-B 1kg_ceu,VFC,CEU.vcf -B 1kg_yri,VCF,YRI.vcf system supported to allow 1KG % (like dbSNP%)

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2632 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/refdata/RODRecordList.java

@@ -64,13 +64,24 @@ public class RODRecordList<ROD extends ReferenceOrderedDatum> implements Iterabl
     public Iterator<ROD> iterator() { return records.iterator() ; }
     public void clear() { records.clear(); }
     public boolean isEmpty() { return records.isEmpty(); }
-    public void add(ROD record) {
+
+    public void add(ROD record) { add(record, false); }
+
+    public void add(ROD record, boolean allowNameMismatch) {
         if ( record != null ) {
-            if ( ! name.equals(record.getName() ) )
+            if ( ! allowNameMismatch && ! name.equals(record.getName() ) )
                 throw new StingException("Attempt to add ROD with non-matching name "+record.getName()+" to the track "+name);
         }
         records.add(record);
     }
+
+    public void add(RODRecordList<ROD> records ) { add( records, false ); }
+
+    public void add(RODRecordList<ROD> records, boolean allowNameMismatch) {
+        for ( ROD record : records )
+            add(record, allowNameMismatch);
+    }    
+
     public int size() { return records.size() ; }
 
     /**
@@ -104,7 +115,7 @@ public class RODRecordList<ROD extends ReferenceOrderedDatum> implements Iterabl
      * <tt>0</tt>, or <tt>1</tt> according to whether the value of
      * <i>expression</i> is negative, zero or positive.
      *
-     * @param o the object to be compared.
+     * @param that the object to be compared.
      * @return a negative integer, zero, or a positive integer as this object
      *         is less than, equal to, or greater than the specified object.
      * @throws ClassCastException if the specified object's type prevents it

java/src/org/broadinstitute/sting/gatk/refdata/RefMetaDataTracker.java

@@ -60,20 +60,41 @@ public class RefMetaDataTracker {
      * @param defaultValue
      * @return
      */
-    public RODRecordList<ReferenceOrderedDatum> getTrackData(final String name, ReferenceOrderedDatum defaultValue) {
+    public RODRecordList<ReferenceOrderedDatum> getTrackData(final String name, ReferenceOrderedDatum defaultValue, boolean requireExactMatch) {
         //logger.debug(String.format("Lookup %s%n", name));
-        final String luName = canonicalName(name);
-        if ( map.containsKey(luName) )
-            return map.get(luName);
-        else {
 
-            if ( defaultValue == null )
+        final String luName = canonicalName(name);
-                return null;
+        RODRecordList<ReferenceOrderedDatum> trackData = null;
-            return new RODRecordList<ReferenceOrderedDatum>(defaultValue.getName(),
+
-                                                            Collections.singletonList(defaultValue),
+        if ( requireExactMatch ) {
-                                                            defaultValue.getLocation());
+            if ( map.containsKey(luName) )
+                trackData = map.get(luName);
+        } else {
+            for ( Map.Entry<String, RODRecordList<ReferenceOrderedDatum>> datum : map.entrySet() ) {
+                final String rodName = datum.getKey();
+                if ( rodName.startsWith(luName) ) {
+                    if ( trackData == null ) trackData = new RODRecordList<ReferenceOrderedDatum>(name);
+                    //System.out.printf("Adding bindings from %s to %s at %s%n", rodName, name, datum.getValue().getLocation());
+                    trackData.add(datum.getValue(), true);
+                }
+            }
         }
+
+        if ( trackData != null )
+            return trackData;
+        else if ( defaultValue == null )
+            return null;
+        else
+            return new RODRecordList<ReferenceOrderedDatum>(defaultValue.getName(),
+                    Collections.singletonList(defaultValue),
+                    defaultValue.getLocation());
     }
+
+    public RODRecordList<ReferenceOrderedDatum> getTrackData(final String name, ReferenceOrderedDatum defaultValue) {
+        return getTrackData(name, defaultValue, true);
+    }
+
+
     /**
      * @see this.lookup
      * @param name

java/src/org/broadinstitute/sting/playground/gatk/walkers/varianteval/VariantDBCoverage.java

