[ADAM-1579] Add unit test coverage for BED12 format.

Resolves #1579. Found small bugs that would lead to last 6 fields not getting written if the thickStart field wasn't written. Also, identified a bug where the strand would be set to unknown if an independent strand was provided.
bigdatagenomics · Jul 4, 2017 · ee331a2 · ee331a2
1 parent 0306717
commit ee331a2
Show file tree

Hide file tree

Showing 5 changed files with 47 additions and 3 deletions.
diff --git a/adam-core/src/main/scala/org/bdgenomics/adam/rdd/feature/FeatureParser.scala b/adam-core/src/main/scala/org/bdgenomics/adam/rdd/feature/FeatureParser.scala
@@ -323,7 +323,7 @@ private[rdd] class BEDParser extends FeatureParser {
 
     if (hasColumn(3)) f.setName(fields(3))
     if (hasColumn(4)) f.setScore(fields(4).toDouble)
-    if (hasColumn(5)) Features.toStrand(fields(5)).foreach(f.setStrand(_))
+    if (fields.length > 5) Features.toStrand(fields(5)).foreach(f.setStrand(_))
 
     val attributes = new ArrayBuffer[(String, String)]()
     if (hasColumn(6)) attributes += ("thickStart" -> fields(6))

diff --git a/adam-core/src/main/scala/org/bdgenomics/adam/rdd/feature/FeatureRDD.scala b/adam-core/src/main/scala/org/bdgenomics/adam/rdd/feature/FeatureRDD.scala
@@ -210,7 +210,9 @@ object FeatureRDD {
     val score = Option(feature.getScore).getOrElse(".")
     val strand = Features.asString(feature.getStrand)
 
-    if (!feature.getAttributes.containsKey("thickStart")) {
+    if (!feature.getAttributes.containsKey("thickStart") &&
+      !feature.getAttributes.containsKey("itemRgb") &&
+      !feature.getAttributes.containsKey("blockCount")) {
       // write BED6 format
       List(chrom, start, end, name, score, strand).mkString("\t")
     } else {

diff --git a/adam-core/src/test/resources/small.1_12.bed b/adam-core/src/test/resources/small.1_12.bed
@@ -0,0 +1,4 @@
+1	143	26423	line1	0.0	.	150	26400	0,0,0	.	.	.
+1	14397230	26472788	line2	100.0	+	14397230	26472700	255,0,0	1	12075558	14397230
+1	169801934	169801939	line3	200.0	-	.	.	0,255,0	2	100,200	169801934,169801739
+1	240997788	240997796	line4 with a space	1000.0	?	.	.	0,0,255	.	.	.
diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/rdd/ADAMContextSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/rdd/ADAMContextSuite.scala
@@ -136,6 +136,13 @@ class ADAMContextSuite extends ADAMFunSuite {
     assert(features.count === 10)
   }
 
+  sparkTest("Can read a BED 12 file") {
+    // note: this .bed doesn't actually conform to the UCSC BED spec...sigh...
+    val path = testFile("small.1_12.bed")
+    val features: RDD[Feature] = sc.loadFeatures(path).rdd
+    assert(features.count === 4)
+  }
+
   sparkTest("Can read a .bed file without cache") {
     val path = testFile("gencode.v7.annotation.trunc10.bed")
     val features: RDD[Feature] = sc.loadFeatures(path, optStorageLevel = Some(StorageLevel.NONE)).rdd

diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/rdd/feature/FeatureRDDSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/rdd/feature/FeatureRDDSuite.scala
@@ -240,7 +240,7 @@ class FeatureRDDSuite extends ADAMFunSuite with TypeCheckedTripleEquals {
     assert(features.rdd.count === reloadedFeatures.rdd.count)
   }
 
-  sparkTest("round trip BED format") {
+  sparkTest("round trip BED6 format") {
     val inputPath = testFile("dvl1.200.bed")
     val expected = sc.loadBed(inputPath)
     val outputPath = tempLocation(".bed")
@@ -263,6 +263,37 @@ class FeatureRDDSuite extends ADAMFunSuite with TypeCheckedTripleEquals {
     })
   }
 
+  sparkTest("round trip BED12 format") {
+    val inputPath = testFile("small.1_12.bed")
+    val expected = sc.loadBed(inputPath)
+    val outputPath = tempLocation(".bed")
+    expected.saveAsBed(outputPath, asSingleFile = true)
+
+    val feature = expected.rdd.first
+    val bedCols = FeatureRDD.toBed(feature).split('\t')
+    assert(bedCols.size === 12)
+    assert(bedCols(0) === "1")
+    assert(bedCols(1) === "143")
+    assert(bedCols(2) === "26423")
+    assert(bedCols(3) === "line1")
+    assert(bedCols(4) === "0.0")
+    assert(bedCols(5) === ".")
+    assert(bedCols(6) === "150")
+    assert(bedCols(7) === "26400")
+    assert(bedCols(8) === "0,0,0")
+    assert(bedCols(9) === ".")
+    assert(bedCols(10) === ".")
+    assert(bedCols(11) === ".")
+
+    val actual = sc.loadBed(outputPath)
+    val pairs = expected.rdd.collect.zip(actual.rdd.collect)
+    pairs.foreach(p => {
+      assert(p._1 === p._2)
+    })
+
+    checkFiles(inputPath, outputPath)
+  }
+
   sparkTest("save IntervalList as GTF format") {
     val inputPath = testFile("SeqCap_EZ_Exome_v3.hg19.interval_list")
     val features = sc.loadIntervalList(inputPath)