>();
final SamReader bamReader = SamReaderFactory.makeDefault().open(BAM_INPUT);
final File bamOutput = File.createTempFile("test", ".bam");
final File samOutput = File.createTempFile("test", ".sam");
final SAMFileWriter samWriter =
new SAMFileWriterFactory().makeWriter(bamReader.getFileHeader(), true, samOutput, null);
final SAMFileWriter bamWriter =
new SAMFileWriterFactory().makeWriter(bamReader.getFileHeader(), true, bamOutput, null);
final SAMRecordIterator iterator = bamReader.iterator();
while (iterator.hasNext()) {
try {
final SAMRecord rec = iterator.next();
samWriter.addAlignment(rec);
bamWriter.addAlignment(rec);
} catch (final Throwable e) {
System.out.println(e.getMessage());
errorMessages.add(e.getMessage());
}
}
CloserUtil.close(bamReader);
samWriter.close();
bamWriter.close();
Assert.assertEquals(errorMessages.size(), 0);
bamOutput.deleteOnExit();
samOutput.deleteOnExit();
}
public void close() {
for (SAMFileWriter w : writerMap.values()) w.close();
}
protected void PicardPreprocess(
Context context, PreprocessingTools tools, SAMRecordIterator input, String output)
throws InterruptedException, QualityException, IOException, URISyntaxException {
outHeader = header.clone();
outHeader.setSortOrder(SAMFileHeader.SortOrder.coordinate);
// tmp files
String tmpOut1 = tmpFileBase + "-p1.bam";
String tmpOut2 = tmpFileBase + "-p2.bam";
String tmpOut3 = tmpFileBase + "-p3.sam";
String fCounts = tmpFileBase + "-features.count";
String tmpMetrics = tmpFileBase + "-p3-metrics.txt";
SAMFileWriterFactory factory = new SAMFileWriterFactory();
if (!inputIsBam) {
outHeader.addReadGroup(bamrg);
}
SAMFileWriter writer = factory.makeBAMWriter(outHeader, true, new File(tmpOut1));
long startTime = System.currentTimeMillis();
int count = 0;
SAMRecord sam;
while (input.hasNext()) {
sam = input.next();
writer.addAlignment(sam);
count++;
}
int reads = input.getCount();
writer.close();
context.getCounter(HalvadeCounters.IN_PREP_READS).increment(reads);
long estimatedTime = System.currentTimeMillis() - startTime;
context.getCounter(HalvadeCounters.TIME_HADOOP_SAMTOBAM).increment(estimatedTime);
Logger.DEBUG("time writing " + count + " records to disk: " + estimatedTime / 1000);
Logger.DEBUG("clean sam");
context.setStatus("clean sam");
tools.runCleanSam(tmpOut1, tmpOut2);
Logger.DEBUG("mark duplicates");
context.setStatus("mark duplicates");
tools.runMarkDuplicates(tmpOut2, tmpOut3, tmpMetrics);
if (gff != null) {
// tmpOut3 is sam for htseq count!
Logger.DEBUG("featureCounts");
context.setStatus("featureCounts");
tools.runFeatureCounts(gff, tmpOut3, fCounts, threads);
HalvadeFileUtils.uploadFileToHDFS(
context,
FileSystem.get(new URI(outputdir), context.getConfiguration()),
fCounts,
outputdir + context.getTaskAttemptID().toString() + ".count");
}
if (!inputIsBam) {
Logger.DEBUG("add read-group");
context.setStatus("add read-group");
tools.runAddOrReplaceReadGroups(tmpOut3, output, RGID, RGLB, RGPL, RGPU, RGSM);
} else {
context.setStatus("convert SAM to BAM");
Logger.DEBUG("convert SAM to BAM");
tools.callSAMToBAM(tmpOut3, output, threads);
}
Logger.DEBUG("build bam index");
context.setStatus("build bam index");
tools.runBuildBamIndex(output);
estimatedTime = System.currentTimeMillis() - startTime;
Logger.DEBUG("estimated time: " + estimatedTime / 1000);
// remove all temporary files now!
HalvadeFileUtils.removeLocalFile(keep, tmpMetrics, context, HalvadeCounters.FOUT_GATK_TMP);
HalvadeFileUtils.removeLocalFile(keep, tmpOut1, context, HalvadeCounters.FOUT_GATK_TMP);
HalvadeFileUtils.removeLocalFile(keep, tmpOut2, context, HalvadeCounters.FOUT_GATK_TMP);
HalvadeFileUtils.removeLocalFile(keep, tmpOut3, context, HalvadeCounters.FOUT_GATK_TMP);
HalvadeFileUtils.removeLocalFile(keep, fCounts);
}
private void fillSmallBam(SAMFileWriter writer) {
final SAMRecordSetBuilder builder = new SAMRecordSetBuilder();
builder.addUnmappedFragment("HiMom!");
for (final SAMRecord rec : builder.getRecords()) writer.addAlignment(rec);
}
@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();
}