/**
* Create random read qualities
*
* @param length the length of the read
* @param allowNs whether or not to allow N's in the read
* @return an array with randomized bases (A-N) with equal probability
*/
public static byte[] createRandomReadBases(int length, boolean allowNs) {
Random random = GenomeAnalysisEngine.getRandomGenerator();
int numberOfBases = allowNs ? 5 : 4;
byte[] bases = new byte[length];
for (int i = 0; i < length; i++) {
switch (random.nextInt(numberOfBases)) {
case 0:
bases[i] = 'A';
break;
case 1:
bases[i] = 'C';
break;
case 2:
bases[i] = 'G';
break;
case 3:
bases[i] = 'T';
break;
case 4:
bases[i] = 'N';
break;
default:
throw new ReviewedStingException("Something went wrong, this is just impossible");
}
}
return bases;
}
private void randomlyAddVariant(int rank, VariantContext vc, byte refBase) {
if (nVariantsAdded < numRandom) variantArray[nVariantsAdded++] = new RandomVariantStructure(vc);
else {
double v = GenomeAnalysisEngine.getRandomGenerator().nextDouble();
double t = (1.0 / (rank - numRandom + 1));
if (v < t) {
variantArray[positionToAdd].set(vc);
nVariantsAdded++;
positionToAdd = nextCircularPosition(positionToAdd);
}
}
}
@Test(enabled = true)
public void testGetBasesReverseComplement() {
int iterations = 1000;
Random random = GenomeAnalysisEngine.getRandomGenerator();
while (iterations-- > 0) {
final int l = random.nextInt(1000);
GATKSAMRecord read = GATKSAMRecord.createRandomRead(l);
byte[] original = read.getReadBases();
byte[] reconverted = new byte[l];
String revComp = ReadUtils.getBasesReverseComplement(read);
for (int i = 0; i < l; i++) {
reconverted[l - 1 - i] = BaseUtils.getComplement((byte) revComp.charAt(i));
}
Assert.assertEquals(reconverted, original);
}
}
private void testPerSampleEqualsFlat(
final String bam1,
final String bam2,
final String persampleFile,
final Double downsampling,
final String md5) {
final String command =
baseCommand3
+ " -I "
+ ArtificalBAMLocation
+ bam1
+ " -I "
+ ArtificalBAMLocation
+ bam2
+ " -o %s ";
WalkerTestSpec spec =
new WalkerTestSpec(command + " -contaminationFile " + persampleFile, 1, Arrays.asList(md5));
final Random rnd = GenomeAnalysisEngine.getRandomGenerator();
rnd.setSeed(123451); // so that the two test cases have a hope of giving the same result
executeTest(
"test contamination on Artificial Contamination, with per-sample file on "
+ bam1
+ " and "
+ bam2
+ " with "
+ persampleFile,
spec);
spec =
new WalkerTestSpec(
command + "-contamination " + downsampling.toString(), 1, Arrays.asList(md5));
rnd.setSeed(123451); // so that the two test cases have a hope of giving the same result
executeTest(
"test contamination on Artificial Contamination, with flat contamination on "
+ bam1
+ " and "
+ bam2
+ " with "
+ downsampling.toString(),
spec);
}
public void writeBeagleOutput(
VariantContext preferredVC, VariantContext otherVC, boolean isValidationSite, double prior) {
GenomeLoc currentLoc =
VariantContextUtils.getLocation(getToolkit().getGenomeLocParser(), preferredVC);
StringBuffer beagleOut = new StringBuffer();
String marker = String.format("%s:%d ", currentLoc.getContig(), currentLoc.getStart());
beagleOut.append(marker);
if (markers != null)
markers.append(marker).append("\t").append(Integer.toString(markerCounter++)).append("\t");
for (Allele allele : preferredVC.getAlleles()) {
String bglPrintString;
if (allele.isNoCall() || allele.isNull()) bglPrintString = "-";
else bglPrintString = allele.getBaseString(); // get rid of * in case of reference allele
beagleOut.append(String.format("%s ", bglPrintString));
if (markers != null) markers.append(bglPrintString).append("\t");
}
if (markers != null) markers.append("\n");
GenotypesContext preferredGenotypes = preferredVC.getGenotypes();
GenotypesContext otherGenotypes = goodSite(otherVC) ? otherVC.getGenotypes() : null;
for (String sample : samples) {
boolean isMaleOnChrX = CHECK_IS_MALE_ON_CHR_X && getSample(sample).getGender() == Gender.MALE;
Genotype genotype;
boolean isValidation;
// use sample as key into genotypes structure
if (preferredGenotypes.containsSample(sample)) {
genotype = preferredGenotypes.get(sample);
isValidation = isValidationSite;
} else if (otherGenotypes != null && otherGenotypes.containsSample(sample)) {
genotype = otherGenotypes.get(sample);
isValidation = !isValidationSite;
} else {
// there is magically no genotype for this sample.
throw new StingException(
"Sample "
+ sample
+ " arose with no genotype in variant or validation VCF. This should never happen.");
}
/*
* Use likelihoods if: is validation, prior is negative; or: is not validation, has genotype key
*/
double[] log10Likelihoods = null;
if ((isValidation && prior < 0.0) || genotype.hasLikelihoods()) {
log10Likelihoods = genotype.getLikelihoods().getAsVector();
// see if we need to randomly mask out genotype in this position.
