Command line fun

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/AssemblyBasedCallerArgumentCollection.java

@@ -29,8 +29,11 @@ public abstract class AssemblyBasedCallerArgumentCollection extends StandardCall
     @ArgumentCollection
     public AssemblyRegionTrimmerArgumentCollection assemblyRegionTrimmerArgs = new AssemblyRegionTrimmerArgumentCollection();
 
+    protected boolean useMutectAssemblerArgumentCollection() { return false; }
+
     @ArgumentCollection
-    public ReadThreadingAssemblerArgumentCollection assemblerArgs = new ReadThreadingAssemblerArgumentCollection();
+    public ReadThreadingAssemblerArgumentCollection assemblerArgs = useMutectAssemblerArgumentCollection() ?
+            new MutectReadThreadingAssemblerArgumentCollection() : new HaplotypeCallerReadThreadingAssemblerArgumentCollection();
 
     @ArgumentCollection
     public LikelihoodEngineArgumentCollection likelihoodArgs = new LikelihoodEngineArgumentCollection();

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/AssemblyBasedCallerUtils.java

@@ -189,20 +189,10 @@ public static ReadLikelihoodCalculationEngine createLikelihoodCalculationEngine(
 
     public static ReadThreadingAssembler createReadThreadingAssembler(final AssemblyBasedCallerArgumentCollection args) {
         final ReadThreadingAssemblerArgumentCollection rtaac = args.assemblerArgs;
-        final ReadThreadingAssembler assemblyEngine = new ReadThreadingAssembler(rtaac.maxNumHaplotypesInPopulation, rtaac.kmerSizes,
-                rtaac.dontIncreaseKmerSizesForCycles, rtaac.allowNonUniqueKmersInRef, rtaac.numPruningSamples, rtaac.minPruneFactor,
-                rtaac.useAdaptivePruning, rtaac.initialErrorRateForPruning, rtaac.pruningLog10OddsThreshold, rtaac.maxUnprunedVariants);
-        assemblyEngine.setErrorCorrectKmers(rtaac.errorCorrectKmers);
+        final ReadThreadingAssembler assemblyEngine = rtaac.makeReadThreadingAssembler();
         assemblyEngine.setDebug(args.debug);
-        assemblyEngine.setDebugGraphTransformations(rtaac.debugGraphTransformations);
-        assemblyEngine.setRecoverDanglingBranches(!rtaac.doNotRecoverDanglingBranches);
-        assemblyEngine.setMinDanglingBranchLength(rtaac.minDanglingBranchLength);
         assemblyEngine.setMinBaseQualityToUseInAssembly(args.minBaseQualityScore);
 
-        if ( rtaac.graphOutput != null ) {
-            assemblyEngine.setGraphWriter(new File(rtaac.graphOutput));
-        }
-
         return assemblyEngine;
     }
 

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/HaplotypeCallerEngine.java

@@ -263,7 +263,7 @@ private void validateAndInitializeArgs() {
             hcArgs.setSampleContamination(AlleleBiasedDownsamplingUtils.loadContaminationFile(hcArgs.CONTAMINATION_FRACTION_FILE, hcArgs.CONTAMINATION_FRACTION, sampleSet, logger));
         }
 
-        if ( hcArgs.genotypingOutputMode == GenotypingOutputMode.GENOTYPE_GIVEN_ALLELES && hcArgs.assemblerArgs.consensusMode ) {
+        if ( hcArgs.genotypingOutputMode == GenotypingOutputMode.GENOTYPE_GIVEN_ALLELES && hcArgs.assemblerArgs.consensusMode() ) {
             throw new UserException("HaplotypeCaller cannot be run in both GENOTYPE_GIVEN_ALLELES mode and in consensus mode at the same time. Please choose one or the other.");
         }
 
@@ -604,7 +604,7 @@ public List<VariantContext> callRegion(final AssemblyRegion region, final Featur
                 assemblyResult.getPaddedReferenceLoc(),
                 regionForGenotyping.getSpan(),
                 features,
-                (hcArgs.assemblerArgs.consensusMode ? Collections.<VariantContext>emptyList() : givenAlleles),
+                (hcArgs.assemblerArgs.consensusMode() ? Collections.<VariantContext>emptyList() : givenAlleles),
                 emitReferenceConfidence(),
                 hcArgs.maxMnpDistance,
                 readsHeader,

