Optimizations across the board, with help from Guillermo, Matt, and JProfiler. Too tired to give details now.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@4535 348d0f76-0448-11de-a6fe-93d51630548a

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

@@ -44,6 +44,7 @@ import org.broadinstitute.sting.utils.sam.ReadUtils;
 public class HaplotypeScore implements InfoFieldAnnotation, StandardAnnotation {
     private final static boolean DEBUG = false;
     private final static int MIN_CONTEXT_WING_SIZE = 10;
+    private final static int MAX_CONSENSUS_HAPLOTYPES_TO_CONSIDER = 20;
     private final static char REGEXP_WILDCARD = '.';
 
     public Map<String, Object> annotate(RefMetaDataTracker tracker, ReferenceContext ref, Map<String, StratifiedAlignmentContext> stratifiedContexts, VariantContext vc) {
@@ -124,7 +125,7 @@ public class HaplotypeScore implements InfoFieldAnnotation, StandardAnnotation {
           */
             }
 
-            if (!foundHaplotypeMatch) {
+            if (!foundHaplotypeMatch && haplotypeList.size() < MAX_CONSENSUS_HAPLOTYPES_TO_CONSIDER) {
                 haplotypeList.add(elem);
             }
         }
@@ -293,7 +294,7 @@ public class HaplotypeScore implements InfoFieldAnnotation, StandardAnnotation {
             byte haplotypeBase = haplotypeBases[i];
             byte readBase = readBases[baseOffset];
 
-            boolean matched = BaseUtils.basesAreEqual(readBase, haplotypeBase );
+            boolean matched = readBase == haplotypeBase;
             double e = QualityUtils.qualToErrorProb(readQuals[baseOffset]);
             expected += e;
             mismatches += matched ? e : 1 - e / 3;

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

@@ -23,7 +23,13 @@ public class RMSMappingQuality implements InfoFieldAnnotation, StandardAnnotatio
         if ( stratifiedContexts.size() == 0 )
             return null;
 
-        ArrayList<Integer> qualities = new ArrayList<Integer>();
+        int totalSize = 0;
+        for ( StratifiedAlignmentContext context : stratifiedContexts.values() )
+            totalSize += context.getContext(StratifiedAlignmentContext.StratifiedContextType.COMPLETE).size();
+
+        int[] qualities = new int[totalSize];
+        int index = 0;
+
         for ( Map.Entry<String, StratifiedAlignmentContext> sample : stratifiedContexts.entrySet() ) {
             AlignmentContext context = sample.getValue().getContext(StratifiedAlignmentContext.StratifiedContextType.COMPLETE);
             ReadBackedPileup pileup = null;
@@ -34,14 +40,11 @@ public class RMSMappingQuality implements InfoFieldAnnotation, StandardAnnotatio
 
             if (pileup != null) {
                 for (PileupElement p : pileup )
-                    qualities.add(p.getRead().getMappingQuality());
+                    qualities[index++] = p.getRead().getMappingQuality();
             }
         }
-        int[] quals = new int[qualities.size()];
+
-        int index = 0;
+        double rms = MathUtils.rms(qualities);
-        for ( Integer i : qualities )
-            quals[index++] = i;
-        double rms = MathUtils.rms(quals);
         Map<String, Object> map = new HashMap<String, Object>();
         map.put(getKeyNames().get(0), String.format("%.2f", rms));
         return map;

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

@@ -318,12 +318,6 @@ public class UnifiedGenotyperEngine {
 
             ReadBackedPileup pileup = rawContext.getBasePileup();
 
-            // filter the reads
-            filterPileup(pileup, badBaseFilter);
-
-            // filter the context based on bad mates and mismatch rate
-            //pileup = pileup.getFilteredPileup(badReadPileupFilter);
-
             // don't call when there is no coverage
             if ( pileup.size() == 0 && !UAC.ALL_BASES_MODE )
                 return null;
@@ -335,6 +329,9 @@ public class UnifiedGenotyperEngine {
 
             // stratify the AlignmentContext and cut by sample
             stratifiedContexts = StratifiedAlignmentContext.splitContextBySample(pileup, UAC.ASSUME_SINGLE_SAMPLE);
+
+            // filter the reads
+            filterPileup(stratifiedContexts, badBaseFilter);
         }
 
         return stratifiedContexts;
@@ -405,11 +402,14 @@ public class UnifiedGenotyperEngine {
         verboseWriter.println();
     }
 
-    private void filterPileup(ReadBackedPileup pileup, BadBaseFilter badBaseFilter) {
+    private void filterPileup(Map<String, StratifiedAlignmentContext> stratifiedContexts, BadBaseFilter badBaseFilter) {
-        for ( PileupElement p : pileup ) {
+        for ( StratifiedAlignmentContext context : stratifiedContexts.values() ) {
-            final SAMRecord read = p.getRead();
+            ReadBackedPileup pileup = context.getContext(StratifiedAlignmentContext.StratifiedContextType.COMPLETE).getBasePileup();
-            if ( read instanceof GATKSAMRecord )
+            for ( PileupElement p : pileup ) {
-                ((GATKSAMRecord)read).setGoodBases(badBaseFilter, true);
+               final SAMRecord read = p.getRead();
+                if ( read instanceof GATKSAMRecord )
+                    ((GATKSAMRecord)read).setGoodBases(badBaseFilter, true);
+            }
         }
     }