From e29e8e52b9be731a065c67cf255f116be7715e7d Mon Sep 17 00:00:00 2001
From: hanna <hanna@348d0f76-0448-11de-a6fe-93d51630548a>
Date: Wed, 23 Dec 2009 00:48:06 +0000
Subject: [PATCH] Multithreading support for the unified genotyper.  Tests on a
 10Mbase region on pilot 1 show a 6.8x improvement when running 8 ways
 parallel.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2430 348d0f76-0448-11de-a6fe-93d51630548a
---
 .../EmpiricalSubstitutionProbabilities.java   | 12 +++---
 .../walkers/genotyper/UnifiedGenotyper.java   | 37 ++++++++++---------
 2 files changed, 25 insertions(+), 24 deletions(-)

diff --git a/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/EmpiricalSubstitutionProbabilities.java b/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/EmpiricalSubstitutionProbabilities.java
index f7e56c7de..4045e2ed4 100755
--- a/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/EmpiricalSubstitutionProbabilities.java
+++ b/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/EmpiricalSubstitutionProbabilities.java
@@ -48,17 +48,17 @@ public class EmpiricalSubstitutionProbabilities extends FourBaseProbabilities {
         }
     }
 
-    private static SAMRecord lastReadForPL = null;
-    private static SequencerPlatform plOfLastRead = null;
+    private static ThreadLocal<SAMRecord> lastReadForPL = new ThreadLocal<SAMRecord>();
+    private static ThreadLocal<SequencerPlatform> plOfLastRead = new ThreadLocal<SequencerPlatform>();
     public static SequencerPlatform getReadSequencerPlatform( SAMRecord read ) {
-        if ( lastReadForPL != read ) {
-            lastReadForPL = read;
+        if ( lastReadForPL.get() != read ) {
+            lastReadForPL.set(read);
             SAMReadGroupRecord readGroup = read.getReadGroup();
             final String platformName = readGroup == null ? null : readGroup.getPlatform();
-            plOfLastRead = standardizeSequencerPlatform(platformName);
+            plOfLastRead.set(standardizeSequencerPlatform(platformName));
         }
 
-        return plOfLastRead;
+        return plOfLastRead.get();
     }
 
     public int getReadSequencerPlatformIndex( SAMRecord read ) {
diff --git a/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyper.java b/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyper.java
index 232d67dcb..626ff2f65 100755
--- a/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyper.java
+++ b/java/src/org/broadinstitute/sting/gatk/walkers/genotyper/UnifiedGenotyper.java
@@ -57,7 +57,7 @@ public class UnifiedGenotyper extends LocusWalker<Pair<VariationCall, List<Genot
     public GenotypeWriter writer = null;
 
     // the model used for calculating genotypes
-    private GenotypeCalculationModel gcm;
+    private ThreadLocal<GenotypeCalculationModel> gcm = new ThreadLocal<GenotypeCalculationModel>();
 
     // samples in input
     private Set<String> samples;
@@ -75,7 +75,7 @@ public class UnifiedGenotyper extends LocusWalker<Pair<VariationCall, List<Genot
      *
      **/
     public void setUnifiedArgumentCollection(UnifiedArgumentCollection UAC) {
-        gcm.close();
+        //gcm.close();
         this.UAC = UAC;
         initialize();
     }
@@ -111,20 +111,6 @@ public class UnifiedGenotyper extends LocusWalker<Pair<VariationCall, List<Genot
         // for ( String sample : samples )
         //     logger.debug("SAMPLE: " + sample);
 
-        GenotypeWriterFactory.GENOTYPE_FORMAT format = GenotypeWriterFactory.GENOTYPE_FORMAT.VCF;
-        if(writer != null) {
-            if(writer instanceof VCFGenotypeWriter)
-                format = GenotypeWriterFactory.GENOTYPE_FORMAT.VCF;
-            else if(writer instanceof GLFGenotypeWriter)
-                format = GenotypeWriterFactory.GENOTYPE_FORMAT.GLF;
-            else if(writer instanceof GeliGenotypeWriter)
-                format = GenotypeWriterFactory.GENOTYPE_FORMAT.GELI;
-            else
-                throw new StingException("Unsupported genotype format: " + writer.getClass().getName());
-        }
-
-        gcm = GenotypeCalculationModelFactory.makeGenotypeCalculation(samples, logger, UAC, format);
-
         // *** If we were called by another walker, then we don't ***
         // *** want to do any of the other initialization steps.  ***
         if ( writer == null )
@@ -191,6 +177,21 @@ public class UnifiedGenotyper extends LocusWalker<Pair<VariationCall, List<Genot
      * @param rawContext contextual information around the locus
      */
     public Pair<VariationCall, List<Genotype>> map(RefMetaDataTracker tracker, ReferenceContext refContext, AlignmentContext rawContext) {
+        GenotypeWriterFactory.GENOTYPE_FORMAT format = GenotypeWriterFactory.GENOTYPE_FORMAT.VCF;
+        if(writer != null) {
+            if(writer instanceof VCFGenotypeWriter)
+                format = GenotypeWriterFactory.GENOTYPE_FORMAT.VCF;
+            else if(writer instanceof GLFGenotypeWriter)
+                format = GenotypeWriterFactory.GENOTYPE_FORMAT.GLF;
+            else if(writer instanceof GeliGenotypeWriter)
+                format = GenotypeWriterFactory.GENOTYPE_FORMAT.GELI;
+            else
+                throw new StingException("Unsupported genotype format: " + writer.getClass().getName());
+        }
+        
+        if(gcm.get() == null)
+            gcm.set(GenotypeCalculationModelFactory.makeGenotypeCalculation(samples, logger, UAC, format));
+
         char ref = Character.toUpperCase(refContext.getBase());
         if ( !BaseUtils.isRegularBase(ref) )
             return null;
@@ -218,7 +219,7 @@ public class UnifiedGenotyper extends LocusWalker<Pair<VariationCall, List<Genot
             return null;
 
         DiploidGenotypePriors priors = new DiploidGenotypePriors(ref, UAC.heterozygosity, DiploidGenotypePriors.PROB_OF_TRISTATE_GENOTYPE);
-        Pair<VariationCall, List<Genotype>> call = gcm.calculateGenotype(tracker, ref, rawContext.getLocation(), stratifiedContexts, priors);
+        Pair<VariationCall, List<Genotype>> call = gcm.get().calculateGenotype(tracker, ref, rawContext.getLocation(), stratifiedContexts, priors);
 
         // annotate the call, if possible
         if ( call != null && call.first != null && call.first instanceof ArbitraryFieldsBacked ) {
@@ -280,7 +281,7 @@ public class UnifiedGenotyper extends LocusWalker<Pair<VariationCall, List<Genot
 
     // Close any file writers
     public void onTraversalDone(Integer sum) {
-        gcm.close();
+        //gcm.close();
         logger.info("Processed " + sum + " loci that are callable for SNPs");
     }