Added a --no_indel_quals argument that when used with -BQSR inhibits the writing of base insertion and base deletion quality tags.

public/java/src/org/broadinstitute/sting/gatk/GenomeAnalysisEngine.java

@@ -197,7 +197,7 @@ public class GenomeAnalysisEngine {
     private BaseRecalibration baseRecalibration = null;
     public BaseRecalibration getBaseRecalibration() { return baseRecalibration; }
     public boolean hasBaseRecalibration() { return baseRecalibration != null; }
-    public void setBaseRecalibration(File recalFile, int quantizationLevels) { baseRecalibration = new BaseRecalibration(recalFile, quantizationLevels); }
+    public void setBaseRecalibration(final File recalFile, final int quantizationLevels, final boolean noIndelQuals) { baseRecalibration = new BaseRecalibration(recalFile, quantizationLevels, noIndelQuals); }
 
     /**
      * Actually run the GATK with the specified walker.
@@ -227,7 +227,7 @@ public class GenomeAnalysisEngine {
 
         // if the use specified an input BQSR recalibration table then enable on the fly recalibration
         if (this.getArguments().BQSR_RECAL_FILE != null)
-            setBaseRecalibration(this.getArguments().BQSR_RECAL_FILE, this.getArguments().quantizationLevels);
+            setBaseRecalibration(this.getArguments().BQSR_RECAL_FILE, this.getArguments().quantizationLevels, this.getArguments().noIndelQuals);
 
         // Determine how the threads should be divided between CPU vs. IO.
         determineThreadAllocation();

public/java/src/org/broadinstitute/sting/gatk/arguments/GATKArgumentCollection.java

@@ -209,6 +209,12 @@ public class GATKArgumentCollection {
     @Argument(fullName="quantize_quals", shortName = "qq", doc = "Quantize quality scores to a given number of levels.", required=false)
     public int quantizationLevels = -1;
 
+    /**
+     * Turns off printing of the base insertion and base deletion tags when using the -BQSR argument.  Only the base substitution qualities will be produced.
+     */
+    @Argument(fullName="no_indel_quals", shortName = "NIQ", doc = "If true, inhibits printing of base insertion and base deletion tags.", required=false)
+    public boolean noIndelQuals = false;
+
     @Argument(fullName="defaultBaseQualities", shortName = "DBQ", doc = "If reads are missing some or all base quality scores, this value will be used for all base quality scores", required=false)
     public byte defaultBaseQualities = -1;
 

public/java/src/org/broadinstitute/sting/gatk/datasources/reads/SAMDataSource.java

@@ -664,8 +664,12 @@ public class SAMDataSource {
                                                         IndexedFastaSequenceFile refReader,
                                                         BaseRecalibration bqsrApplier,
                                                         byte defaultBaseQualities) {
-        // **** NOTE: ALL FILTERING SHOULD BE DONE BEFORE ANY ITERATORS THAT MODIFY THE READS! ****
+
-        //         (otherwise we will process something that we may end up throwing away)
+        // *********************************************************************************** //
+        // *  NOTE: ALL FILTERING SHOULD BE DONE BEFORE ANY ITERATORS THAT MODIFY THE READS! * //
+        // *     (otherwise we will process something that we may end up throwing away)      * //
+        // *********************************************************************************** //
+
         if (downsamplingFraction != null)
             wrappedIterator = new DownsampleIterator(wrappedIterator, downsamplingFraction);
 

public/java/src/org/broadinstitute/sting/utils/recalibration/BaseRecalibration.java

@@ -49,6 +49,8 @@ public class BaseRecalibration {
     private final RecalibrationTables recalibrationTables;
     private final Covariate[] requestedCovariates;                                                                      // list of all covariates to be used in this calculation
 
+    private final boolean noIndelQuals;
+
     private static final NestedHashMap[] qualityScoreByFullCovariateKey = new NestedHashMap[EventType.values().length]; // Caches the result of performSequentialQualityCalculation(..) for all sets of covariate values.
     static {
         for (int i = 0; i < EventType.values().length; i++)
@@ -58,10 +60,11 @@ public class BaseRecalibration {
     /**
      * Constructor using a GATK Report file
      * 
-     * @param RECAL_FILE a GATK Report file containing the recalibration information
+     * @param RECAL_FILE         a GATK Report file containing the recalibration information
      * @param quantizationLevels number of bins to quantize the quality scores
+     * @param noIndelQuals       if true, do not emit base indel qualities
      */
-    public BaseRecalibration(final File RECAL_FILE, int quantizationLevels) {
+    public BaseRecalibration(final File RECAL_FILE, final int quantizationLevels, final boolean noIndelQuals) {
         RecalibrationReport recalibrationReport = new RecalibrationReport(RECAL_FILE);
 
         recalibrationTables = recalibrationReport.getRecalibrationTables();
@@ -73,6 +76,7 @@ public class BaseRecalibration {
             quantizationInfo.quantizeQualityScores(quantizationLevels);
 
         readCovariates = new ReadCovariates(MAXIMUM_RECALIBRATED_READ_LENGTH, requestedCovariates.length);
+        this.noIndelQuals = noIndelQuals;
     }
 
     /**
@@ -87,6 +91,7 @@ public class BaseRecalibration {
         this.recalibrationTables = recalibrationTables;
         this.requestedCovariates = requestedCovariates;
         readCovariates = new ReadCovariates(MAXIMUM_RECALIBRATED_READ_LENGTH, requestedCovariates.length);
+        noIndelQuals = false;
     }
 
     /**
@@ -99,6 +104,11 @@ public class BaseRecalibration {
     public void recalibrateRead(final GATKSAMRecord read) {
         RecalDataManager.computeCovariates(read, requestedCovariates, readCovariates);                                  // compute all covariates for the read
         for (final EventType errorModel : EventType.values()) {                                                         // recalibrate all three quality strings
+            if (noIndelQuals && errorModel != EventType.BASE_SUBSTITUTION) {
+                read.setBaseQualities(null, errorModel);
+                continue;
+            }
+
             final byte[] quals = read.getBaseQualities(errorModel);
             final int[][] fullReadKeySet = readCovariates.getKeySet(errorModel);                                        // get the keyset for this base using the error model
 

public/java/src/org/broadinstitute/sting/utils/sam/GATKSAMRecord.java

@@ -173,10 +173,10 @@ public class GATKSAMRecord extends BAMRecord {
                 setBaseQualities(quals);
                 break;
             case BASE_INSERTION:
-                setAttribute( GATKSAMRecord.BQSR_BASE_INSERTION_QUALITIES, SAMUtils.phredToFastq(quals) );
+                setAttribute( GATKSAMRecord.BQSR_BASE_INSERTION_QUALITIES, quals == null ? null : SAMUtils.phredToFastq(quals) );
                 break;
             case BASE_DELETION:
-                setAttribute( GATKSAMRecord.BQSR_BASE_DELETION_QUALITIES, SAMUtils.phredToFastq(quals) );
+                setAttribute( GATKSAMRecord.BQSR_BASE_DELETION_QUALITIES, quals == null ? null : SAMUtils.phredToFastq(quals) );
                 break;
             default:
                 throw new ReviewedStingException("Unrecognized Base Recalibration type: " + errorModel );