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/HaplotypeCallerReadThreadingAssemblerArgumentCollection.java

@@ -0,0 +1,60 @@
+package org.broadinstitute.hellbender.tools.walkers.haplotypecaller;
+
+import org.broadinstitute.barclay.argparser.Advanced;
+import org.broadinstitute.barclay.argparser.Argument;
+import org.broadinstitute.barclay.argparser.Hidden;
+import org.broadinstitute.hellbender.tools.walkers.haplotypecaller.readthreading.ReadThreadingAssembler;
+
+import java.io.File;
+
+public class HaplotypeCallerReadThreadingAssemblerArgumentCollection extends ReadThreadingAssemblerArgumentCollection {
+    /**
+     * A single edge multiplicity cutoff for pruning doesn't work in samples with variable depths, for example exomes
+     * and RNA.  This parameter enables the probabilistic algorithm for pruning the assembly graph that considers the
+     * likelihood that each chain in the graph comes from real variation.
+     */
+    @Advanced
+    @Argument(fullName="adaptive-pruning", doc = "Use Mutect2's adaptive graph pruning algorithm", optional = true)
+    public boolean useAdaptivePruning = false;
+
+    /**
+     * By default, the read threading assembler will attempt to recover dangling heads and tails. See the `minDanglingBranchLength` argument documentation for more details.
+     */
+    @Hidden
+    @Argument(fullName="do-not-recover-dangling-branches", doc="Disable dangling head and tail recovery", optional = true)
+    public boolean doNotRecoverDanglingBranches = false;
+
+    /**
+     * As of version 3.3, this argument is no longer needed because dangling end recovery is now the default behavior. See GATK 3.3 release notes for more details.
+     */
+    @Deprecated
+    @Argument(fullName="recover-dangling-heads", doc="This argument is deprecated since version 3.3", optional = true)
+    public boolean DEPRECATED_RecoverDanglingHeads = false;
+
+    /**
+     * This argument is specifically intended for 1000G consensus analysis mode. Setting this flag will inject all
+     * provided alleles to the assembly graph but will not forcibly genotype all of them.
+     */
+    @Advanced
+    @Argument(fullName="consensus", doc="1000G consensus mode", optional = true)
+    public boolean consensusMode = false;
+
+    @Override
+    public ReadThreadingAssembler makeReadThreadingAssembler() {
+        final ReadThreadingAssembler assemblyEngine = new ReadThreadingAssembler(maxNumHaplotypesInPopulation, kmerSizes,
+                dontIncreaseKmerSizesForCycles, allowNonUniqueKmersInRef, numPruningSamples, useAdaptivePruning ? 0 : minPruneFactor,
+                useAdaptivePruning, initialErrorRateForPruning, pruningLog10OddsThreshold, maxUnprunedVariants);
+        assemblyEngine.setDebugGraphTransformations(debugGraphTransformations);
+        assemblyEngine.setRecoverDanglingBranches(!doNotRecoverDanglingBranches);
+        assemblyEngine.setMinDanglingBranchLength(minDanglingBranchLength);
+
+        if ( graphOutput != null ) {
+            assemblyEngine.setGraphWriter(new File(graphOutput));
+        }
+
+        return assemblyEngine;
+    }
+
+    @Override
+    public boolean consensusMode() { return consensusMode; }
+}

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/MutectReadThreadingAssemblerArgumentCollection.java

