We no longer require dbSNP files to be of the dbsnp rod-type; VCFs will do (provided they are bound to the name 'dbsnp')

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@4753 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/refdata/utils/helpers/DbSNPHelper.java

@@ -3,6 +3,7 @@ package org.broadinstitute.sting.gatk.refdata.utils.helpers;
 import net.sf.samtools.util.SequenceUtil;
 import org.broad.tribble.annotation.Strand;
 import org.broad.tribble.dbsnp.DbSNPFeature;
+import org.broad.tribble.util.variantcontext.VariantContext;
 import org.broadinstitute.sting.utils.Utils;
 
 import java.util.ArrayList;
@@ -33,6 +34,28 @@ public class DbSNPHelper {
         return dbsnp;
     }
 
+    public static String rsIDOfFirstRealSNP(List<Object> featureList) {
+        if (featureList == null)
+            return null;
+
+        String rsID = null;
+        for ( Object d : featureList ) {
+            if ( d instanceof DbSNPFeature ) {
+                if ( DbSNPHelper.isSNP((DbSNPFeature)d) ) {
+                    rsID = ((DbSNPFeature)d).getRsID();
+                    break;
+                }
+            } else if ( d instanceof VariantContext) {
+                if ( ((VariantContext)d).isSNP() ) {
+                    rsID = ((VariantContext)d).getID();
+                    break;
+                }
+            }
+        }
+
+        return rsID;
+    }
+
     /**
      * get the -1 * (log 10 of the error value)
      *

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

@@ -25,7 +25,6 @@
 
 package org.broadinstitute.sting.gatk.walkers;
 
-import org.broad.tribble.dbsnp.DbSNPFeature;
 import org.broad.tribble.hapmap.HapMapFeature;
 import org.broad.tribble.util.variantcontext.Genotype;
 import org.broad.tribble.util.variantcontext.VariantContext;
@@ -69,14 +68,14 @@ public class VariantsToVCF extends RodWalker<Integer, Integer> {
         if ( tracker == null || !BaseUtils.isRegularBase(ref.getBase()) )
             return 0;
 
-        DbSNPFeature dbsnp = DbSNPHelper.getFirstRealSNP(tracker.getReferenceMetaData(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME));
+        String rsID = DbSNPHelper.rsIDOfFirstRealSNP(tracker.getReferenceMetaData(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME));
 
         Collection<VariantContext> contexts = tracker.getVariantContexts(ref, INPUT_ROD_NAME, ALLOWED_VARIANT_CONTEXT_TYPES, context.getLocation(), true, false);
 
         for ( VariantContext vc : contexts ) {
             Map<String, Object> attrs = new HashMap<String, Object>(vc.getAttributes());
-            if ( dbsnp != null )
+            if ( rsID != null )
-                attrs.put(VariantContext.ID_KEY, dbsnp.getRsID());
+                attrs.put(VariantContext.ID_KEY, rsID);
             vc = VariantContext.modifyAttributes(vc, attrs);
 
             // set the appropriate sample name if necessary

java/src/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotatorEngine.java

@@ -37,7 +37,6 @@ import java.util.Map;
 import java.util.Set;
 import java.util.Map.Entry;
 
-import org.broad.tribble.dbsnp.DbSNPFeature;
 import org.broad.tribble.util.variantcontext.Genotype;
 import org.broad.tribble.util.variantcontext.VariantContext;
 import org.broad.tribble.vcf.*;
@@ -212,11 +211,11 @@ public class VariantAnnotatorEngine {
     private void annotateDBs(RefMetaDataTracker tracker, ReferenceContext ref, VariantContext vc, Map<String, Object> infoAnnotations) {
         for ( Map.Entry<String, String> dbSet : dbAnnotations.entrySet() ) {
             if ( dbSet.getKey().equals(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME) ) {
-                DbSNPFeature dbsnp = DbSNPHelper.getFirstRealSNP(tracker.getReferenceMetaData(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME));
+                String rsID = DbSNPHelper.rsIDOfFirstRealSNP(tracker.getReferenceMetaData(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME));
-                infoAnnotations.put(VCFConstants.DBSNP_KEY, dbsnp == null ? false : true);
+                infoAnnotations.put(VCFConstants.DBSNP_KEY, rsID == null ? false : true);
                 // annotate dbsnp id if available and not already there
-                if ( dbsnp != null && (!vc.hasAttribute(VariantContext.ID_KEY) || vc.getAttribute(VariantContext.ID_KEY).equals(VCFConstants.EMPTY_ID_FIELD)) )
+                if ( rsID != null && (!vc.hasID() || vc.getID().equals(VCFConstants.EMPTY_ID_FIELD)) )
-                    infoAnnotations.put(VariantContext.ID_KEY, dbsnp.getRsID());
+                    infoAnnotations.put(VariantContext.ID_KEY, rsID);
             } else {
                 boolean overlapsComp = false;
                 for ( VariantContext comp : tracker.getVariantContexts(ref, dbSet.getKey(), null, ref.getLocus(), false, false) ) {

java/src/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyperEngine.java

@@ -45,7 +45,6 @@ import org.broadinstitute.sting.utils.pileup.*;
 import org.broadinstitute.sting.utils.sam.GATKSAMRecordFilter;
 import org.broadinstitute.sting.utils.sam.GATKSAMRecord;
 import org.broad.tribble.vcf.VCFConstants;
-import org.broad.tribble.dbsnp.DbSNPFeature;
 
 import java.io.PrintStream;
 import java.util.*;
@@ -312,9 +311,9 @@ public class UnifiedGenotyperEngine {
         // *** note that calculating strand bias involves overwriting data structures, so we do that last
         HashMap<String, Object> attributes = new HashMap<String, Object>();
 
-        DbSNPFeature dbsnp = getDbSNP(tracker);
+        String rsID = DbSNPHelper.rsIDOfFirstRealSNP(tracker.getReferenceMetaData(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME));
-        if ( dbsnp != null )
+        if ( rsID != null )
-            attributes.put(VariantContext.ID_KEY, dbsnp.getRsID());
+            attributes.put(VariantContext.ID_KEY, rsID);
 
         // if the site was downsampled, record that fact
         if ( rawContext.hasPileupBeenDownsampled() )
@@ -598,15 +597,6 @@ public class UnifiedGenotyperEngine {
         }
     }
 
-    /**
-     * @param tracker   rod data
-     *
-     * @return the dbsnp rod if there is one at this position
-     */
-    protected static DbSNPFeature getDbSNP(RefMetaDataTracker tracker) {
-        return DbSNPHelper.getFirstRealSNP(tracker.getReferenceMetaData(DbSNPHelper.STANDARD_DBSNP_TRACK_NAME));
-    }
-
     protected boolean passesEmitThreshold(double conf, int bestAFguess, boolean atTriggerTrack) {
         return (atTriggerTrack ?
                 (conf >= Math.min(UAC.TRIGGER_CONFIDENCE_FOR_CALLING, UAC.TRIGGER_CONFIDENCE_FOR_EMITTING)) :