Increasing unit test coverage for VariantContextConverter

bigdatagenomics · Nov 16, 2016 · 6835e2e · 6835e2e
1 parent c1b4b7d
commit 6835e2e
Show file tree

Hide file tree

Showing 4 changed files with 159 additions and 8 deletions.
diff --git a/adam-core/src/main/scala/org/bdgenomics/adam/converters/VariantContextConverter.scala b/adam-core/src/main/scala/org/bdgenomics/adam/converters/VariantContextConverter.scala
@@ -397,6 +397,9 @@ private[adam] class VariantContextConverter(dict: Option[SequenceDictionary] = N
     if (vc.isFiltered) {
       builder.setFiltersFailed(new java.util.ArrayList(vc.getFilters));
     }
+    if (vc.getAttributeAsBoolean("SOMATIC", false)) {
+      builder.setSomatic(true)
+    }
     builder.build
   }
 
@@ -607,6 +610,21 @@ private[adam] class VariantContextConverter(dict: Option[SequenceDictionary] = N
       case Some(s) => vcb.id(s)
     }
 
+    val filtersApplied = Option(variant.getFiltersApplied).getOrElse(false)
+    val filtersPassed = Option(variant.getFiltersPassed).getOrElse(false)
+
+    (filtersApplied, filtersPassed) match {
+      case (false, false) => vcb.unfiltered
+      case (false, true)  => vcb.passFilters // log warning?
+      case (true, false)  => vcb.filters(new java.util.HashSet(variant.getFiltersFailed()))
+      case (true, true)   => vcb.passFilters
+    }
+
+    val somatic: java.lang.Boolean = Option(variant.getSomatic).getOrElse(false)
+    if (somatic) {
+      vcb.attribute("SOMATIC", true)
+    }
+
     // TODO: Extract provenance INFO fields
     try {
       vcb.genotypes(vc.genotypes.map(g => {

diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/converters/VariantContextConverterSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/converters/VariantContextConverterSuite.scala
@@ -86,6 +86,27 @@ class VariantContextConverterSuite extends ADAMFunSuite {
 
     assert(variant.getReferenceAllele === "A")
     assert(variant.getStart === 0L)
+    assert(variant.getSomatic === false)
+  }
+
+  test("Convert somatic htsjdk site-only SNV to ADAM") {
+    val converter = new VariantContextConverter
+
+    // not sure why this doesn't work
+    //htsjdkSNVBuilder.attribute("SOMATIC", true)
+    //val adamVCs = converter.convert(htsjdkSNVBuilder.make)
+
+    val vcb: VariantContextBuilder = new VariantContextBuilder()
+      .alleles(List(Allele.create("A", true), Allele.create("T")))
+      .start(1L)
+      .stop(1L)
+      .chr("1")
+      .attribute("SOMATIC", true)
+
+    val adamVCs = converter.convert(vcb.make)
+    val adamVC = adamVCs.head
+    val variant = adamVC.variant.variant
+    assert(variant.getSomatic === true)
   }
 
   test("Convert htsjdk site-only SNV to ADAM with contig conversion") {
@@ -413,4 +434,94 @@ class VariantContextConverterSuite extends ADAMFunSuite {
     assert(htsjdkVC.hasID)
     assert(htsjdkVC.getID === "rs3131972;rs201888535")
   }
+
+  test("Convert ADAM variant context with null filters applied to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setFiltersApplied(null)
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(!htsjdkVC.filtersWereApplied)
+    assert(!htsjdkVC.isFiltered)
+    assert(htsjdkVC.getFilters.isEmpty)
+  }
+
+  test("Convert ADAM variant context with no filters applied to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setFiltersApplied(false)
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(!htsjdkVC.filtersWereApplied)
+    assert(!htsjdkVC.isFiltered)
+    assert(htsjdkVC.getFilters.isEmpty)
+  }
+
+  test("Convert ADAM variant context with passing filters to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setFiltersApplied(true)
+      .setFiltersPassed(true)
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(htsjdkVC.filtersWereApplied)
+    assert(!htsjdkVC.isFiltered)
+    assert(htsjdkVC.getFilters.isEmpty)
+  }
+
+  test("Convert ADAM variant context with failing filters to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setFiltersApplied(true)
+      .setFiltersPassed(false)
+      .setFiltersFailed(ImmutableList.of("FILTER1", "FILTER2"))
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(htsjdkVC.filtersWereApplied)
+    assert(htsjdkVC.isFiltered)
+    assert(htsjdkVC.getFilters.contains("FILTER1"))
+    assert(htsjdkVC.getFilters.contains("FILTER2"))
+  }
+
+  test("Convert ADAM variant context with null somatic flag to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setSomatic(null)
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(!htsjdkVC.hasAttribute("SOMATIC"))
+  }
+
+  test("Convert ADAM variant context with non-somatic variant to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setSomatic(false)
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(!htsjdkVC.hasAttribute("SOMATIC"))
+  }
+
+  test("Convert ADAM variant context with somatic variant to htsjdk") {
+    val variant = adamSNVBuilder()
+      .setSomatic(true)
+      .build
+
+    val converter = new VariantContextConverter
+
+    val htsjdkVC = converter.convert(ADAMVariantContext(variant))
+    assert(htsjdkVC.hasAttribute("SOMATIC"))
+    assert(htsjdkVC.getAttributeAsBoolean("SOMATIC", false) === true)
+  }
 }
diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/rdd/ADAMContextSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/rdd/ADAMContextSuite.scala
@@ -152,6 +152,22 @@ class ADAMContextSuite extends ADAMFunSuite {
     assert(vcs.size === 6)
 
