Merge branch 'master' of ssh://gsa4.broadinstitute.org/humgen/gsa-scr1/gsa-engineering/git/unstable

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/DeBruijnEdge.java

@@ -2,6 +2,9 @@ package org.broadinstitute.sting.gatk.walkers.haplotypecaller;
 
 import org.jgrapht.graph.DefaultDirectedGraph;
 
+import java.io.Serializable;
+import java.util.Comparator;
+
 /**
  * Created by IntelliJ IDEA.
  * User: ebanks
@@ -9,7 +12,7 @@ import org.jgrapht.graph.DefaultDirectedGraph;
  */
 
 // simple edge class for connecting nodes in the graph
-public class DeBruijnEdge implements Comparable<DeBruijnEdge> {
+public class DeBruijnEdge {
 
     private int multiplicity;
     private boolean isRef;
@@ -53,8 +56,10 @@ public class DeBruijnEdge implements Comparable<DeBruijnEdge> {
         return (graph.getEdgeSource(this).equals(graph2.getEdgeSource(edge))) && (graph.getEdgeTarget(this).equals(graph2.getEdgeTarget(edge)));
     }
 
+    public static class EdgeWeightComparator implements Comparator<DeBruijnEdge>, Serializable {
         @Override
-    public int compareTo( final DeBruijnEdge that ) {
+        public int compare(final DeBruijnEdge edge1, final DeBruijnEdge edge2) {
-        return this.multiplicity - that.multiplicity;
+            return edge1.multiplicity - edge2.multiplicity;
+        }
     }
 }

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/DeBruijnVertex.java

@@ -14,7 +14,7 @@ public class DeBruijnVertex {
     public final int kmer;
 
     public DeBruijnVertex( final byte[] sequence, final int kmer ) {
-        this.sequence = sequence;
+        this.sequence = sequence.clone();
         this.kmer = kmer;
     }
 
@@ -37,7 +37,7 @@ public class DeBruijnVertex {
     }
 
     public byte[] getSequence() {
-        return sequence;
+        return sequence.clone();
     }
 
     public byte[] getSuffix() {

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/KBestPaths.java

@@ -4,6 +4,7 @@ import org.apache.commons.lang.ArrayUtils;
 import org.broadinstitute.sting.utils.exceptions.ReviewedStingException;
 import org.jgrapht.graph.DefaultDirectedGraph;
 
+import java.io.Serializable;
 import java.util.*;
 
 /**
@@ -76,13 +77,15 @@ public class KBestPaths {
         }
     }
 
-    protected static class PathComparatorTotalScore implements Comparator<Path> {
+    protected static class PathComparatorTotalScore implements Comparator<Path>, Serializable {
+        @Override
         public int compare(final Path path1, final Path path2) {
             return path1.totalScore - path2.totalScore;
         }
     }
 
-    //protected static class PathComparatorLowestEdge implements Comparator<Path> {
+    //protected static class PathComparatorLowestEdge implements Comparator<Path>, Serializable {
+    //    @Override
     //    public int compare(final Path path1, final Path path2) {
     //        return path2.lowestEdge - path1.lowestEdge;
     //    }
@@ -124,7 +127,7 @@ public class KBestPaths {
             // recursively run DFS
             final ArrayList<DeBruijnEdge> edgeArrayList = new ArrayList<DeBruijnEdge>();
             edgeArrayList.addAll(graph.outgoingEdgesOf(path.lastVertex));
-            Collections.sort(edgeArrayList);
+            Collections.sort(edgeArrayList, new DeBruijnEdge.EdgeWeightComparator());
             Collections.reverse(edgeArrayList);
             for ( final DeBruijnEdge edge : edgeArrayList ) {
                 // make sure the edge is not already in the path

protected/java/src/org/broadinstitute/sting/gatk/walkers/haplotypecaller/LikelihoodCalculationEngine.java

@@ -78,10 +78,10 @@ public class LikelihoodCalculationEngine {
         PairHMM.initializeArrays(matchMetricArray, XMetricArray, YMetricArray, X_METRIC_LENGTH);
 
         // for each sample's reads
-        for( final String sample : perSampleReadList.keySet() ) {
+        for( final Map.Entry<String, ArrayList<GATKSAMRecord>> sampleEntry : perSampleReadList.entrySet() ) {
             //if( DEBUG ) { System.out.println("Evaluating sample " + sample + " with " + perSampleReadList.get( sample ).size() + " passing reads"); }
             // evaluate the likelihood of the reads given those haplotypes
-            computeReadLikelihoods( haplotypes, perSampleReadList.get(sample), sample, matchMetricArray, XMetricArray, YMetricArray );
+            computeReadLikelihoods( haplotypes, sampleEntry.getValue(), sampleEntry.getKey(), matchMetricArray, XMetricArray, YMetricArray );
         }
     }
 

public/java/src/org/broadinstitute/sting/gatk/traversals/TraverseActiveRegions.java

