/**
* Method that combines the fingerprint evidence across all the read groups for the same sample
* and then produces a matrix of LOD scores for comparing every sample with every other sample.
*/
private void crossCheckSamples(final List<Fingerprint> fingerprints, final PrintStream out) {
final SortedMap<String, Fingerprint> sampleFps =
FingerprintChecker.mergeFingerprintsBySample(fingerprints);
final SortedSet<String> samples = (SortedSet<String>) sampleFps.keySet();
// Print header row
out.print("\t");
for (final String sample : samples) {
out.print(sample);
out.print("\t");
}
out.println();
// Print results rows
for (final String sample : samples) {
out.print(sample);
final Fingerprint fp = sampleFps.get(sample);
for (final String otherSample : samples) {
final MatchResults results =
FingerprintChecker.calculateMatchResults(
fp, sampleFps.get(otherSample), GENOTYPING_ERROR_RATE, LOSS_OF_HET_RATE);
out.print("\t");
out.print(formatUtil.format(results.getLOD()));
}
out.println();
}
}
/**
* Method that pairwise checks every pair of read groups and reports a LOD score for the two read
* groups coming from the same sample.
*/
private int crossCheckReadGroups(
final Map<SAMReadGroupRecord, Fingerprint> fingerprints, final PrintStream out) {
int mismatches = 0;
int unexpectedMatches = 0;
final List<SAMReadGroupRecord> readGroupRecords = new ArrayList<>(fingerprints.keySet());
final List<String> output = new ArrayList<>();
for (int i = 0; i < readGroupRecords.size(); i++) {
final SAMReadGroupRecord lhsRg = readGroupRecords.get(i);
for (int j = i + 1; j < readGroupRecords.size(); j++) {
final SAMReadGroupRecord rhsRg = readGroupRecords.get(j);
final boolean expectedToMatch =
EXPECT_ALL_READ_GROUPS_TO_MATCH || lhsRg.getSample().equals(rhsRg.getSample());
final MatchResults results =
FingerprintChecker.calculateMatchResults(
fingerprints.get(lhsRg),
fingerprints.get(rhsRg),
GENOTYPING_ERROR_RATE,
LOSS_OF_HET_RATE);
if (expectedToMatch) {
if (results.getLOD() < LOD_THRESHOLD) {
mismatches++;
output.add(getMatchDetails(UNEXPECTED_MISMATCH, results, lhsRg, rhsRg));
} else {
if (!OUTPUT_ERRORS_ONLY) {
output.add(getMatchDetails(EXPECTED_MATCH, results, lhsRg, rhsRg));
}
}
} else {
if (results.getLOD() > -LOD_THRESHOLD) {
unexpectedMatches++;
output.add(getMatchDetails(UNEXPECTED_MATCH, results, lhsRg, rhsRg));
} else {
if (!OUTPUT_ERRORS_ONLY) {
output.add(getMatchDetails(EXPECTED_MISMATCH, results, lhsRg, rhsRg));
}
}
}
}
}
if (!output.isEmpty()) {
out.println(
"RESULT\tLOD_SCORE\tLOD_SCORE_TUMOR_NORMAL\tLOD_SCORE_NORMAL_TUMOR\tLEFT_RUN_BARCODE\tLEFT_LANE\tLEFT_MOLECULAR_BARCODE_SEQUENCE\tLEFT_LIBRARY\tLEFT_SAMPLE\t"
+ "RIGHT_RUN_BARCODE\tRIGHT_LANE\tRIGHT_MOLECULAR_BARCODE_SEQUENCE\tRIGHT_LIBRARY\tRIGHT_SAMPLE");
out.println(String.join("\n", output));
}
if (mismatches + unexpectedMatches > 0) {
log.info("WARNING: At least two read groups did not relate as expected.");
return EXIT_CODE_WHEN_MISMATCH;
} else {
log.info("All read groups related as expected.");
return 0;
}
}
@Override
protected int doWork() {
// Check inputs
for (final File f : INPUT) IOUtil.assertFileIsReadable(f);
IOUtil.assertFileIsReadable(HAPLOTYPE_MAP);
if (OUTPUT != null) IOUtil.assertFileIsWritable(OUTPUT);
final HaplotypeMap map = new HaplotypeMap(HAPLOTYPE_MAP);
final FingerprintChecker checker = new FingerprintChecker(map);
checker.setAllowDuplicateReads(ALLOW_DUPLICATE_READS);
log.info("Done checking input files, moving onto fingerprinting files.");
List<File> unrolledFiles =
IOUtil.unrollFiles(INPUT, BamFileIoUtils.BAM_FILE_EXTENSION, IOUtil.SAM_FILE_EXTENSION);
final Map<SAMReadGroupRecord, Fingerprint> fpMap =
checker.fingerprintSamFiles(unrolledFiles, NUM_THREADS, 1, TimeUnit.DAYS);
final List<Fingerprint> fingerprints = new ArrayList<>(fpMap.values());
log.info("Finished generating fingerprints from BAM files, moving on to cross-checking.");
// Setup the output
final PrintStream out;
if (OUTPUT != null) {
out = new PrintStream(IOUtil.openFileForWriting(OUTPUT), true);
} else {
out = System.out;
}
if (this.CROSSCHECK_SAMPLES) {
crossCheckSamples(fingerprints, out);
return 0;
} else if (this.CROSSCHECK_LIBRARIES) {
crossCheckLibraries(fpMap, out);
return 0;
} else {
return crossCheckReadGroups(fpMap, out);
}
}