Let the DuplexConsensusCaller support UMI pairs with one absent

fulcrumgenomics · Dec 30, 2023 · 9d5a422 · 9d5a422
1 parent d510954
commit 9d5a422
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 2 deletions.
diff --git a/src/main/scala/com/fulcrumgenomics/umi/DuplexConsensusCaller.scala b/src/main/scala/com/fulcrumgenomics/umi/DuplexConsensusCaller.scala
@@ -25,13 +25,13 @@
 package com.fulcrumgenomics.umi
 
 import java.lang.Math.min
-
 import com.fulcrumgenomics.FgBioDef._
 import com.fulcrumgenomics.bam.api.SamRecord
 import com.fulcrumgenomics.commons.util.LazyLogging
 import com.fulcrumgenomics.umi.DuplexConsensusCaller._
 import com.fulcrumgenomics.umi.UmiConsensusCaller.ReadType.{ReadType, _}
 import com.fulcrumgenomics.umi.UmiConsensusCaller.{SimpleRead, SourceRead}
+import com.fulcrumgenomics.umi.Umis.UmiSeparatorPattern
 import com.fulcrumgenomics.util.NumericTypes.PhredScore
 
 /**
@@ -324,7 +324,7 @@ class DuplexConsensusCaller(override val readNamePrefix: String,
     // UMI bases are present, `None` otherwise.
     reads.flatMap(_.sam).flatMap { rec =>
       rec.get[String](ConsensusTags.UmiBases).map { umi =>
-        if (rec.firstOfPair == firstOfPair) umi else umi.split('-').reverse.mkString("-")
+        if (rec.firstOfPair == firstOfPair) umi else UmiSeparatorPattern.split(umi, -1).reverse.mkString("-")
       }
     }
   }

diff --git a/src/test/scala/com/fulcrumgenomics/umi/DuplexConsensusCallerTest.scala b/src/test/scala/com/fulcrumgenomics/umi/DuplexConsensusCallerTest.scala
@@ -88,6 +88,29 @@ class DuplexConsensusCallerTest extends UnitSpec with OptionValues {
     r2.quals(0) > 30 shouldBe true
   }
 
+  Seq(
+    Seq("right", "ACT-", "-ACT"),
+    Seq("left", "-ACT", "ACT-"),
+  ).foreach { case Seq(orientation, leftUmi, rightUmi) =>
+    it should s"create a simple double stranded consensus for a pair of A and a pair of B reads in which the $orientation UMI is absent" in {
+      val builder = new SamBuilder(readLength = 10, baseQuality = 20)
+      builder.addPair(name = "q1", start1 = 100, start2 = 200, strand1 = Plus, strand2 = Minus, bases1 = "AAAAAAAAAA", bases2 = "CCCCCCCCCC", attrs = Map(MI -> "foo/A", RX -> leftUmi))
+      builder.addPair(name = "q2", start1 = 200, start2 = 100, strand1 = Minus, strand2 = Plus, bases1 = "CCCCCCCCCC", bases2 = "AAAAAAAAAA", attrs = Map(MI -> "foo/B", RX -> rightUmi))
+
+      val recs = c.consensusReadsFromSamRecords(builder.toSeq)
+      recs should have size 2
+      val r1 = recs.find(_.firstOfPair).getOrElse(fail("No first of pair."))
+      val r2 = recs.find(_.secondOfPair).getOrElse(fail("No second of pair."))
+      r1.basesString shouldBe "AAAAAAAAAA"
+      r2.basesString shouldBe "GGGGGGGGGG" // after un-revcomping
+      r1.quals(0) > 30 shouldBe true
+      r2.quals(0) > 30 shouldBe true
+      // Both R1 and R2 inherit the RX pattern of the "A" family which is the left-oriented read pair in this test.
+      r1.get[String](RX) shouldBe Some(leftUmi)
+      r2.get[String](RX) shouldBe Some(leftUmi)
+    }
+  }
+
   Seq(true).foreach { allowSingleStrand =>
     val extraName: String = if (allowSingleStrand) "" else "not "
     Seq("A", "B").foreach { duplexStrand =>