Following up Mark's recent commit: hooking up the --maxPositionalMoveAllowed argument into the indel realigner and through to the SAM writer. We now ensure that no read is realigned more than N bases (200 by default, which is nowhere close to realistically possible). If anyone ever sees a warning message about this with the default value then please let me know because I need to see it for myself.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5331 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/io/StingSAMFileWriter.java

@@ -30,5 +30,6 @@ public interface StingSAMFileWriter extends SAMFileWriter {
     public void setMaxRecordsInRam(int maxRecordsInRam);
 
     public void setMaxInsertSizeForMovingReadPairs(int maxInsertSizeForMovingReadPairs);
+    public void setMaxPositionalMoveAllowed(int maxPositionalMoveAllowed);
     public void setUseConstrainedFileWriter(boolean useConstrainedFileWriter);
 }

java/src/org/broadinstitute/sting/gatk/io/storage/SAMFileWriterStorage.java

@@ -82,7 +82,7 @@ public class SAMFileWriterStorage implements SAMFileWriter, Storage<SAMFileWrite
             throw new UserException("Unable to write to SAM file; neither a target file nor a stream has been specified");
 
         if ( stub.useConstrainedFileWriter() ) {
-            this.writer = new ConstrainedMateFixingSAMFileWriter(writer, stub.getMaxInsertSizeForMovingReadPairs());
+            this.writer = new ConstrainedMateFixingSAMFileWriter(writer, stub.getMaxInsertSizeForMovingReadPairs(), stub.getMaxPositionalMoveAllowed());
         }
     }
 

java/src/org/broadinstitute/sting/gatk/io/stubs/SAMFileWriterStub.java

@@ -101,6 +101,14 @@ public class SAMFileWriterStub implements Stub<SAMFileWriter>, StingSAMFileWrite
      */
     private boolean writeStarted = false;
 
+    /**
+     * Arguments used for the constrained mate pair fixing file writer
+     */
+    private boolean useConstrainedFileWriter = false;
+    private int maxInsertSizeForMovingReadPairs = -1;
+    private int maxPositionalMoveAllowed = -1;
+
+
     /**
      * HMM for BAQ, if needed
      */
@@ -258,12 +266,6 @@ public class SAMFileWriterStub implements Stub<SAMFileWriter>, StingSAMFileWrite
         outputTracker.getStorage(this).close();    
     }
 
-    //
-    // Experimental arguments for the constrained SAMFileWriter
-    //
-    private boolean useConstrainedFileWriter = false;
-    private int maxInsertSizeForMovingReadPairs = -1;
-
     public int getMaxInsertSizeForMovingReadPairs() {
         return maxInsertSizeForMovingReadPairs;
     }
@@ -272,6 +274,14 @@ public class SAMFileWriterStub implements Stub<SAMFileWriter>, StingSAMFileWrite
         this.maxInsertSizeForMovingReadPairs = maxInsertSizeForMovingReadPairs;
     }
 
+    public int getMaxPositionalMoveAllowed() {
+        return maxPositionalMoveAllowed;
+    }
+
+    public void setMaxPositionalMoveAllowed(int maxPositionalMoveAllowed) {
+        this.maxPositionalMoveAllowed = maxPositionalMoveAllowed;
+    }
+
     public boolean useConstrainedFileWriter() {
         return useConstrainedFileWriter;
     }

java/src/org/broadinstitute/sting/gatk/walkers/indels/IndelRealigner.java

@@ -101,6 +101,9 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
     @Argument(fullName="maxIsizeForMovement", shortName="maxIsize", doc="maximum insert size of read pairs that we attempt to realign", required=false)
     protected int MAX_ISIZE_FOR_MOVEMENT = 3000;
 
+    @Argument(fullName="maxPositionalMoveAllowed", shortName="maxPosMove", doc="maximum positional move in basepairs that a read can be adjusted during realignment", required=false)
+    protected int MAX_POS_MOVE_ALLOWED = 200;
+
     @Argument(fullName="maxConsensuses", shortName="maxConsensuses", doc="max alternate consensuses to try (necessary to improve performance in deep coverage)", required=false)
     protected int MAX_CONSENSUSES = 30;
 
@@ -378,6 +381,7 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
 
         writer.setUseConstrainedFileWriter(true);
         writer.setMaxInsertSizeForMovingReadPairs(MAX_ISIZE_FOR_MOVEMENT);
+        writer.setMaxPositionalMoveAllowed(MAX_POS_MOVE_ALLOWED);
     }
 
     private void emit(final SAMRecord read) {
@@ -406,7 +410,7 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
 
     private void emitReadLists() {
         // pre-merge lists with priority queue for constrained SAMFileWriter
-        logger.warn("EMIT currentInterval " + currentInterval);
+        //logger.warn("EMIT currentInterval " + currentInterval);
         readsNotToClean.addAll(readsToClean.getReads());
         emit(ReadUtils.coordinateSortReads(readsNotToClean));
         readsToClean.clear();
@@ -1430,6 +1434,10 @@ public class IndelRealigner extends ReadWalker<Integer, Integer> {
                 return false;
             if ( newStart == -1 )
                 newStart = read.getAlignmentStart();
+            else if ( Math.abs(newStart - read.getAlignmentStart()) > MAX_POS_MOVE_ALLOWED ) {
+                logger.warn(String.format("Attempting to realign read %s at %d more than %d bases to %d.", read.getReadName(), read.getAlignmentStart(), MAX_POS_MOVE_ALLOWED, newStart));
+                return false;
+            }
 
