From bf5cbad04c61948ac6e2504aebc40e683ba4709b Mon Sep 17 00:00:00 2001
From: ebanks <ebanks@348d0f76-0448-11de-a6fe-93d51630548a>
Date: Mon, 28 Jun 2010 13:28:37 +0000
Subject: [PATCH] Make the target creator a rod walker (that allows reads) so
 that we can easily trigger the cleaner on only known indel sites.  Adding an
 integration test to cover this case.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@3651 348d0f76-0448-11de-a6fe-93d51630548a
---
 .../walkers/indels/RealignerTargetCreator.java | 18 ++++++++++++------
 .../RealignerTargetCreatorIntegrationTest.java |  6 ++++++
 2 files changed, 18 insertions(+), 6 deletions(-)

diff --git a/java/src/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreator.java b/java/src/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreator.java
index 24cc8790b..b127ff7b0 100755
--- a/java/src/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreator.java
+++ b/java/src/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreator.java
@@ -32,8 +32,7 @@ import org.broadinstitute.sting.gatk.contexts.variantcontext.VariantContext;
 import org.broadinstitute.sting.gatk.filters.Platform454Filter;
 import org.broadinstitute.sting.gatk.filters.ZeroMappingQualityReadFilter;
 import org.broadinstitute.sting.gatk.refdata.RefMetaDataTracker;
-import org.broadinstitute.sting.gatk.walkers.LocusWalker;
-import org.broadinstitute.sting.gatk.walkers.ReadFilters;
+import org.broadinstitute.sting.gatk.walkers.*;
 import org.broadinstitute.sting.utils.GenomeLoc;
 import org.broadinstitute.sting.utils.GenomeLocParser;
 import org.broadinstitute.sting.utils.StingException;
@@ -49,7 +48,10 @@ import java.util.ArrayList;
  * Emits intervals for the Local Indel Realigner to target for cleaning.  Ignores 454 and MQ0 reads.
  */
 @ReadFilters({Platform454Filter.class, ZeroMappingQualityReadFilter.class})
-public class RealignerTargetCreator extends LocusWalker<RealignerTargetCreator.Event, RealignerTargetCreator.Event> {
+@Reference(window=@Window(start=-1,stop=50))
+@Allows(value={DataSource.READS, DataSource.REFERENCE})
+@By(DataSource.REFERENCE)
+public class RealignerTargetCreator extends RodWalker<RealignerTargetCreator.Event, RealignerTargetCreator.Event> {
 
     // mismatch/entropy/SNP arguments
     @Argument(fullName="windowSize", shortName="window", doc="window size for calculating entropy or SNP clusters", required=false)
@@ -86,7 +88,7 @@ public class RealignerTargetCreator extends LocusWalker<RealignerTargetCreator.E
         long furthestStopPos = -1;
 
         // look for insertions in the extended context (we'll get deletions from the normal context)
-        if ( context.hasExtendedEventPileup() ) {
+        if ( context != null && context.hasExtendedEventPileup() ) {
             ReadBackedExtendedEventPileup pileup = context.getExtendedEventPileup();
             if ( pileup.getNumberOfInsertions() > 0 ) {
                 hasIndel = hasInsertion = true;
@@ -115,7 +117,9 @@ public class RealignerTargetCreator extends LocusWalker<RealignerTargetCreator.E
                             hasInsertion = true;
                         break;
                 }
-             }
+                if ( hasIndel )
+                    furthestStopPos = vc.getLocation().getStop();
+            }
         }
 
         // look at the normal context to get deletions and positions with high entropy
@@ -154,13 +158,15 @@ public class RealignerTargetCreator extends LocusWalker<RealignerTargetCreator.E
         if ( !hasIndel && !hasPointEvent )
             return null;
 
-        // return null if we didn't find any usable reads associated with the event
+        // return null if we didn't find any usable reads/rods associated with the event
         if ( furthestStopPos == -1 )
             return null;
 
         GenomeLoc eventLoc = context.getLocation();
         if ( hasInsertion )
             eventLoc =  GenomeLocParser.createGenomeLoc(eventLoc.getContigIndex(), eventLoc.getStart(), eventLoc.getStart()+1);
+        else if ( hasIndel && context.getBasePileup().size() == 0 )
+            eventLoc =  GenomeLocParser.createGenomeLoc(eventLoc.getContigIndex(), eventLoc.getStart(), furthestStopPos);        
 
         EVENT_TYPE eventType = (hasIndel ? (hasPointEvent ? EVENT_TYPE.BOTH : EVENT_TYPE.INDEL_EVENT) : EVENT_TYPE.POINT_EVENT);
 
diff --git a/java/test/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreatorIntegrationTest.java b/java/test/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreatorIntegrationTest.java
index b288fb260..f1b338789 100755
--- a/java/test/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreatorIntegrationTest.java
+++ b/java/test/org/broadinstitute/sting/gatk/walkers/indels/RealignerTargetCreatorIntegrationTest.java
@@ -20,5 +20,11 @@ public class RealignerTargetCreatorIntegrationTest extends WalkerTest {
                  1,
                  Arrays.asList("bfccfa50f62d10ee2fe8cfa68fb70002"));
         executeTest("test dbsnp", spec2);
+
+        WalkerTest.WalkerTestSpec spec3 = new WalkerTest.WalkerTestSpec(
+                "-T RealignerTargetCreator -R " + oneKGLocation + "reference/human_b36_both.fasta -B indels,VCF," + validationDataLocation + "NA12878.chr1_10mb_11mb.slx.indels.vcf -BTI indels -o %s",
+                 1,
+                 Arrays.asList("1a11cfc9cc713617c82bdec503ebe02a"));
+        executeTest("test rods only", spec3);
     }
 }