@@ -12,6 +12,7 @@ import org.broadinstitute.sting.gatk.walkers.DataSource;
 import org.broadinstitute.sting.gatk.walkers.Walker;
 import org.broadinstitute.sting.utils.GenomeLoc;
 import org.broadinstitute.sting.utils.GenomeLocSortedSet;
+import org.broadinstitute.sting.utils.activeregion.ActiveRegion;
 import org.broadinstitute.sting.utils.activeregion.ActivityProfile;
 import org.broadinstitute.sting.utils.activeregion.ActivityProfileResult;
 import org.broadinstitute.sting.utils.pileup.PileupElement;
@@ -31,7 +32,7 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
      */
     protected final static Logger logger = Logger.getLogger(TraversalEngine.class);
 
-    private final LinkedList<org.broadinstitute.sting.utils.activeregion.ActiveRegion> workQueue = new LinkedList<org.broadinstitute.sting.utils.activeregion.ActiveRegion>();
+    private final LinkedList<ActiveRegion> workQueue = new LinkedList<ActiveRegion>();
     private final LinkedHashSet<GATKSAMRecord> myReads = new LinkedHashSet<GATKSAMRecord>();
 
     @Override
@@ -110,18 +111,18 @@ public class TraverseActiveRegions <M,T> extends TraversalEngine<M,T,ActiveRegio
             // add these blocks of work to the work queue
             // band-pass filter the list of isActive probabilities and turn into active regions
             final ActivityProfile bandPassFiltered = profile.bandPassFilter();
-            final List<org.broadinstitute.sting.utils.activeregion.ActiveRegion> activeRegions = bandPassFiltered.createActiveRegions( activeRegionExtension, maxRegionSize );
+            final List<ActiveRegion> activeRegions = bandPassFiltered.createActiveRegions( activeRegionExtension, maxRegionSize );
 
             // add active regions to queue of regions to process
             // first check if can merge active regions over shard boundaries
             if( !activeRegions.isEmpty() ) {
                 if( !workQueue.isEmpty() ) {
-                    final org.broadinstitute.sting.utils.activeregion.ActiveRegion last = workQueue.getLast();
+                    final ActiveRegion last = workQueue.getLast();
-                    final org.broadinstitute.sting.utils.activeregion.ActiveRegion first = activeRegions.get(0);
+                    final ActiveRegion first = activeRegions.get(0);
                     if( last.isActive == first.isActive && last.getLocation().contiguousP(first.getLocation()) && last.getLocation().size() + first.getLocation().size() <= maxRegionSize ) {
                         workQueue.removeLast();
                         activeRegions.remove(first);
-                        workQueue.add( new org.broadinstitute.sting.utils.activeregion.ActiveRegion(last.getLocation().union(first.getLocation()), first.isActive, this.engine.getGenomeLocParser(), activeRegionExtension) );
+                        workQueue.add( new ActiveRegion(last.getLocation().union(first.getLocation()), first.isActive, this.engine.getGenomeLocParser(), activeRegionExtension) );
                     }
                 }
                 workQueue.addAll( activeRegions );

public/java/src/org/broadinstitute/sting/gatk/walkers/annotator/HaplotypeScore.java

@@ -48,6 +48,7 @@ import org.broadinstitute.sting.utils.variantcontext.Allele;
 import org.broadinstitute.sting.utils.variantcontext.Genotype;
 import org.broadinstitute.sting.utils.variantcontext.VariantContext;
 
+import java.io.Serializable;
 import java.util.*;
 
 /**
@@ -111,7 +112,7 @@ public class HaplotypeScore extends InfoFieldAnnotation implements StandardAnnot
         return map;
     }
 
-    private static class HaplotypeComparator implements Comparator<Haplotype> {
+    private static class HaplotypeComparator implements Comparator<Haplotype>, Serializable {
 
         public int compare(Haplotype a, Haplotype b) {
             if (a.getQualitySum() < b.getQualitySum())

public/java/src/org/broadinstitute/sting/gatk/walkers/bqsr/StandardRecalibrationEngine.java

@@ -42,7 +42,7 @@ public class StandardRecalibrationEngine implements RecalibrationEngine, PublicP
     protected RecalibrationTables recalibrationTables;
 
     public void initialize(final Covariate[] covariates, final RecalibrationTables recalibrationTables) {
-        this.covariates = covariates;
+        this.covariates = covariates.clone();
         this.recalibrationTables = recalibrationTables;
     }
 

public/java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/Tranche.java

@@ -29,10 +29,7 @@ import org.broadinstitute.sting.utils.exceptions.ReviewedStingException;
 import org.broadinstitute.sting.utils.exceptions.UserException;
 import org.broadinstitute.sting.utils.text.XReadLines;
 
