Not nearly as badass as it looks. The problem I mentioned yesterday with "bleeding in" of samples comes from VCFUtils and SampleUtils looking for all VCF-class RODs in the tracker, and stealing the name from them. I have introduced a new HapmapVCF - type rod for use

when you want to protect your VCF header from being infected by the samples in a bound hapmap VCF. Changes are as follows:

VCFRecord - minor change to adapt isNovel() to the case where the dbsnp ID field is empty, but the info field has DB=1

HapmapVCFRod - introduced for the reason at the top

RODRecordIterator - was: catch ( Exception e ) { throw new StingException("long ass message") }
                 is now: catch ( Exception e ) { throw new StingException("long ass message",e) }
                    to permit full stack ejaculation.

RodVCF - Now with more brackets!

ReferenceOrderedData - registering HapmapVCF as a bindable string

VariantAnnotator - There's an extra space on a line. And some new brackets.



git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2733 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/refdata/HapmapVCFROD.java

@@ -0,0 +1,170 @@
+package org.broadinstitute.sting.gatk.refdata;
+
+import org.broadinstitute.sting.utils.GenomeLoc;
+import org.broadinstitute.sting.utils.StingException;
+import org.broadinstitute.sting.utils.genotype.DiploidGenotype;
+import org.broadinstitute.sting.utils.genotype.Genotype;
+import org.broadinstitute.sting.utils.genotype.VariantBackedByGenotype;
+import org.broadinstitute.sting.utils.genotype.Variation;
+import org.broadinstitute.sting.utils.genotype.vcf.VCFHeader;
+import org.broadinstitute.sting.utils.genotype.vcf.VCFReader;
+import org.broadinstitute.sting.utils.genotype.vcf.VCFRecord;
+
+import java.io.File;
+import java.io.FileNotFoundException;
+import java.io.IOException;
+import java.util.Iterator;
+import java.util.List;
+
+/**
+ * Created by IntelliJ IDEA.
+ * User: chartl
+ * Date: Jan 29, 2010
+ * Time: 8:57:54 AM
+ * To change this template use File | Settings | File Templates.
+ */
+public class HapmapVCFROD extends BasicReferenceOrderedDatum implements VariationRod, VariantBackedByGenotype, Iterator<HapmapVCFROD> {
+// This is a (hopefully temporary) wrapper class for certain VCF files that we want to protect from
+// utilities that grab genotypes or sample names across all VCF files
+
+    private RodVCF rod;
+
+    public VCFReader getReader() {
+        return rod.getReader();
+    }
+
+    public VCFRecord getRecord() {
+        return rod.getRecord();
+    }
+
+    public HapmapVCFROD(String name) {
+        super(name);
+    }
+
+    public HapmapVCFROD(String name, VCFRecord currentRecord, VCFReader reader) {
+        super(name);
+        rod = new RodVCF(name,currentRecord,reader);
+    }
+
+    public HapmapVCFROD(String name, RodVCF rod) {
+        super(name);
+        this.rod = rod;
+    }
+
+    public Object initialize(final File source) throws FileNotFoundException {
+        rod = new RodVCF(name);
+        rod.initialize(source);
+        return rod.getHeader();
+    }
+
+    public boolean parseLine(Object obj, String[] args) {
+        try {
+            return rod.parseLine(obj,args);
+        } catch (Exception e) {
+            throw new UnsupportedOperationException("Parse line not supported",e);
+        }
+    }
+
+    public double getNegLog10PError() {
+        return rod.getNegLog10PError();
+    }
+
+    public List<Genotype> getGenotypes() {
+        return rod.getGenotypes();
+    }
+
+    public String getReference() {
+        return rod.getReference();
+    }
+
+    public String toString() {
+        return rod.toString();
+    }
+
+    public List<String> getAlternateAlleleList() {
+        return rod.getAlternateAlleleList();
+    }
+
+    public boolean isDeletion() {
+        return rod.isDeletion();
+    }
+
+    public GenomeLoc getLocation() {
+        return rod.getLocation();
+    }
+
+    public boolean isBiallelic() {
+        return rod.isBiallelic();
+    }
+
+    public boolean isIndel() {
+        return rod.isIndel();
+    }
+
+    public Variation.VARIANT_TYPE getType() {
+        return rod.getType();
+    }
+
+    public boolean isSNP() {
+        return rod.isSNP();
+    }
+
+    public boolean isReference() {
+        return rod.isReference();
+    }
+
+    public double getNonRefAlleleFrequency() {
+        return rod.getNonRefAlleleFrequency();
+    }
+
+    public char getAlternativeBaseForSNP() {
+        return rod.getAlternativeBaseForSNP();
+    }
+
+    public boolean isInsertion() {
+        return rod.isInsertion();
+    }
+
+    public List<String> getAlleleList() {
+        return rod.getAlleleList();
+    }
+
+    public Genotype getCalledGenotype() {
+        return rod.getCalledGenotype();
+    }
+
+    public char getReferenceForSNP() {
+        return rod.getReferenceForSNP();
+    }
+
+    public boolean hasGenotype(DiploidGenotype g) {
+        return rod.hasGenotype(g);
+    }
+
+    public VCFHeader getHeader() {
+        return rod.getHeader();
+    }
+
+    public boolean hasNext() {
+        return rod.hasNext();
+    }
+
+    public HapmapVCFROD next() {
+        return new HapmapVCFROD(name,rod.next());
+    }
+
+    public void remove() {
+        rod.remove();
+    }
+
+    public static HapmapVCFROD createIterator(String name, File file) {
+        RodVCF vcf = new RodVCF(name);
+        try {
+            vcf.initialize(file);
+        } catch (FileNotFoundException e) {
+            throw new StingException("Unable to find file " + file);
+        }
+        return new HapmapVCFROD(name,vcf);
+    }
+
+}

