/** Initialize the stratifications, evaluations, evaluation contexts, and reporting object */
public void initialize() {
// Just list the modules, and exit quickly.
if (LIST) {
variantEvalUtils.listModulesAndExit();
}
// maintain the full list of comps
comps.addAll(compsProvided);
if (dbsnp.dbsnp.isBound()) {
comps.add(dbsnp.dbsnp);
knowns.add(dbsnp.dbsnp);
}
// Add a dummy comp track if none exists
if (comps.size() == 0)
comps.add(
new RodBinding<VariantContext>(VariantContext.class, "none", "UNBOUND", "", new Tags()));
// Set up set of additional knowns
for (RodBinding<VariantContext> compRod : comps) {
if (KNOWN_NAMES.contains(compRod.getName())) knowns.add(compRod);
}
// Now that we have all the rods categorized, determine the sample list from the eval rods.
Map<String, VCFHeader> vcfRods = GATKVCFUtils.getVCFHeadersFromRods(getToolkit(), evals);
Set<String> vcfSamples =
SampleUtils.getSampleList(
vcfRods, GATKVariantContextUtils.GenotypeMergeType.REQUIRE_UNIQUE);
// Load the sample list, using an intermediate tree set to sort the samples
final Set<String> allSampleNames = SampleUtils.getSamplesFromCommandLineInput(vcfSamples);
sampleNamesForEvaluation.addAll(
new TreeSet<String>(
SampleUtils.getSamplesFromCommandLineInput(vcfSamples, SAMPLE_EXPRESSIONS)));
isSubsettingSamples = !sampleNamesForEvaluation.containsAll(allSampleNames);
if (Arrays.asList(STRATIFICATIONS_TO_USE).contains("Sample")) {
sampleNamesForStratification.addAll(sampleNamesForEvaluation);
}
sampleNamesForStratification.add(ALL_SAMPLE_NAME);
// Initialize select expressions
for (VariantContextUtils.JexlVCMatchExp jexl :
VariantContextUtils.initializeMatchExps(SELECT_NAMES, SELECT_EXPS)) {
SortableJexlVCMatchExp sjexl = new SortableJexlVCMatchExp(jexl.name, jexl.exp);
jexlExpressions.add(sjexl);
}
// Initialize the set of stratifications and evaluations to use
// The list of stratifiers and evaluators to use
final List<VariantStratifier> stratificationObjects =
variantEvalUtils.initializeStratificationObjects(
NO_STANDARD_STRATIFICATIONS, STRATIFICATIONS_TO_USE);
final Set<Class<? extends VariantEvaluator>> evaluationClasses =
variantEvalUtils.initializeEvaluationObjects(NO_STANDARD_MODULES, MODULES_TO_USE);
checkForIncompatibleEvaluatorsAndStratifiers(stratificationObjects, evaluationClasses);
for (VariantStratifier vs : stratificationObjects) {
if (vs.getName().equals("Filter")) byFilterIsEnabled = true;
else if (vs.getName().equals("Sample")) perSampleIsEnabled = true;
}
if (intervalsFile != null) {
boolean fail = true;
for (final VariantStratifier vs : stratificationObjects) {
if (vs.getClass().equals(IntervalStratification.class)) fail = false;
}
if (fail)
throw new UserException.BadArgumentValue(
"ST", "stratIntervals argument provided but -ST IntervalStratification not provided");
}
// Initialize the evaluation contexts
createStratificationStates(stratificationObjects, evaluationClasses);
// Load ancestral alignments
if (ancestralAlignmentsFile != null) {
try {
ancestralAlignments = new IndexedFastaSequenceFile(ancestralAlignmentsFile);
} catch (FileNotFoundException e) {
throw new ReviewedStingException(
String.format(
"The ancestral alignments file, '%s', could not be found",
ancestralAlignmentsFile.getAbsolutePath()));
}
}
// initialize CNVs
if (knownCNVsFile != null) {
knownCNVsByContig = createIntervalTreeByContig(knownCNVsFile);
}
}
public void initialize() {
for (final Tranche t : Tranche.readTranches(TRANCHES_FILE)) {
if (t.ts >= TS_FILTER_LEVEL) {
tranches.add(t);
}
logger.info(String.format("Read tranche " + t));
}
Collections.reverse(
tranches); // this algorithm wants the tranches ordered from best (lowest truth sensitivity)
// to worst (highest truth sensitivity)
for (final RodBinding rod : input) {
inputNames.add(rod.getName());
}
if (IGNORE_INPUT_FILTERS != null) {
ignoreInputFilterSet.addAll(Arrays.asList(IGNORE_INPUT_FILTERS));
}
// setup the header fields
final Set<VCFHeaderLine> hInfo = new HashSet<VCFHeaderLine>();
hInfo.addAll(GATKVCFUtils.getHeaderFields(getToolkit(), inputNames));
addVQSRStandardHeaderLines(hInfo);
final TreeSet<String> samples = new TreeSet<String>();
samples.addAll(SampleUtils.getUniqueSamplesFromRods(getToolkit(), inputNames));
if (tranches.size() >= 2) {
for (int iii = 0; iii < tranches.size() - 1; iii++) {
final Tranche t = tranches.get(iii);
hInfo.add(
new VCFFilterHeaderLine(
t.name,
String.format(
"Truth sensitivity tranche level for "
+ t.model.toString()
+ " model at VQS Lod: "
+ t.minVQSLod
+ " <= x < "
+ tranches.get(iii + 1).minVQSLod)));
}
}
if (tranches.size() >= 1) {
hInfo.add(
new VCFFilterHeaderLine(
tranches.get(0).name + "+",
String.format(
"Truth sensitivity tranche level for "
+ tranches.get(0).model.toString()
+ " model at VQS Lod < "
+ tranches.get(0).minVQSLod)));
} else {
throw new UserException(
"No tranches were found in the file or were above the truth sensitivity filter level "
+ TS_FILTER_LEVEL);
}
logger.info("Keeping all variants in tranche " + tranches.get(tranches.size() - 1));
final VCFHeader vcfHeader = new VCFHeader(hInfo, samples);
vcfWriter.writeHeader(vcfHeader);
}