@@ -58,11 +58,26 @@ public class VariantDBCoverage extends BasicVariantAnalysis implements GenotypeA
         return nConcordant() / (1.0 * nSNPsAtdbSNPs());
     }
 
+    public static Variation getFirstRealSNP(RODRecordList<ReferenceOrderedDatum> dbsnpList) {
+        if (dbsnpList == null)
+            return null;
+
+        Variation dbsnp = null;
+        for (ReferenceOrderedDatum d : dbsnpList) {
+            if (((Variation) d).isSNP()) {
+                dbsnp = (Variation)d;
+                break;
+            }
+        }
+
+        return dbsnp;
+    }
+
     public String update(Variation eval, RefMetaDataTracker tracker, char ref, AlignmentContext context) {
-        rodDbSNP dbSNP = rodDbSNP.getFirstRealSNP(tracker.getTrackData( dbName, null ));
+        Variation dbSNP = getFirstRealSNP(tracker.getTrackData( dbName, null, false ));
         String result = null;
 
-        if (dbSNP != null) nDBSNPs++;               // count the number of real dbSNP events
+        if ( dbSNP != null ) nDBSNPs++;               // count the number of real dbSNP events
         if ( eval != null && eval.isSNP() ) {       // ignore indels right now
             nEvalObs++;                             // count the number of eval snps we've seen
 
@@ -76,13 +91,13 @@ public class VariantDBCoverage extends BasicVariantAnalysis implements GenotypeA
             }
         }
 
-        if ( dbSNP != null && dbSNP.isSNP() ) {
+//        if ( dbSNP != null && dbSNP.isSNP() ) {
-            BrokenRODSimulator.attach("dbSNP");
+//            BrokenRODSimulator.attach("dbSNP");
-            rodDbSNP dbsnp = (rodDbSNP) BrokenRODSimulator.simulate_lookup("dbSNP", context.getLocation(), tracker);
+//            rodDbSNP dbsnp = (rodDbSNP) BrokenRODSimulator.simulate_lookup("dbSNP", context.getLocation(), tracker);
-            if ( ! dbSNP.getRS_ID().equals(dbsnp.getRS_ID()) && dbsnp.isSNP() ) {
+//            if ( ! dbSNP.getRS_ID().equals(dbsnp.getRS_ID()) && dbsnp.isSNP() ) {
-                System.out.printf("Discordant site! %n%s%n vs.%n%s%n", dbSNP, dbsnp);
+//                System.out.printf("Discordant site! %n%s%n vs.%n%s%n", dbSNP, dbsnp);
-            }
+//            }
-        }
+//        }
 
         return result;
     }
@@ -97,7 +112,7 @@ public class VariantDBCoverage extends BasicVariantAnalysis implements GenotypeA
         s.add(String.format("n_concordant             %d", nConcordant()));
         s.add(String.format("n_novel_sites            %d", nNovelSites()));
 
-        s.add(String.format("dbsnp_rate               %.2f       # percent eval snps at dbsnp snps", 100 * dbSNPRate()));
+        s.add(String.format("%s_rate               %.2f       # percent eval snps at dbsnp snps", dbName.toLowerCase(), 100 * dbSNPRate()));
         s.add(String.format("concordance_rate         %.2f", 100 * concordanceRate()));
 
         return s;

java/src/org/broadinstitute/sting/playground/gatk/walkers/varianteval/VariantEvalWalker.java

@@ -71,6 +71,7 @@ public class VariantEvalWalker extends RodWalker<Integer, Integer> {
     String analysisFilenameBase = null;
 
     final String knownSNPDBName = "dbSNP";
+    final String One1KGSNPNames = "1kg";
     final String genotypeChipName = "hapmap-chip";
 
     HashMap<ANALYSIS_TYPE, ArrayList<VariantAnalysis>> analysisSets;
@@ -164,6 +165,7 @@ public class VariantEvalWalker extends RodWalker<Integer, Integer> {
 
         analyses.add(new VariantCounter());
         analyses.add(new VariantDBCoverage(knownSNPDBName));
+        analyses.add(new VariantDBCoverage(One1KGSNPNames));
 
         if ( samplesFile != null ) {
             //if ( numPeopleInPool < 1 )

java/test/org/broadinstitute/sting/playground/gatk/walkers/varianteval/VariantEvalWalkerIntegrationTest.java