     val vc = vcs.head
+
+    /*
+1	14397	.	CTGT	C	139.12	IndelQD	AC=2;AF=0.333;AN=6;BaseQRankSum=1.800;ClippingRankSum=0.138;DP=69;FS=7.786;MLEAC=2;MLEAF=0.333;MQ=26.84;MQ0=0;MQRankSum=-1.906;QD=1.55;ReadPosRankSum=0.384	GT:AD:DP:FT:GQ:PL	0/1:16,4:20:rd:99:120,0,827	0/1:8,2:10:dp;rd:60:60,0,414	0/0:39,0:39:PASS:99:0,116,2114
+     */
+    val variant = vc.variant.variant
+    assert(variant.getContigName === "1")
+    assert(variant.getStart === 14396L)
+    assert(variant.getEnd === 14400L)
+    assert(variant.getReferenceAllele === "CTGT")
+    assert(variant.getAlternateAllele === "C")
+    assert(variant.getNames.isEmpty)
+    assert(variant.getFiltersApplied === true)
+    assert(variant.getFiltersPassed === false)
+    assert(variant.getFiltersFailed.contains("IndelQD"))
+    assert(variant.getSomatic === false)
+
     assert(vc.genotypes.size === 3)
 
     val gt = vc.genotypes.head

diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/rdd/variant/VariantContextRDDSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/rdd/variant/VariantContextRDDSuite.scala
@@ -45,6 +45,8 @@ class VariantContextRDDSuite extends ADAMFunSuite {
       .setReferenceAllele("T")
       .setAlternateAllele("C")
       .setNames(ImmutableList.of("rs3131972", "rs201888535"))
+      .setFiltersApplied(true)
+      .setFiltersPassed(true)
       .build
 
     val g0 = Genotype.newBuilder().setVariant(v0)
@@ -67,14 +69,18 @@ class VariantContextRDDSuite extends ADAMFunSuite {
     val vcRdd = sc.loadVcf("%s/test.vcf/part-r-00000".format(tempDir))
     assert(vcRdd.rdd.count === 1)
 
-    val variant = vcRdd.rdd.first.variant
-    assert(variant.variant.getContigName === "chr11")
-    assert(variant.variant.getStart === 17409572)
-    assert(variant.variant.getReferenceAllele === "T")
-    assert(variant.variant.getAlternateAllele === "C")
-    assert(variant.variant.getNames.length === 2)
-    assert(variant.variant.getNames.get(0) === "rs3131972")
-    assert(variant.variant.getNames.get(1) === "rs201888535")
+    val variant = vcRdd.rdd.first.variant.variant
+    assert(variant.getContigName === "chr11")
+    assert(variant.getStart === 17409572)
+    assert(variant.getReferenceAllele === "T")
+    assert(variant.getAlternateAllele === "C")
+    assert(variant.getNames.length === 2)
+    assert(variant.getNames.get(0) === "rs3131972")
+    assert(variant.getNames.get(1) === "rs201888535")
+    assert(variant.getFiltersApplied === true)
+    assert(variant.getFiltersPassed === true)
+    assert(variant.getFiltersFailed.isEmpty)
+    assert(variant.getSomatic === false)
 
     assert(vcRdd.sequences.records.size === 1)
     assert(vcRdd.sequences.records(0).name === "chr11")