java/src/org/broadinstitute/sting/gatk/refdata/RODRecordIterator.java

@@ -173,7 +173,7 @@ public class RODRecordIterator<ROD extends ReferenceOrderedDatum> implements Ite
              catch ( Exception e ) {
                      throw new StingException("Failed to parse ROD data ("+type.getName()+") from file "+ file + " at line #"+linenum+
                                              "\nOffending line: "+line+
-                                             "\nReason ("+e.getClass().getName()+"): "+e.getMessage());
+                                             "\nReason ("+e.getClass().getName()+"): "+e.getMessage(),e);
              }
         }
 

java/src/org/broadinstitute/sting/gatk/refdata/ReferenceOrderedData.java

@@ -77,6 +77,7 @@ public class ReferenceOrderedData<ROD extends ReferenceOrderedDatum> implements
         addModule("GLF", RodGLF.class);
         addModule("VCF", RodVCF.class);
         addModule("PicardDbSNP", rodPicardDbSNP.class);
+        addModule("HapmapVCF",HapmapVCFROD.class);
 
     }
 

java/src/org/broadinstitute/sting/gatk/refdata/RodVCF.java

@@ -56,8 +56,9 @@ public class RodVCF extends BasicReferenceOrderedDatum implements VariationRod,
     }
 
     public Object initialize(final File source) throws FileNotFoundException {
-        if ( mReader == null )
+        if ( mReader == null ) {
             mReader = new VCFReader(source);
+        }
         return mReader.getHeader();
     }
 

java/src/org/broadinstitute/sting/gatk/walkers/annotator/VariantAnnotator.java

@@ -95,14 +95,15 @@ public class VariantAnnotator extends LocusWalker<Integer, Integer> {
         SampleUtils.getUniquifiedSamplesFromRods(getToolkit(), samples, new HashMap<Pair<String, String>, String>());
 
         // add the non-VCF sample from the command-line, if applicable
-        if ( sampleName != null ) {
+        if ( sampleName != null  ) {
             nonVCFsampleName.put(sampleName.toUpperCase(), "variant");
             samples.add(sampleName.toUpperCase());
         }
 
         // if there are no valid samples, warn the user
-        if ( samples.size() == 0 )
+        if ( samples.size() == 0 ) {
             logger.warn("There are no samples input at all; use the --sampleName argument to specify one if desired.");
+        }
 
         determineAllAnnotations();
 

java/src/org/broadinstitute/sting/utils/genotype/vcf/VCFRecord.java

@@ -325,7 +325,11 @@ public class VCFRecord implements Variation, VariantBackedByGenotype {
     }
 
     public boolean isNovel() {
-        return mID == null && ! isInHapmap();
+        return ( ! isInDBSNP() ) && ( ! isInHapmap() );
+    }
+
+    public boolean isInDBSNP() {
+        return ( mID != null || ( mInfoFields.get(DBSNP_KEY) != null && mInfoFields.get(DBSNP_KEY).equals("1") ) );
     }
 
     public boolean isInHapmap() {