BaseQualityHistogramAnnotation

src/main/java/org/broadinstitute/hellbender/tools/walkers/annotator/BaseQuality.java

@@ -40,6 +40,10 @@ protected int aggregate(final List<Integer> values) {
 
     @Override
     protected OptionalInt getValueForRead(final GATKRead read, final VariantContext vc) {
+        return getBaseQuality(read, vc);
+    }
+
+    public static OptionalInt getBaseQuality(final GATKRead read, final VariantContext vc) {
         if (vc.getStart() < read.getStart() || read.getEnd() < vc.getStart()) {
             return OptionalInt.empty();
         }

src/main/java/org/broadinstitute/hellbender/tools/walkers/annotator/BaseQualityHistogram.java

@@ -0,0 +1,72 @@
+package org.broadinstitute.hellbender.tools.walkers.annotator;
+
+import com.google.common.collect.ImmutableMap;
+import com.google.common.collect.Multiset;
+import com.google.common.collect.TreeMultiset;
+import htsjdk.variant.variantcontext.Allele;
+import htsjdk.variant.variantcontext.VariantContext;
+import htsjdk.variant.vcf.VCFHeaderLineCount;
+import htsjdk.variant.vcf.VCFHeaderLineType;
+import htsjdk.variant.vcf.VCFInfoHeaderLine;
+import org.apache.commons.lang.mutable.MutableInt;
+import org.broadinstitute.hellbender.engine.ReferenceContext;
+import org.broadinstitute.hellbender.utils.QualityUtils;
+import org.broadinstitute.hellbender.utils.Utils;
+import org.broadinstitute.hellbender.utils.genotyper.ReadLikelihoods;
+import org.broadinstitute.hellbender.utils.read.GATKRead;
+import scala.Int;
+
+import java.util.*;
+import java.util.stream.Collectors;
+
+public class BaseQualityHistogram extends InfoFieldAnnotation {
+
+    public static final String KEY = "BQHIST";
+
+    public Map<String, Object> annotate(final ReferenceContext ref,
+                                        final VariantContext vc,
+                                        final ReadLikelihoods<Allele> likelihoods) {
+        Utils.nonNull(vc);
+        if ( likelihoods == null ) {
+            return Collections.emptyMap();
+        }
+
+        final Map<Allele, TreeMultiset<Integer>> values = likelihoods.alleles().stream()
+                .collect(Collectors.toMap(a -> a, a -> TreeMultiset.create()));
+
+        Utils.stream(likelihoods.bestAllelesBreakingTies())
+                .filter(ba -> ba.isInformative() && isUsableRead(ba.read))
+                .forEach(ba -> BaseQuality.getBaseQuality(ba.read, vc).ifPresent(v -> values.get(ba.allele).add(v)));
+
+
+        final List<Integer> distinctBaseQualities = likelihoods.alleles().stream()
+                .flatMap(a -> values.get(a).stream())
+                .distinct()
+                .sorted()
+                .collect(Collectors.toList());
+
+        final List<Integer> output = new ArrayList<>();
+
+        for (final int qual : distinctBaseQualities) {
+            output.add(qual);
+            likelihoods.alleles().forEach(allele -> output.add(values.get(allele).count(qual)));
+        }
+
+
+        return ImmutableMap.of(KEY, output);
+    }
+
+    @Override
+    public List<VCFInfoHeaderLine> getDescriptions() {
+        return Arrays.asList(new VCFInfoHeaderLine(KEY, VCFHeaderLineCount.A, VCFHeaderLineType.Integer,
+                "Base quality counts for each allele represented sparsely as alternating entries of qualities and counts for each allele." +
+                "For example [10,1,0,20,0,1] means one ref base with quality 10 and one alt base with quality 20."));
+    }
+
+    @Override
+    public List<String> getKeyNames() { return Arrays.asList(KEY); }
+
+    private static boolean isUsableRead(final GATKRead read) {
+        return read.getMappingQuality() != 0 && read.getMappingQuality() != QualityUtils.MAPPING_QUALITY_UNAVAILABLE;
+    }
+}