Merge pull request #289 from broadinstitute/md_fix_bq

Bugfix: defaultBaseQualities actually works now

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

@@ -187,6 +187,12 @@ public class GATKArgumentCollection {
     @Argument(fullName = "allow_potentially_misencoded_quality_scores", shortName="allowPotentiallyMisencodedQuals", doc="Do not fail when encountering base qualities that are too high and that seemingly indicate a problem with the base quality encoding of the BAM file", required = false)
     public boolean ALLOW_POTENTIALLY_MISENCODED_QUALS = false;
 
+    @Argument(fullName="useOriginalQualities", shortName = "OQ", doc = "If set, use the original base quality scores from the OQ tag when present instead of the standard scores", required=false)
+    public Boolean useOriginalBaseQualities = 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;
+
     // --------------------------------------------------------------------------------------------------------------
     //
     // performance log arguments
@@ -201,9 +207,6 @@ public class GATKArgumentCollection {
     @Argument(fullName = "performanceLog", shortName="PF", doc="If provided, a GATK runtime performance log will be written to this file", required = false)
     public File performanceLog = null;
 
-    @Argument(fullName="useOriginalQualities", shortName = "OQ", doc = "If set, use the original base quality scores from the OQ tag when present instead of the standard scores", required=false)
-    public Boolean useOriginalBaseQualities = false;
-
     // --------------------------------------------------------------------------------------------------------------
     //
     // BQSR arguments
@@ -267,9 +270,6 @@ public class GATKArgumentCollection {
     //
     // --------------------------------------------------------------------------------------------------------------
 
-    @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;
-
     @Argument(fullName = "validation_strictness", shortName = "S", doc = "How strict should we be with validation", required = false)
     public SAMFileReader.ValidationStringency strictnessLevel = SAMFileReader.ValidationStringency.SILENT;
 

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

@@ -630,6 +630,10 @@ public class SAMDataSource {
         // *     (otherwise we will process something that we may end up throwing away)                   * //
         // ************************************************************************************************ //
 
+        if (useOriginalBaseQualities || defaultBaseQualities >= 0)
+            // only wrap if we are replacing the original qualities or using a default base quality
+            wrappedIterator = new ReadFormattingIterator(wrappedIterator, useOriginalBaseQualities, defaultBaseQualities);
+
         // Filters:
         wrappedIterator = StingSAMIteratorAdapter.adapt(new CountingFilteringIterator(readMetrics,wrappedIterator,supplementalFilters));
 
@@ -654,10 +658,6 @@ public class SAMDataSource {
         if (!noValidationOfReadOrder && enableVerification)
             wrappedIterator = new VerifyingSamIterator(wrappedIterator);
 
-        if (useOriginalBaseQualities || defaultBaseQualities >= 0)
-            // only wrap if we are replacing the original qualities or using a default base quality
-            wrappedIterator = new ReadFormattingIterator(wrappedIterator, useOriginalBaseQualities, defaultBaseQualities);
-
         // set up read transformers
         for ( final ReadTransformer readTransformer : readTransformers ) {
             if ( readTransformer.enabled() && readTransformer.getApplicationTime() == ReadTransformer.ApplicationTime.ON_INPUT )

public/java/src/org/broadinstitute/sting/gatk/filters/MalformedReadFilter.java

@@ -234,7 +234,10 @@ public class MalformedReadFilter extends ReadFilter {
         else if (filterMismatchingBaseAndQuals)
             result = false;
         else
-            throw new UserException.MalformedBAM(read, String.format("BAM file has a read with mismatching number of bases and base qualities. Offender: %s [%d bases] [%d quals]", read.getReadName(), read.getReadLength(), read.getBaseQualities().length));
+            throw new UserException.MalformedBAM(read,
+                    String.format("BAM file has a read with mismatching number of bases and base qualities. Offender: %s [%d bases] [%d quals].%s",
+                            read.getReadName(), read.getReadLength(), read.getBaseQualities().length,
+                            read.getBaseQualities().length == 0 ? " You can use --defaultBaseQualities to assign a default base quality for all reads, but this can be dangerous in you don't know what you are doing." : ""));
 
         return result;
     }

public/java/test/org/broadinstitute/sting/gatk/EngineFeaturesIntegrationTest.java

@@ -227,4 +227,32 @@ public class EngineFeaturesIntegrationTest extends WalkerTest {
             nLines++;
         Assert.assertTrue(nLines > 0);
     }
+
+    // --------------------------------------------------------------------------------
+    //
+    // Test that defaultBaseQualities actually works
+    //
+    // --------------------------------------------------------------------------------
+
+    public WalkerTestSpec testDefaultBaseQualities(final Integer value, final String md5) {
+        return new WalkerTestSpec("-T PrintReads -R " + b37KGReference + " -I " + privateTestDir + "/baseQualitiesToFix.bam -o %s"
+                + (value != null ? " --defaultBaseQualities " + value : ""),
+                1, Arrays.asList(md5));
+    }
+
+    @Test()
+    public void testDefaultBaseQualities20() {
+        executeTest("testDefaultBaseQualities20", testDefaultBaseQualities(20, "7d254a9d0ec59c66ee3e137f56f4c78f"));
+    }
+
+    @Test()
+    public void testDefaultBaseQualities30() {
+        executeTest("testDefaultBaseQualities30", testDefaultBaseQualities(30, "0f50def6cbbbd8ccd4739e2b3998e503"));
+    }
+
+    @Test(expectedExceptions = Exception.class)
+    public void testDefaultBaseQualitiesNoneProvided() {
+        executeTest("testDefaultBaseQualitiesNoneProvided", testDefaultBaseQualities(null, ""));
+    }
+
 }