@@ -0,0 +1,35 @@
+package org.broadinstitute.hellbender.tools.walkers.haplotypecaller;
+
+import org.broadinstitute.barclay.argparser.Advanced;
+import org.broadinstitute.barclay.argparser.Argument;
+import org.broadinstitute.hellbender.tools.walkers.haplotypecaller.readthreading.ReadThreadingAssembler;
+
+import java.io.File;
+
+public class MutectReadThreadingAssemblerArgumentCollection extends ReadThreadingAssemblerArgumentCollection {
+
+    /**
+     * A single edge multiplicity cutoff for pruning doesn't work in samples with variable depths, for example exomes
+     * and RNA.  This parameter disables the probabilistic algorithm for pruning the assembly graph that considers the
+     * likelihood that each chain in the graph comes from real variation, and instead uses a simple multiplicity cutoff.
+     */
+    @Advanced
+    @Argument(fullName="disable-adaptive-pruning", doc = "Disable the adaptive algorithm for pruning paths in the graph", optional = true)
+    public boolean disableAdaptivePruning = false;
+
+    @Override
+    public ReadThreadingAssembler makeReadThreadingAssembler() {
+        final ReadThreadingAssembler assemblyEngine = new ReadThreadingAssembler(maxNumHaplotypesInPopulation, kmerSizes,
+                dontIncreaseKmerSizesForCycles, allowNonUniqueKmersInRef, numPruningSamples, disableAdaptivePruning ? minPruneFactor : 0,
+                !disableAdaptivePruning, initialErrorRateForPruning, pruningLog10OddsThreshold, maxUnprunedVariants);
+        assemblyEngine.setDebugGraphTransformations(debugGraphTransformations);
+        assemblyEngine.setRecoverDanglingBranches(true);
+        assemblyEngine.setMinDanglingBranchLength(minDanglingBranchLength);
+
+        if ( graphOutput != null ) {
+            assemblyEngine.setGraphWriter(new File(graphOutput));
+        }
+
+        return assemblyEngine;
+    }
+}

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/ReadThreadingAssemblerArgumentCollection.java

@@ -4,14 +4,15 @@
 import org.broadinstitute.barclay.argparser.Advanced;
 import org.broadinstitute.barclay.argparser.Argument;
 import org.broadinstitute.barclay.argparser.Hidden;
+import org.broadinstitute.hellbender.tools.walkers.haplotypecaller.readthreading.ReadThreadingAssembler;
 
 import java.io.Serializable;
 import java.util.List;
 
 /**
  * Set of arguments related to the {@link org.broadinstitute.hellbender.tools.walkers.haplotypecaller.readthreading.ReadThreadingAssembler}
  */
-public final class ReadThreadingAssemblerArgumentCollection implements Serializable {
+public abstract class ReadThreadingAssemblerArgumentCollection implements Serializable {
     private static final long serialVersionUID = 1L;
 
     // -----------------------------------------------------------------------------------------------
@@ -48,20 +49,6 @@ public final class ReadThreadingAssemblerArgumentCollection implements Serializa
     @Argument(fullName="num-pruning-samples", doc="Number of samples that must pass the minPruning threshold", optional = true)
     public int numPruningSamples = 1;
 
-    /**
-     * As of version 3.3, this argument is no longer needed because dangling end recovery is now the default behavior. See GATK 3.3 release notes for more details.
-     */
-    @Deprecated
-    @Argument(fullName="recover-dangling-heads", doc="This argument is deprecated since version 3.3", optional = true)
-    public boolean DEPRECATED_RecoverDanglingHeads = false;
-
-    /**
-     * By default, the read threading assembler will attempt to recover dangling heads and tails. See the `minDanglingBranchLength` argument documentation for more details.
-     */
-    @Hidden
-    @Argument(fullName="do-not-recover-dangling-branches", doc="Disable dangling head and tail recovery", optional = true)
-    public boolean doNotRecoverDanglingBranches = false;
-
     /**
      * When constructing the assembly graph we are often left with "dangling" branches.  The assembly engine attempts to rescue these branches
      * by merging them back into the main graph.  This argument describes the minimum length of a dangling branch needed for the engine to
@@ -71,13 +58,7 @@ public final class ReadThreadingAssemblerArgumentCollection implements Serializa
     @Argument(fullName="min-dangling-branch-length", doc="Minimum length of a dangling branch to attempt recovery", optional = true)
     public int minDanglingBranchLength = 4;
 
-    /**
-     * This argument is specifically intended for 1000G consensus analysis mode. Setting this flag will inject all
-     * provided alleles to the assembly graph but will not forcibly genotype all of them.
-     */
-    @Advanced
-    @Argument(fullName="consensus", doc="1000G consensus mode", optional = true)
-    public boolean consensusMode = false;
+
 
     /**
      * The assembly graph can be quite complex, and could imply a very large number of possible haplotypes.  Each haplotype
@@ -91,13 +72,6 @@ public final class ReadThreadingAssemblerArgumentCollection implements Serializa
     @Argument(fullName="max-num-haplotypes-in-population", doc="Maximum number of haplotypes to consider for your population", optional = true)
     public int maxNumHaplotypesInPopulation = 128;
 
