Some minor changes and test:

- DepthOfCoverage is now by reference (so locus-by-locus output correctly reports zero-coverage bases)
  - VariantsToVCF now lets you bind variants with any string except intervals and dbsnp (not just NA######)
  - A PileupWalker integration test on a particularly nasty FHS site
  - Two second-base annotation related integration tests on that same site
       + outputs were all hand-validated in matlab; within a certain tolerance for the annotations




git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2197 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/walkers/DepthOfCoverageWalker.java

@@ -23,7 +23,8 @@
  */
 
 package org.broadinstitute.sting.gatk.walkers;
-
+import org.broadinstitute.sting.gatk.walkers.DataSource;
+import org.broadinstitute.sting.gatk.walkers.By;
 import org.broadinstitute.sting.gatk.contexts.AlignmentContext;
 import org.broadinstitute.sting.gatk.contexts.ReferenceContext;
 import org.broadinstitute.sting.gatk.refdata.RefMetaDataTracker;
@@ -39,6 +40,7 @@ import java.util.ArrayList;
 /**
  * Display the depth of coverage at a given locus.
  */
+@By(DataSource.REFERENCE)
 public class DepthOfCoverageWalker extends LocusWalker<Integer, Pair<Long, Long>> {
     enum printType {
         NONE,
@@ -131,4 +133,4 @@ public class DepthOfCoverageWalker extends LocusWalker<Integer, Pair<Long, Long>
         out.printf("Average depth of coverage is: %.2f in %d total coverage over %d sites\n", 
                 ((double)result.getFirst() / (double)result.getSecond()), result.getFirst(), result.getSecond());
     }
-}
+}

java/src/org/broadinstitute/sting/playground/gatk/walkers/variantstovcf/VariantsToVCF.java

@@ -39,8 +39,10 @@ public class VariantsToVCF extends RefWalker<Integer, Integer> {
             String[] rodPieces = rodName.split(",");
             String sampleName = rodPieces[0];
 
-            if (sampleName.startsWith("NA"))
-                sampleNames.put(sampleName.toUpperCase(), sampleName.toUpperCase());
+            if (! ( sampleName.equals("dbSNP") || sampleName.equals("interval") ) ) {
+                sampleNames.put(sampleName, sampleName);
+            }
+
         }
 
         vcfheader = getHeader(args, sampleNames.keySet());

java/test/org/broadinstitute/sting/gatk/walkers/PileupWalkerIntegrationTest.java

@@ -0,0 +1,26 @@
+package org.broadinstitute.sting.gatk.walkers;
+
+import org.broadinstitute.sting.WalkerTest;
+import org.junit.Test;
+
+import java.util.Arrays;
+
+/**
+ * Created by IntelliJ IDEA.
+ * User: chartl
+ * Date: Dec 1, 2009
+ * Time: 9:03:34 AM
+ * To change this template use File | Settings | File Templates.
+ */
+public class PileupWalkerIntegrationTest extends WalkerTest {
+
+    @Test
+    public void testGnarleyFHSPileup() {
+        String gatk_args = "-T Pileup -I /humgen/gsa-scr1/GATK_Data/Validation_Data/FHS_Pileup_Test.bam "
+                 + "-R /seq/references/Homo_sapiens_assembly18/v0/Homo_sapiens_assembly18.fasta"
+                 +  " -L chr15:46,347,148 -o %s";
+        String expected_md5 = "98040c47b623bb133cd296ce12768d49";
+        WalkerTestSpec spec = new WalkerTestSpec(gatk_args, 1, Arrays.asList(expected_md5));
+        executeTest("Testing the standard (no-indel) pileup on three merged FHS pools with 27 deletions in 969 bases", spec);
+    }
+}

java/test/org/broadinstitute/sting/gatk/walkers/annotator/SecondBaseSkewIntegrationTest.java

@@ -55,4 +55,26 @@ public class SecondBaseSkewIntegrationTest extends WalkerTest {
         executeTest("Testing on bam file without 2bb annotations ",spec);
     }
 
+    @Test
+    public void testOnIndels() {
+        String test_args = "-T VariantAnnotator -I /humgen/gsa-scr1/GATK_Data/Validation_Data/FHS_Pileup_Test.bam"
+                     + " -R /seq/references/Homo_sapiens_assembly18/v0/Homo_sapiens_assembly18.fasta -A SecondBaseSkew"
+                     + " -sample variant -B variant,VCF,/humgen/gsa-scr1/GATK_Data/Validation_Data/FHS_pileup_test_chr15.vcf"
+                     + " -vcf %s -L chr15:46347148";
+        String expected_md5 = "fe7f6d2b48ecf2c1340b5de98a9d5614";
+        WalkerTestSpec spec = new WalkerTestSpec(test_args,1,Arrays.asList(expected_md5));
+        executeTest("Testing on locus with many indels", spec);
+    }
+
+    @Test
+    public void testPrimaryBaseSecondaryBaseOnIndels() {
+        String test_args = "-T VariantAnnotator -I /humgen/gsa-scr1/GATK_Data/Validation_Data/FHS_Pileup_Test.bam"
+                     + " -R /seq/references/Homo_sapiens_assembly18/v0/Homo_sapiens_assembly18.fasta -A PrimaryBaseSecondaryBaseSymmetry"
+                     + " -sample variant -B variant,VCF,/humgen/gsa-scr1/GATK_Data/Validation_Data/FHS_pileup_test_chr15.vcf"
+                     + " -vcf %s -L chr15:46347148";
+        String expected_md5 = "9b587be7a270c6df7e0affcfc61a861a";
+        WalkerTestSpec spec = new WalkerTestSpec(test_args,1,Arrays.asList(expected_md5));
+        executeTest("Testing PrimaryBaseSecondaryBaseSymmetry on locus with many indels", spec);
+    }
+
 }