Adding support to annotate CNVs with DoC (#73)

Closes: #73
Related-Issue: #73
Projected-Results-Impact: none

.gitattributes

@@ -2,3 +2,4 @@
 /tests/**/*.ser filter=lfs diff=lfs merge=lfs -text
 /tests/**/*.vcf* filter=lfs diff=lfs merge=lfs -text
 /varfish-annotator-cli/src/test/resources/input/real-world-37/* filter=lfs diff=lfs merge=lfs -text
+/varfish-annotator-cli/src/test/resources/input/*/*.vcf* filter=lfs diff=lfs merge=lfs -text

CHANGES.md

@@ -2,6 +2,9 @@
 
 ## v0.27 (work in progress)
 
+- Writing out proper SV type for Dragen CNV (#76)
+- Adding support for depth of coverage annotation (#73)
+
 ## v0.26
 
 - Explicitely model support for SV callers (#68)

varfish-annotator-cli/src/main/java/com/github/bihealth/varfish_annotator/annotate_svs/AnnotateSvsArgs.java

@@ -2,6 +2,8 @@
 
 import com.beust.jcommander.Parameter;
 import com.beust.jcommander.Parameters;
+import java.util.ArrayList;
+import java.util.List;
 
 /**
  * JCommander command for <tt>varfish_annotator annotate-sv</tt>.
@@ -110,6 +112,12 @@ public final class AnnotateSvsArgs {
           "Skip records with this filter value (comma-separated list), default is 'LowQual'")
   private String skipFilters = "LowQual";
 
+  @Parameter(
+      names = "--coverage-vcf",
+      description =
+          "Annotate CNV with coverage and mapping quality from maelstrom-core coverage VCF file")
+  private List<String> coverageVcfs = new ArrayList<>();
+
   public String getRefseqSerPath() {
     return refseqSerPath;
   }
@@ -186,6 +194,10 @@ public String getSkipFilters() {
     return skipFilters;
   }
 
+  public List<String> getCoverageVcfs() {
+    return coverageVcfs;
+  }
+
   @Override
   public String toString() {
     return "AnnotateSvsArgs{"
@@ -236,16 +248,17 @@ public String toString() {
         + '\''
         + ", sequentialUuids="
         + sequentialUuids
-        + '\''
-        + ", optOutFeatures=\'"
+        + ", optOutFeatures='"
         + optOutFeatures
-        + "\'"
-        + ", writeBndMates=\'"
+        + '\''
+        + ", writeBndMates='"
         + writeBndMates
-        + "\'"
-        + ", skipFilters=\'"
+        + '\''
+        + ", skipFilters='"
         + skipFilters
-        + "\'"
+        + '\''
+        + ", coverageVcfs="
+        + coverageVcfs
         + '}';
   }
 }

varfish-annotator-cli/src/main/java/com/github/bihealth/varfish_annotator/annotate_svs/AnnotateSvsVcf.java

@@ -11,6 +11,7 @@
 import com.google.common.collect.ImmutableList;
 import com.google.common.collect.ImmutableMap;
 import com.google.common.collect.ImmutableSet;
+import com.google.common.io.Closer;
 import de.charite.compbio.jannovar.annotation.SVAnnotation;
 import de.charite.compbio.jannovar.annotation.SVAnnotations;
 import de.charite.compbio.jannovar.data.JannovarData;
@@ -36,10 +37,7 @@
 import java.nio.file.Files;
 import java.nio.file.Paths;
 import java.sql.*;
-import java.util.HashSet;
-import java.util.List;
-import java.util.Set;
-import java.util.UUID;
+import java.util.*;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 
@@ -67,14 +65,10 @@ public final class AnnotateSvsVcf {
   /** Pedigree to use for annotation. */
   private Pedigree pedigree;
 
-  /** Helper to use for creating genotypes and feature effects files. */
-  private CallerSupport callerSupport;
-
   /** Construct with the given configuration. */
   public AnnotateSvsVcf(AnnotateSvsArgs args) {
     this.args = args;
     this.pedigree = null;
-    this.callerSupport = CallerSupportFactory.getFor(new File(args.getInputVcf()));
   }
 
   /** UUID counter for sequential UUID generation. */
@@ -139,19 +133,36 @@ public void run() {
                 featureEffectsStream, args.getOutputFeatureEffects());
         Writer dbInfoWriter =
             GzipUtil.maybeOpenGzipOutputStream(dbInfoStream, args.getOutputDbInfos());
-        BufferedWriter dbInfoBufWriter = new BufferedWriter(dbInfoWriter); ) {
+        BufferedWriter dbInfoBufWriter = new BufferedWriter(dbInfoWriter);
+        Closer covVcfCloser = Closer.create(); ) {
       // Guess genome version.
       GenomeVersion genomeVersion = new VcfCompatibilityChecker(reader).guessGenomeVersion();
 
       new VcfCompatibilityChecker(reader).check(args.getRelease());
       new DatabaseSelfTest(conn)
           .selfTest(args.getRelease(), args.isSelfTestChr1Only(), args.isSelfTestChr22Only());
 
+      final Map<String, CoverageFromMaelstromReader> covReaders = new TreeMap<>();
+      for (String covVcf : args.getCoverageVcfs()) {
+        final CoverageFromMaelstromReader covReader =
+            new CoverageFromMaelstromReader(new File(covVcf));
+        covReaders.put(covReader.getSample(), covReader);
+        covVcfCloser.register(covReader);
+      }
+      final CallerSupport callerSupport =
+          new CallerSupportFactory(covReaders).getFor(new File(args.getInputVcf()));
+
       System.err.println("Deserializing Jannovar file...");
       JannovarData refseqJvData = new JannovarDataSerializer(args.getRefseqSerPath()).load();
       JannovarData ensemblJvData = new JannovarDataSerializer(args.getEnsemblSerPath()).load();
       annotateSvVcf(
-          genomeVersion, reader, refseqJvData, ensemblJvData, gtWriter, featureEffectsWriter);
+          genomeVersion,
+          reader,
+          refseqJvData,
+          ensemblJvData,
+          callerSupport,
+          gtWriter,
+          featureEffectsWriter);
       new DbInfoWriterHelper()
           .writeDbInfos(conn, dbInfoBufWriter, args.getRelease(), AnnotateVcf.class);
     } catch (SQLException e) {
@@ -217,6 +228,7 @@ private void checkOptOutFeatures() {
    * @param reader Reader for the input VCF file.
    * @param refseqJv Deserialized RefSeq transcript database for Jannovar.
    * @param ensemblJv Deserialized ENSEMBL transcript database for Jannovar.
+   * @param callerSupport Helper to use for adapting to SV caller.
    * @param gtWriter Writer for variant call ("genotype") TSV file.
    * @param featureEffectsWriter Writer for gene-wise feature effects.
    * @throws VarfishAnnotatorException in case of problems
@@ -226,6 +238,7 @@ private void annotateSvVcf(
       VCFFileReader reader,
       JannovarData refseqJv,
       JannovarData ensemblJv,
+      CallerSupport callerSupport,
       Writer gtWriter,
       Writer featureEffectsWriter)
       throws VarfishAnnotatorException {