-import java.io.ByteArrayOutputStream;
+import java.io.*;
-import java.io.File;
-import java.io.FileNotFoundException;
-import java.io.PrintStream;
 import java.util.*;
 
 /**
@@ -41,7 +38,7 @@ import java.util.*;
  * Date: Mar 10, 2011
  */
 
-public class Tranche implements Comparable<Tranche> {
+public class Tranche {
     private static final int CURRENT_VERSION = 5;
 
     public double ts, minVQSLod, knownTiTv, novelTiTv;
@@ -83,10 +80,14 @@ public class Tranche implements Comparable<Tranche> {
         return accessibleTruthSites > 0 ? callsAtTruthSites / (1.0*accessibleTruthSites) : 0.0;
     }
 
-    public int compareTo(Tranche other) {
+    public static class TrancheTruthSensitivityComparator implements Comparator<Tranche>, Serializable {
-        return Double.compare(this.ts,  other.ts);
+        @Override
+        public int compare(final Tranche tranche1, final Tranche tranche2) {
+            return Double.compare(tranche1.ts, tranche2.ts);
+        }
     }
 
+    @Override
     public String toString() {
         return String.format("Tranche ts=%.2f minVQSLod=%.4f known=(%d @ %.4f) novel=(%d @ %.4f) truthSites(%d accessible, %d called), name=%s]",
                 ts, minVQSLod, numKnown, knownTiTv, numNovel, novelTiTv, accessibleTruthSites, callsAtTruthSites, name);
@@ -102,7 +103,7 @@ public class Tranche implements Comparable<Tranche> {
         final ByteArrayOutputStream bytes = new ByteArrayOutputStream();
         final PrintStream stream = new PrintStream(bytes);
 
-        Collections.sort(tranches);
+        Collections.sort( tranches, new TrancheTruthSensitivityComparator() );
 
         stream.println("# Variant quality score tranches file");
         stream.println("# Version number " + CURRENT_VERSION);
@@ -183,7 +184,7 @@ public class Tranche implements Comparable<Tranche> {
                 }
             }
 
-            Collections.sort(tranches);
+            Collections.sort( tranches, new TrancheTruthSensitivityComparator() );
             return tranches;
         } catch( FileNotFoundException e ) {
             throw new UserException.CouldNotReadInputFile(f, e);

public/java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/TrancheManager.java

@@ -146,7 +146,7 @@ public class TrancheManager {
     public static List<Tranche> findTranches( final ArrayList<VariantDatum> data, final double[] trancheThresholds, final SelectionMetric metric, final VariantRecalibratorArgumentCollection.Mode model, final File debugFile ) {
         logger.info(String.format("Finding %d tranches for %d variants", trancheThresholds.length, data.size()));
 
-        Collections.sort(data);
+        Collections.sort( data, new VariantDatum.VariantDatumLODComparator() );
         metric.calculateRunningMetric(data);
 
         if ( debugFile != null) { writeTranchesDebuggingInfo(debugFile, data, metric); }

public/java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantDataManager.java

@@ -158,7 +158,7 @@ public class VariantDataManager {
         logger.info( "Found " + numBadSitesAdded + " variants overlapping bad sites training tracks." );
 
         // Next sort the variants by the LOD coming from the positive model and add to the list the bottom X percent of variants
-        Collections.sort( data );
+        Collections.sort( data, new VariantDatum.VariantDatumLODComparator() );
         final int numToAdd = Math.max( minimumNumber - trainingData.size(), Math.round((float)bottomPercentage * data.size()) );
         if( numToAdd > data.size() ) {
             throw new UserException.BadInput( "Error during negative model training. Minimum number of variants to use in training is larger than the whole call set. One can attempt to lower the --minNumBadVariants arugment but this is unsafe." );

public/java/src/org/broadinstitute/sting/gatk/walkers/variantrecalibration/VariantDatum.java

@@ -27,13 +27,16 @@ package org.broadinstitute.sting.gatk.walkers.variantrecalibration;
 
 import org.broadinstitute.sting.utils.GenomeLoc;
 
+import java.io.Serializable;
+import java.util.Comparator;
+
 /**
  * Created by IntelliJ IDEA.
  * User: rpoplin
  * Date: Mar 4, 2011
  */
 
-public class VariantDatum implements Comparable<VariantDatum> {
+public class VariantDatum {
 
     public double[] annotations;
     public boolean[] isNull;
@@ -52,8 +55,10 @@ public class VariantDatum implements Comparable<VariantDatum> {
     public int worstAnnotation;
     public MultivariateGaussian assignment; // used in K-means implementation 
 
