/**
* Provides the next record from the underlying iterator after applying filter strings generated
* by the set of filters in use by the iterator.
*/
@Override
public VariantContext next() {
final VariantContext ctx = this.iterator.next();
final Set<String> filterStrings = new HashSet<String>();
// Collect variant level filters
for (final VariantFilter filter : this.filters) {
final String val = filter.filter(ctx);
if (val != null) filterStrings.add(val);
}
// Collect genotype level filters in a Map of Sample -> List<filter string>
final ListMap<String, String> gtFilterStrings = new ListMap<String, String>();
final Set<String> variantSamples = new HashSet<String>();
for (final Genotype gt : ctx.getGenotypes()) {
if (gt.isCalled() && !gt.isHomRef()) variantSamples.add(gt.getSampleName());
for (final GenotypeFilter filter : gtFilters) {
final String filterString = filter.filter(ctx, gt);
if (filterString != null) gtFilterStrings.add(gt.getSampleName(), filterString);
}
}
// If all genotypes are filtered apply a site level filter
if (gtFilterStrings.keySet().containsAll(variantSamples)) {
filterStrings.add(ALL_GTS_FILTERED);
}
// Make a builder and set the site level filter appropriately
final VariantContextBuilder builder = new VariantContextBuilder(ctx);
if (filterStrings.isEmpty()) {
builder.passFilters();
} else {
builder.filters(filterStrings);
}
// Apply filters to the necessary genotypes
builder.noGenotypes();
final List<Genotype> newGenotypes = new ArrayList<Genotype>(ctx.getNSamples());
for (final Genotype gt : ctx.getGenotypes()) {
final GenotypeBuilder gtBuilder = new GenotypeBuilder(gt);
final List<String> filtersLocal = gtFilterStrings.get(gt.getSampleName());
if (filtersLocal == null || filtersLocal.isEmpty()) {
gtBuilder.filter(PASS_FILTER);
} else {
gtBuilder.filters(filtersLocal);
}
newGenotypes.add(gtBuilder.make());
}
builder.genotypes(newGenotypes);
return builder.make();
}
/**
* Takes the interval, finds it in the stash, prints it to the VCF
*
* @param stats The statistics of the interval
* @param refAllele the reference allele
*/
private void outputStatsToVCF(final IntervalStratification stats, final Allele refAllele) {
GenomeLoc interval = stats.getInterval();
final List<Allele> alleles = new ArrayList<>();
final Map<String, Object> attributes = new HashMap<>();
final ArrayList<Genotype> genotypes = new ArrayList<>();
for (String sample : samples) {
final GenotypeBuilder gb = new GenotypeBuilder(sample);
SampleStratification sampleStat = stats.getSampleStatistics(sample);
gb.attribute(
GATKVCFConstants.AVG_INTERVAL_DP_BY_SAMPLE_KEY,
sampleStat.averageCoverage(interval.size()));
gb.attribute(GATKVCFConstants.LOW_COVERAGE_LOCI, sampleStat.getNLowCoveredLoci());
gb.attribute(GATKVCFConstants.ZERO_COVERAGE_LOCI, sampleStat.getNUncoveredLoci());
gb.filters(statusToStrings(stats.getSampleStatistics(sample).callableStatuses(), false));
genotypes.add(gb.make());
}
alleles.add(refAllele);
alleles.add(SYMBOLIC_ALLELE);
VariantContextBuilder vcb =
new VariantContextBuilder(
"DiagnoseTargets",
interval.getContig(),
interval.getStart(),
interval.getStop(),
alleles);
vcb = vcb.log10PError(VariantContext.NO_LOG10_PERROR);
vcb.filters(new LinkedHashSet<>(statusToStrings(stats.callableStatuses(), true)));
attributes.put(VCFConstants.END_KEY, interval.getStop());
attributes.put(GATKVCFConstants.AVG_INTERVAL_DP_KEY, stats.averageCoverage(interval.size()));
attributes.put(GATKVCFConstants.INTERVAL_GC_CONTENT_KEY, stats.gcContent());
vcb = vcb.attributes(attributes);
vcb = vcb.genotypes(genotypes);
vcfWriter.add(vcb.make());
}