-    /**
-     * Enabling this argument may cause fundamental problems with the assembly graph itself.
-     */
-    @Hidden
-    @Argument(fullName="error-correct-kmers", doc = "Use an exploratory algorithm to error correct the kmers used during assembly", optional = true)
-    public boolean errorCorrectKmers = false;
-
     /**
      * Paths with fewer supporting kmers than the specified threshold will be pruned from the graph.
      *
@@ -111,15 +85,6 @@ public final class ReadThreadingAssemblerArgumentCollection implements Serializa
     @Argument(fullName="min-pruning", doc = "Minimum support to not prune paths in the graph", optional = true)
     public int minPruneFactor = 2;
 
-    /**
-     * A single edge multiplicity cutoff for pruning doesn't work in samples with variable depths, for example exomes
-     * and RNA.  This parameter activates a probabilistic algorithm for pruning the assembly graph that considers the
-     * likelihood that each chain in the graph comes from real variation.
-     */
-    @Advanced
-    @Argument(fullName="adaptive-pruning", doc = "Use an adaptive algorithm for pruning paths in the graph", optional = true)
-    public boolean useAdaptivePruning = false;
-
     /**
      * Initial base error rate guess for the probabilistic adaptive pruning model.  Results are not very sensitive to this
      * parameter because it is only a starting point from which the algorithm discovers the true error rate.
@@ -168,4 +133,8 @@ public final class ReadThreadingAssemblerArgumentCollection implements Serializa
     @Hidden
     @Argument(fullName="min-observations-for-kmer-to-be-solid", doc = "A k-mer must be seen at least these times for it considered to be solid", optional = true)
     public int minObservationsForKmerToBeSolid = 20;
+
+    public abstract ReadThreadingAssembler makeReadThreadingAssembler();
+
+    public boolean consensusMode() { return false; }
 }

src/main/java/org/broadinstitute/hellbender/tools/walkers/haplotypecaller/readthreading/ReadThreadingAssembler.java

@@ -62,8 +62,6 @@ public final class ReadThreadingAssembler {
     private int pruneFactor;
     private final ChainPruner<MultiDeBruijnVertex, MultiSampleEdge> chainPruner;
 
-    protected boolean errorCorrectKmers = false;
-
     private File debugGraphOutputPath = null;  //Where to write debug graphs, if unset it defaults to the current working dir
     private File graphOutputPath = null;
 
@@ -572,18 +570,6 @@ private void printGraphs(final List<SeqGraph> graphs) {
     //
     // -----------------------------------------------------------------------------------------------
 
-    public int getPruneFactor() {
-        return pruneFactor;
-    }
-
-    public boolean shouldErrorCorrectKmers() {
-        return errorCorrectKmers;
-    }
-
-    public void setErrorCorrectKmers(boolean errorCorrectKmers) {
-        this.errorCorrectKmers = errorCorrectKmers;
-    }
-
     public void setGraphWriter(File graphOutputPath) {
         this.graphOutputPath = graphOutputPath;
     }
@@ -610,10 +596,6 @@ public boolean isDebugGraphTransformations() {
 
     public boolean isRecoverDanglingBranches() { return recoverDanglingBranches; }
 
-    public void setPruneFactor(final int pruneFactor) {
-        this.pruneFactor = pruneFactor;
-    }
-
     public void setDebugGraphTransformations(final boolean debugGraphTransformations) {
         this.debugGraphTransformations = debugGraphTransformations;
     }

src/main/java/org/broadinstitute/hellbender/tools/walkers/mutect/M2ArgumentCollection.java

@@ -49,6 +49,9 @@ public class M2ArgumentCollection extends AssemblyBasedCallerArgumentCollection
     public static final double DEFAULT_INITIAL_LOD = 2.0;
     public static final double DEFAULT_MITO_INITIAL_LOD = 0;
 
+    @Override
+    protected boolean useMutectAssemblerArgumentCollection() { return true; }
+
     //TODO: HACK ALERT HACK ALERT HACK ALERT
     //TODO: GATK4 does not yet have a way to tag inputs, eg -I:tumor tumor.bam -I:normal normal.bam,
     //TODO: so for now we require the user to specify bams *both* as inputs, with -I tumor.bam -I normal.bam