if (GenomeAnalysisEngine.getRandomGenerator().nextDouble() <= insertedNoCallRate) {
// we are masking out this genotype
log10Likelihoods =
isMaleOnChrX ? HAPLOID_FLAT_LOG10_LIKELIHOODS : DIPLOID_FLAT_LOG10_LIKELIHOODS;
}
if (isMaleOnChrX) {
log10Likelihoods[1] = -255; // todo -- warning this is dangerous for multi-allele case
}
}
/** otherwise, use the prior uniformly */
else if (!isValidation && genotype.isCalled() && !genotype.hasLikelihoods()) {
// hack to deal with input VCFs with no genotype likelihoods. Just assume the called
// genotype
// is confident. This is useful for Hapmap and 1KG release VCFs.
double AA = (1.0 - prior) / 2.0;
double AB = (1.0 - prior) / 2.0;
double BB = (1.0 - prior) / 2.0;
if (genotype.isHomRef()) {
AA = prior;
} else if (genotype.isHet()) {
AB = prior;
} else if (genotype.isHomVar()) {
BB = prior;
}
log10Likelihoods = MathUtils.toLog10(new double[] {AA, isMaleOnChrX ? 0.0 : AB, BB});
} else {
log10Likelihoods =
isMaleOnChrX ? HAPLOID_FLAT_LOG10_LIKELIHOODS : DIPLOID_FLAT_LOG10_LIKELIHOODS;
}
writeSampleLikelihoods(beagleOut, preferredVC, log10Likelihoods);
}
beagleWriter.println(beagleOut.toString());
}
/**
* Create random read qualities
*
* @param length the length of the read
* @return an array with randomized base qualities between 0 and 50
*/
public static byte[] createRandomReadQuals(int length) {
Random random = GenomeAnalysisEngine.getRandomGenerator();
byte[] quals = new byte[length];
for (int i = 0; i < length; i++) quals[i] = (byte) random.nextInt(50);
return quals;
}
/**
* Subset VC record if necessary and emit the modified record (provided it satisfies criteria for
* printing)
*
* @param tracker the ROD tracker
* @param ref reference information
* @param context alignment info
* @return 1 if the record was printed to the output file, 0 if otherwise
*/
@Override
public Integer map(RefMetaDataTracker tracker, ReferenceContext ref, AlignmentContext context) {
if (tracker == null) return 0;
Collection<VariantContext> vcs =
tracker.getValues(variantCollection.variants, context.getLocation());
if (vcs == null || vcs.size() == 0) {
return 0;
}
for (VariantContext vc : vcs) {
if (MENDELIAN_VIOLATIONS) {
boolean foundMV = false;
for (MendelianViolation mv : mvSet) {
if (mv.isViolation(vc)) {
foundMV = true;
// System.out.println(vc.toString());
if (outMVFile != null)
outMVFileStream.format(
"MV@%s:%d. REF=%s, ALT=%s, AC=%d, momID=%s, dadID=%s, childID=%s, momG=%s, momGL=%s, dadG=%s, dadGL=%s, "
+ "childG=%s childGL=%s\n",
vc.getChr(),
vc.getStart(),
vc.getReference().getDisplayString(),
vc.getAlternateAllele(0).getDisplayString(),
vc.getChromosomeCount(vc.getAlternateAllele(0)),
mv.getSampleMom(),
mv.getSampleDad(),
mv.getSampleChild(),
vc.getGenotype(mv.getSampleMom()).toBriefString(),
vc.getGenotype(mv.getSampleMom()).getLikelihoods().getAsString(),
vc.getGenotype(mv.getSampleDad()).toBriefString(),
vc.getGenotype(mv.getSampleMom()).getLikelihoods().getAsString(),
vc.getGenotype(mv.getSampleChild()).toBriefString(),
vc.getGenotype(mv.getSampleChild()).getLikelihoods().getAsString());
}
}
if (!foundMV) break;
}
if (DISCORDANCE_ONLY) {
Collection<VariantContext> compVCs =
tracker.getValues(discordanceTrack, context.getLocation());
if (!isDiscordant(vc, compVCs)) return 0;
}
if (CONCORDANCE_ONLY) {
Collection<VariantContext> compVCs =
tracker.getValues(concordanceTrack, context.getLocation());
if (!isConcordant(vc, compVCs)) return 0;
}
if (alleleRestriction.equals(NumberAlleleRestriction.BIALLELIC) && !vc.isBiallelic())
continue;
if (alleleRestriction.equals(NumberAlleleRestriction.MULTIALLELIC) && vc.isBiallelic())
continue;
if (!selectedTypes.contains(vc.getType())) continue;
VariantContext sub = subsetRecord(vc, samples);
if ((sub.isPolymorphic() || !EXCLUDE_NON_VARIANTS)
&& (!sub.isFiltered() || !EXCLUDE_FILTERED)) {
for (VariantContextUtils.JexlVCMatchExp jexl : jexls) {
if (!VariantContextUtils.match(sub, jexl)) {
return 0;
}
}
if (SELECT_RANDOM_NUMBER) {
randomlyAddVariant(++variantNumber, sub, ref.getBase());
} else if (!SELECT_RANDOM_FRACTION
|| (GenomeAnalysisEngine.getRandomGenerator().nextDouble() < fractionRandom)) {
vcfWriter.add(sub);
}
}
}
return 1;
}