+    public static class VariantDatumLODComparator implements Comparator<VariantDatum>, Serializable {
         @Override
-    public int compareTo( final VariantDatum other ) {
+        public int compare(final VariantDatum datum1, final VariantDatum datum2) {
-        return Double.compare(this.lod, other.lod);
+            return Double.compare(datum1.lod, datum2.lod);
+        }
     }
 }

public/java/src/org/broadinstitute/sting/utils/Haplotype.java

@@ -55,22 +55,22 @@ public class Haplotype {
      * @param bases bases
      * @param qual  qual
      */
-    public Haplotype(byte[] bases, int qual) {
+    public Haplotype( final byte[] bases, final int qual ) {
-        this.bases = bases;
+        this.bases = bases.clone();
         quals = new double[bases.length];
         Arrays.fill(quals, (double)qual);
     }
 
-    public Haplotype(byte[] bases, double[] quals) {
+    public Haplotype( final byte[] bases, final double[] quals ) {
-        this.bases = bases;
+        this.bases = bases.clone();
-        this.quals = quals;
+        this.quals = quals.clone();
     }
 
-    public Haplotype(byte[] bases) {
+    public Haplotype( final byte[] bases ) {
         this(bases, 0);
     }
 
-    public Haplotype(byte[] bases, GenomeLoc loc) {
+    public Haplotype( final byte[] bases, final GenomeLoc loc ) {
         this(bases);
         this.genomeLocation = loc;
     }
@@ -140,10 +140,10 @@ public class Haplotype {
     }
 
     public double[] getQuals() {
-        return quals;
+        return quals.clone();
     }
     public byte[] getBases() {
-        return bases;
+        return bases.clone();
     }
 
     public long getStartPosition() {

public/java/src/org/broadinstitute/sting/utils/MannWhitneyU.java

@@ -11,6 +11,7 @@ import org.broadinstitute.sting.gatk.GenomeAnalysisEngine;
 import org.broadinstitute.sting.utils.collections.Pair;
 import org.broadinstitute.sting.utils.exceptions.StingException;
 
+import java.io.Serializable;
 import java.util.Comparator;
 import java.util.TreeSet;
 
@@ -434,12 +435,14 @@ public class MannWhitneyU {
      * A comparator class which uses dithering on tie-breaking to ensure that the internal treeset drops no values
      * and to ensure that rank ties are broken at random.
      */
-    private class DitheringComparator implements Comparator<Pair<Number,USet>> {
+    private static class DitheringComparator implements Comparator<Pair<Number,USet>>, Serializable {
 
         public DitheringComparator() {}
 
+        @Override
         public boolean equals(Object other) { return false; }
 
+        @Override
         public int compare(Pair<Number,USet> left, Pair<Number,USet> right) {
             double comp = Double.compare(left.first.doubleValue(),right.first.doubleValue());
             if ( comp > 0 ) { return 1; }

public/java/src/org/broadinstitute/sting/utils/activeregion/ActiveRegion.java

@@ -15,7 +15,7 @@ import java.util.ArrayList;
  * Date: 1/4/12
  */
 
-public class ActiveRegion implements HasGenomeLocation, Comparable<ActiveRegion> {
+public class ActiveRegion implements HasGenomeLocation {
 
     private final ArrayList<GATKSAMRecord> reads = new ArrayList<GATKSAMRecord>();
     private final GenomeLoc activeRegionLoc;
@@ -73,10 +73,6 @@ public class ActiveRegion implements HasGenomeLocation, Comparable<ActiveRegion>
                Math.min(referenceReader.getSequenceDictionary().getSequence(fullExtentReferenceLoc.getContig()).getSequenceLength(), fullExtentReferenceLoc.getStop() + padding) ).getBases();
     }
 
-    @Override
-    public int compareTo( final ActiveRegion other ) {
-        return this.getLocation().compareTo(other.getLocation());
-    }
 
     @Override
     public GenomeLoc getLocation() { return activeRegionLoc; }
@@ -97,4 +93,19 @@ public class ActiveRegion implements HasGenomeLocation, Comparable<ActiveRegion>
         if ( extendedLoc.compareTo(other.extendedLoc) != 0 ) return false;
         return true;
     }
+
+    /**
+     * A comparator class which is used to sort ActiveRegions by their start location
+     */
+    /*
+    public static class ActiveRegionStartLocationComparator implements Comparator<ActiveRegion> {
+
+        public ActiveRegionStartLocationComparator() {}
+
+        @Override
+        public int compare(final ActiveRegion left, final ActiveRegion right) {
+            return left.getLocation().compareTo(right.getLocation());
+        }
+    }
+    */
 }