Merge pull request #1460 from broadinstitute/rhl_rank_sum_test_del_return

Make getElementForRead() in RankSumTest robust

protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/tools/walkers/annotator/AS_RankSumTest.java

@@ -278,7 +278,7 @@ public abstract class AS_RankSumTest extends RankSumTest implements ReducibleAnn
             if ( isUsableRead(read, refLoc) ) {
                 final Double value = getElementForRead(read, refLoc, a);
                 // Bypass read if the clipping goal is not reached or the refloc is inside a spanning deletion
-                if ( value == null || value < 0.0 )
+                if ( value == null || value == INVALID_ELEMENT_FROM_READ )
                     continue;
 
                 if(perAlleleValues.containsKey(a.getMostLikelyAllele()))

protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/tools/walkers/annotator/AS_ReadPosRankSumTest.java

@@ -102,10 +102,13 @@ public class AS_ReadPosRankSumTest extends AS_RankSumTest implements AS_Standard
         if ( offset == ReadUtils.CLIPPING_GOAL_NOT_REACHED )
             return null;
 
+        // If the offset inside a deletion, it does not lie on a read.
+        if ( AlignmentUtils.isInsideDeletion(read.getCigar(), offset) ) {
+            return INVALID_ELEMENT_FROM_READ;
+        }
+
         int readPos = AlignmentUtils.calcAlignmentByteArrayOffset(read.getCigar(), offset, false, 0, 0);
         final int numAlignedBases = AlignmentUtils.getNumAlignedBasesCountingSoftClips( read );
-        // Note: For a spanning deletion, readPos is at the upstream end of the deletion and is greater than numAlignedBases (which does not include deletions).
-        // Hence, the resulting readPos will have a negative value.
         if (readPos > numAlignedBases / 2)
             readPos = numAlignedBases - (readPos + 1);
         return (double)readPos;

protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/tools/walkers/annotator/RankSumTest.java

@@ -76,7 +76,7 @@ import java.util.*;
 //TODO: will eventually implement ReducibleAnnotation in order to preserve accuracy for CombineGVCFs and GenotypeGVCFs -- see RMSAnnotation.java for an example of an abstract ReducibleAnnotation
 public abstract class RankSumTest extends InfoFieldAnnotation implements ActiveRegionBasedAnnotation {
     static final boolean DEBUG = false;
-    protected static double INVALID_READ_POSITION = -1; // No mapping to a read position
+    protected static double INVALID_ELEMENT_FROM_READ = Double.NEGATIVE_INFINITY;
 
     public Map<String, Object> annotate(final RefMetaDataTracker tracker,
                                         final AnnotatorCompatible walker,
@@ -185,7 +185,7 @@ public abstract class RankSumTest extends InfoFieldAnnotation implements ActiveR
             if ( isUsableRead(read, refLoc) ) {
                 final Double value = getElementForRead(read, refLoc, a);
                 // Bypass read if the clipping goal is not reached or the refloc is inside a spanning deletion
-                if ( value == null || value == INVALID_READ_POSITION )
+                if ( value == null || value == INVALID_ELEMENT_FROM_READ )
                     continue;
 
                 if ( a.getMostLikelyAllele().isReference() )

protected/gatk-tools-protected/src/main/java/org/broadinstitute/gatk/tools/walkers/annotator/ReadPosRankSumTest.java

@@ -106,7 +106,7 @@ public class ReadPosRankSumTest extends RankSumTest implements StandardAnnotatio
 
         // If the offset inside a deletion, it does not lie on a read.
         if ( AlignmentUtils.isInsideDeletion(read.getCigar(), offset) ) {
-            return INVALID_READ_POSITION;
+            return INVALID_ELEMENT_FROM_READ;
         }
 
         int readPos = AlignmentUtils.calcAlignmentByteArrayOffset( read.getCigar(), offset, false, 0, 0 );

protected/gatk-tools-protected/src/test/java/org/broadinstitute/gatk/tools/walkers/haplotypecaller/HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest.java

@@ -72,7 +72,7 @@ public class HaplotypeCallerComplexAndSymbolicVariantsIntegrationTest extends Wa
 
     @Test
     public void testHaplotypeCallerMultiSampleComplex1() {
-        HCTestComplexVariants(privateTestDir + "AFR.complex.variants.bam", "", "86528820f8c102c712d9562b83204c05");
+        HCTestComplexVariants(privateTestDir + "AFR.complex.variants.bam", "", "b01df95864808dc67295efc6db37983d");
     }
 
     private void HCTestSymbolicVariants(String bam, String args, String md5) {

protected/gatk-tools-protected/src/test/java/org/broadinstitute/gatk/tools/walkers/haplotypecaller/HaplotypeCallerGVCFIntegrationTest.java

@@ -458,7 +458,7 @@ public class HaplotypeCallerGVCFIntegrationTest extends WalkerTest {
     public void testHaplotypeCallerGVCSpanDel() {
         final String commandLine = String.format("-T HaplotypeCaller -R %s -I %s -L 1:26357667 -ERC GVCF --no_cmdline_in_header -A AS_ReadPosRankSumTest -A ReadPosRankSumTest -variant_index_type %s -variant_index_parameter %d",
                 b37KGReference, privateTestDir + "NexPond-377866-1:26357600-26357700.bam", GATKVCFUtils.DEFAULT_GVCF_INDEX_TYPE, GATKVCFUtils.DEFAULT_GVCF_INDEX_PARAMETER);
-        final WalkerTestSpec spec = new WalkerTestSpec(commandLine + " -o %s", Arrays.asList("93bc22340e6a4b01a7b96e5a3a12dfc3"));
+        final WalkerTestSpec spec = new WalkerTestSpec(commandLine + " -o %s", Arrays.asList("b8f0bb74bc099a8f78d600d88861e1b6"));
         spec.disableShadowBCF();
         executeTest("testHaplotypeCallerGVCSpanDel", spec);
     }