             // annotate the record with the original cigar (and optionally the alignment start)
             if ( !NO_ORIGINAL_ALIGNMENT_TAGS ) {

java/src/org/broadinstitute/sting/utils/sam/ConstrainedMateFixingSAMFileWriter.java

@@ -75,17 +75,16 @@ import java.util.*;
  * @version 0.1
  */
 public class ConstrainedMateFixingSAMFileWriter implements SAMFileWriter {
-    final protected static Logger logger = Logger.getLogger(ConstrainedMateFixingSAMFileWriter.class);
+    protected static final Logger logger = Logger.getLogger(ConstrainedMateFixingSAMFileWriter.class);
-    private final static boolean DEBUG = false;
+    private static final boolean DEBUG = false;
 
     /** How often do we check whether we want to emit reads? */
     private final static int EMIT_FREQUENCY = 1000;
 
     /**
      * How much could a single read move in position from its original position?
-     * todo -- this really should be a provided parameter
      */
-    private final static int MAX_POS_MOVE_ALLOWED = 200;
+    private int MAX_POS_MOVE_ALLOWED;
 
     /** how we order our SAM records */
     private final SAMRecordComparator comparer = new SAMRecordCoordinateComparator();
@@ -122,15 +121,19 @@ public class ConstrainedMateFixingSAMFileWriter implements SAMFileWriter {
     public ConstrainedMateFixingSAMFileWriter(final SAMFileHeader header,
                                               final File outputFile,
                                               final int compressionLevel,
-                                              final int maxInsertSizeForMovingReadPairs) {
+                                              final int maxInsertSizeForMovingReadPairs,
+                                              final int maxMoveAllowed) {
         this(new SAMFileWriterFactory().makeBAMWriter(header, true, outputFile, compressionLevel),
-                maxInsertSizeForMovingReadPairs);
+                maxInsertSizeForMovingReadPairs,
+                maxMoveAllowed);
     }
 
     public ConstrainedMateFixingSAMFileWriter(final SAMFileWriter finalDestination,
-                                              final int maxInsertSizeForMovingReadPairs) {
+                                              final int maxInsertSizeForMovingReadPairs,
+                                              final int maxmoveAllowed) {
         this.finalDestination = finalDestination;
         this.maxInsertSizeForMovingReadPairs = maxInsertSizeForMovingReadPairs;
+        this.MAX_POS_MOVE_ALLOWED = maxmoveAllowed;
 
         //timer.start();
         //lastProgressPrintTime = timer.currentTime();

java/test/org/broadinstitute/sting/utils/sam/ConstrainedMateFixingSAMFileWriterUnitTest.java

@@ -1,16 +1,12 @@
 package org.broadinstitute.sting.utils.sam;
 
-import com.sun.xml.internal.messaging.saaj.packaging.mime.util.OutputUtil;
 import net.sf.picard.reference.IndexedFastaSequenceFile;
 import net.sf.picard.sam.SamFileValidator;
 import net.sf.samtools.*;
 import org.broadinstitute.sting.BaseTest;
-import org.broadinstitute.sting.utils.GenomeLocParser;
-import org.broadinstitute.sting.utils.Utils;
 import org.broadinstitute.sting.utils.exceptions.UserException;
 import org.testng.Assert;
 import org.testng.annotations.BeforeMethod;
-import org.testng.annotations.BeforeTest;
 import org.testng.annotations.Test;
 
 import java.io.File;
@@ -21,9 +17,6 @@ import java.util.Arrays;
 import java.util.Collection;
 import java.util.List;
 
-import static org.testng.Assert.assertEquals;
-import static org.testng.Assert.assertTrue;
-
 
 /*
  * Copyright (c) 2009 The Broad Institute
@@ -86,7 +79,7 @@ public class ConstrainedMateFixingSAMFileWriterUnitTest extends BaseTest {
     }
 
     private ConstrainedMateFixingSAMFileWriter makeWriter(final int maxInsertSizeForMovingReadPairs) {
-        return new ConstrainedMateFixingSAMFileWriter(bamIn.getFileHeader(), OUTPUT_FILE, 5, maxInsertSizeForMovingReadPairs);
+        return new ConstrainedMateFixingSAMFileWriter(bamIn.getFileHeader(), OUTPUT_FILE, 5, maxInsertSizeForMovingReadPairs, 200);
     }
 
     private List<SAMRecord> readBAM(File file) {