BAMScheduler: detect contigs from the interval list that are not in the merged BAM header's sequence dictionary

This is a quick-and-dirty patch for the null pointer error Mauricio reported earlier.

Later on we might want to address in a more general way the fact that we validate user intervals
against the reference but not against the merged BAM header produced by the engine at runtime.

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

@@ -34,6 +34,8 @@ import net.sf.samtools.SAMSequenceRecord;
 import org.broadinstitute.sting.utils.GenomeLoc;
 import org.broadinstitute.sting.utils.GenomeLocParser;
 import org.broadinstitute.sting.utils.GenomeLocSortedSet;
+import org.broadinstitute.sting.utils.exceptions.UserException;
+import org.broadinstitute.sting.utils.sam.ReadUtils;
 
 import java.util.*;
 
@@ -245,7 +247,14 @@ public class BAMScheduler implements Iterator<FilePointer> {
         // This will ensure that if the two sets of contigs don't quite match (b36 male vs female ref, hg19 Epstein-Barr), then
         // we'll be using the correct contig index for the BAMs.
         // TODO: Warning: assumes all BAMs use the same sequence dictionary!  Get around this with contig aliasing.
-        final int currentContigIndex = dataSource.getHeader().getSequence(currentLocus.getContig()).getSequenceIndex();
+        SAMSequenceRecord currentContigSequenceRecord = dataSource.getHeader().getSequence(currentLocus.getContig());
+        if ( currentContigSequenceRecord == null ) {
+            throw new UserException(String.format("Contig %s not present in sequence dictionary for merged BAM header: %s",
+                                                  currentLocus.getContig(),
+                                                  ReadUtils.prettyPrintSequenceRecords(dataSource.getHeader().getSequenceDictionary())));
+        }
+
+        final int currentContigIndex = currentContigSequenceRecord.getSequenceIndex();
 
         // Stale reference sequence or first invocation.  (Re)create the binTreeIterator.
         if(lastReferenceSequenceLoaded == null || lastReferenceSequenceLoaded != currentContigIndex) {

public/java/src/org/broadinstitute/sting/utils/exceptions/UserException.java

@@ -30,6 +30,7 @@ import net.sf.samtools.SAMSequenceDictionary;
 import net.sf.samtools.SAMSequenceRecord;
 import org.broadinstitute.sting.utils.GenomeLoc;
 import org.broadinstitute.sting.utils.help.DocumentedGATKFeature;
+import org.broadinstitute.sting.utils.sam.ReadUtils;
 import org.broadinstitute.sting.utils.variantcontext.VariantContext;
 
 import java.io.File;
@@ -273,7 +274,7 @@ public class UserException extends ReviewedStingException {
     public static class IncompatibleSequenceDictionaries extends UserException {
         public IncompatibleSequenceDictionaries(String message, String name1, SAMSequenceDictionary dict1, String name2, SAMSequenceDictionary dict2) {
             super(String.format("Input files %s and %s have incompatible contigs: %s.\n  %s contigs = %s\n  %s contigs = %s",
-                    name1, name2, message, name1, prettyPrintSequenceRecords(dict1), name2, prettyPrintSequenceRecords(dict2)));
+                    name1, name2, message, name1, ReadUtils.prettyPrintSequenceRecords(dict1), name2, ReadUtils.prettyPrintSequenceRecords(dict2)));
         }
     }
 
@@ -284,17 +285,11 @@ public class UserException extends ReviewedStingException {
                     + "\nThis is because all distributed GATK resources are sorted in karyotypic order, and your processing will fail when you need to use these files."
                     + "\nYou can use the ReorderSam utility to fix this problem: http://www.broadinstitute.org/gsa/wiki/index.php/ReorderSam"
                     + "\n  %s contigs = %s",
-                    name, name, prettyPrintSequenceRecords(dict)));
+                    name, name, ReadUtils.prettyPrintSequenceRecords(dict)));
         }
     }
 
-    private static String prettyPrintSequenceRecords(SAMSequenceDictionary sequenceDictionary) {
-        String[] sequenceRecordNames = new String[sequenceDictionary.size()];
-        int sequenceRecordIndex = 0;
-        for (SAMSequenceRecord sequenceRecord : sequenceDictionary.getSequences())
-            sequenceRecordNames[sequenceRecordIndex++] = sequenceRecord.getSequenceName();
-        return Arrays.deepToString(sequenceRecordNames);
-    }
+
 
     public static class MissingWalker extends UserException {
         public MissingWalker(String walkerName, String message) {

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

@@ -648,4 +648,12 @@ public class ReadUtils {
         }
         return new Pair<HashMap<Integer, HashSet<GATKSAMRecord>>, HashMap<GATKSAMRecord, Boolean[]>>(locusToReadMap, readToLocusMap);
     }
+
+    public static String prettyPrintSequenceRecords ( SAMSequenceDictionary sequenceDictionary ) {
+        String[] sequenceRecordNames = new String[sequenceDictionary.size()];
+        int sequenceRecordIndex = 0;
+        for (SAMSequenceRecord sequenceRecord : sequenceDictionary.getSequences())
+            sequenceRecordNames[sequenceRecordIndex++] = sequenceRecord.getSequenceName();
+        return Arrays.deepToString(sequenceRecordNames);
+    }
 }