@@ -18,8 +18,8 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
     @Test
     public void testEvalVariantROD() {
         HashMap<String, String> md5 = new HashMap<String, String>();
-        md5.put("", "d6b8c2d6c37d42d1ca2288799a8bd8e4");
+        md5.put("", "4b5a43cb4fa1b82e1bb361632754eaa1");
-        md5.put("-A", "0294b2e3915e88dfe2547e9db64ed1b3");
+        md5.put("-A", "7dd6aa5379ea5422fe7db4e79baad1f7");
 
         /**
          * the above MD5 was calculated from running the following command:
@@ -52,7 +52,7 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
     @Test
     public void testEvalVariantRODConfSix() {
         List<String> md5 = new ArrayList<String>();
-        md5.add("85cfefcac2dfb06545792605a3043a52");
+        md5.add("cfcc2a8e0587cb288f0472118a9bdc1d");
 
         /**
          * the above MD5 was calculated from running the following command:
@@ -84,7 +84,7 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
      @Test
     public void testEvalVariantRODOutputViolations() {
         List<String> md5 = new ArrayList<String>();
-        md5.add("e24732ffd95a78385a2c6986d1d3a359");
+        md5.add("ce708258676bbd4e71e8714828e7d695");
 
         /**
          * the above MD5 was calculated from running the following command:
@@ -116,7 +116,7 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
     @Test
     public void testEvalGenotypeROD() {
         List<String> md5 = new ArrayList<String>();
-        md5.add("010d1c7ce773b39f3de1355eb9682e4d");
+        md5.add("4df6a9c9d09a7bf034f29db24fb7a923");
         /**
          * the above MD5 was calculated after running the following command:
          *
@@ -150,7 +150,7 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
     @Test
     public void testEvalMarksGenotypingExample() {
         List<String> md5 = new ArrayList<String>();
-        md5.add("7d5a98c01051f96a684a383786da3d76");
+        md5.add("f79448f380cfb56be3585547b794a48c");
         /**
          * Run with the following commands:
          * 
@@ -177,7 +177,7 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
     @Test
 	public void testEvalRuntimeWithLotsOfIntervals() {
         List<String> md5 = new ArrayList<String>();
-        md5.add("d11ea079fc1835514d392056a2c2a28d");
+        md5.add("4d01485ed2d3cb5edbff24cc0c05b205");
         WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(
                 "-T VariantEval -R " + oneKGLocation + "reference/human_b36_both.fasta " +
                 "-B eval,Variants," + validationDataLocation + "NA12878.pilot_3.all.geli.calls " +
@@ -193,11 +193,11 @@ public class VariantEvalWalkerIntegrationTest extends WalkerTest {
     @Test
     public void testVCFVariantEvals() {
         HashMap<String, String> md5 = new HashMap<String, String>();
-        md5.put("", "3dda57ac7a9c8f3800726c9affb9d9bd");
+        md5.put("", "9702cba801cf216324607c8ecb7871e4");
-        md5.put("-A", "d985e61fd0d7fc34c9c1a553e2881c67");
+        md5.put("-A", "e9e02eba2e3423bde18ddb0a33a50e34");
-        md5.put("-A --includeFilteredRecords", "434c60986aa54c5fd07c22df1910ec44");
+        md5.put("-A --includeFilteredRecords", "f0cb51e2c7917e502ca510cc9a3382a9");
-        md5.put("-A --sampleName NA12878", "aff844b88f71824a6cd3cce553325b17");
+        md5.put("-A --sampleName NA12878", "b936723b0de69ed2698ab2af4b70a1e8");
-        md5.put("-A -vcfInfoSelector AF=0.50", "9ab9fa5d89cd6e3278d0d2b13cabbd51");
+        md5.put("-A -vcfInfoSelector AF=0.50", "a8299c6ba370bf9334327ec03e8885b7");
 
         for ( Map.Entry<String, String> e : md5.entrySet() ) {
             WalkerTest.WalkerTestSpec spec = new WalkerTest.WalkerTestSpec(