[ADAM-1469] Don't filter on whether reads have mismatches during real…

…ignment Resolves #1469. Provides approximately a 1% bump in both INDEL calling precision and recall.
bigdatagenomics · Apr 3, 2017 · 8df3aa2 · 8df3aa2
1 parent 15aa5cd
commit 8df3aa2
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 6 deletions.
diff --git a/adam-core/src/main/scala/org/bdgenomics/adam/rdd/read/realignment/RealignIndels.scala b/adam-core/src/main/scala/org/bdgenomics/adam/rdd/read/realignment/RealignIndels.scala
@@ -289,7 +289,6 @@ private[read] class RealignIndels(
         val (targetIdx, _) = target.get
         val startTime = System.nanoTime()
         // bootstrap realigned read set with the reads that need to be realigned
-        val (realignedReads, readsToRealign) = reads.partition(r => r.mdTag.exists(!_.hasMismatches))
 
         // get reference from reads
         val refStart = reads.map(_.getStart).min
@@ -299,7 +298,7 @@ private[read] class RealignIndels(
 
         // preprocess reads and get consensus
         val readsToClean = consensusModel.preprocessReadsForRealignment(
-          readsToRealign,
+          reads,
           reference,
           refRegion
         ).zipWithIndex
@@ -318,7 +317,7 @@ private[read] class RealignIndels(
           observedConsensus
         }
 
-        val finalReads = if (readsToClean.size > 0 && consensus.size > 0) {
+        val finalReads = if (reads.size > 0 && consensus.size > 0) {
 
           // do not check realigned reads - they must match
           val mismatchQualities = ComputingOriginalScores.time {
@@ -443,7 +442,7 @@ private[read] class RealignIndels(
               log.info("On " + refRegion + ", realigned " + realignedReadCount + " reads to " +
                 bestConsensus + " due to LOD improvement of " + lodImprovement)
 
-              cleanedReads ++ realignedReads
+              cleanedReads
             }
           } else {
             log.info("On " + refRegion + ", skipping realignment due to insufficient LOD improvement (" +

diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/rdd/read/realignment/RealignIndelsSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/rdd/read/realignment/RealignIndelsSuite.scala
@@ -307,7 +307,7 @@ class RealignIndelsSuite extends ADAMFunSuite {
       .collect()
 
     val movedReads = result.filter(pair => pair._1 != pair._2)
-    assert(movedReads.size === 22)
+    assert(movedReads.size === 41)
     val read = movedReads.map(_._2)
       .filter(_.getReadName === "H06HDADXX130110:1:1114:19044:27806")
       .head
@@ -552,7 +552,7 @@ class RealignIndelsSuite extends ADAMFunSuite {
     val realignedReads = rdd.realignIndels(lodThreshold = 0.0)
       .rdd
       .collect
-    assert(realignedReads.count(_.getMapq >= 50) === 6)
+    assert(realignedReads.count(_.getMapq >= 50) === 7)
     val realignedExtRead = realignedReads.filter(_.getMapq == 50).head
     assert(realignedExtRead.getStart === 8L)
     assert(realignedExtRead.getEnd === 14L)