JAL-2617 unit tests for CDS + stop codon, CDS with start phase

diff --git a/src/jalview/analysis/AlignmentUtils.java b/src/jalview/analysis/AlignmentUtils.java
index 311a9a3..90d9197 100644 (file)
--- a/src/jalview/analysis/AlignmentUtils.java
+++ b/src/jalview/analysis/AlignmentUtils.java
@@ -2164,7 +2164,10 @@ public class AlignmentUtils
 
   /**
    * Returns a mapping from dna to protein by inspecting sequence features of
-   * type "CDS" on the dna.
+   * type "CDS" on the dna. A mapping is constructed if the total CDS feature
+   * length is 3 times the peptide length (optionally after dropping a trailing
+   * stop codon). This method does not check whether the CDS nucleotide sequence
+   * translates to the peptide sequence.
    * 
    * @param dnaSeq
    * @param proteinSeq
@@ -2176,6 +2179,15 @@ public class AlignmentUtils
     List<int[]> ranges = findCdsPositions(dnaSeq);
     int mappedDnaLength = MappingUtils.getLength(ranges);
 
+    /*
+     * if not a whole number of codons, something is wrong,
+     * abort mapping
+     */
+    if (mappedDnaLength % CODON_LENGTH > 0)
+    {
+      return null;
+    }
+
     int proteinLength = proteinSeq.getLength();
     int proteinStart = proteinSeq.getStart();
     int proteinEnd = proteinSeq.getEnd();
@@ -2199,6 +2211,7 @@ public class AlignmentUtils
     if (codesForResidues == (proteinLength + 1))
     {
       // assuming extra codon is for STOP and not in peptide
+      // todo: check trailing codon is indeed a STOP codon
       codesForResidues--;
       mappedDnaLength -= CODON_LENGTH;
       MappingUtils.removeEndPositions(CODON_LENGTH, ranges);
@@ -2233,7 +2246,6 @@ public class AlignmentUtils
       return result;
     }
     SequenceFeatures.sortFeatures(sfs, true);
-    int startPhase = 0;
 
     for (SequenceFeature sf : sfs)
     {
@@ -2251,7 +2263,7 @@ public class AlignmentUtils
        */
       int begin = sf.getBegin();
       int end = sf.getEnd();
-      if (result.isEmpty())
+      if (result.isEmpty() && phase > 0)
       {
         begin += phase;
         if (begin > end)
@@ -2266,16 +2278,6 @@ public class AlignmentUtils
     }
 
     /*
-     * remove 'startPhase' positions (usually 0) from the first range 
-     * so we begin at the start of a complete codon
-     */
-    if (!result.isEmpty())
-    {
-      // TODO JAL-2022 correctly model start phase > 0
-      result.get(0)[0] += startPhase;
-    }
-
-    /*
      * Finally sort ranges by start position. This avoids a dependency on 
      * keeping features in order on the sequence (if they are in order anyway,
      * the sort will have almost no work to do). The implicit assumption is CDS

diff --git a/test/jalview/analysis/AlignmentUtilsTests.java b/test/jalview/analysis/AlignmentUtilsTests.java
index 4439bb9..06b51e6 100644 (file)
--- a/test/jalview/analysis/AlignmentUtilsTests.java
+++ b/test/jalview/analysis/AlignmentUtilsTests.java
@@ -2533,4 +2533,71 @@ public class AlignmentUtilsTests
     assertEquals(s_as3, uas3.getSequenceAsString());
   }
 
+  /**
+   * Tests for the method that maps nucleotide to protein based on CDS features
+   */
+  @Test(groups = "Functional")
+  public void testMapCdsToProtein()
+  {
+    SequenceI peptide = new Sequence("pep", "KLQ");
+
+    /*
+     * Case 1: CDS 3 times length of peptide
+     * NB method only checks lengths match, not translation
+     */
+    SequenceI dna = new Sequence("dna", "AACGacgtCTCCT");
+    dna.createDatasetSequence();
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 1, 4, null));
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 9, 13, null));
+    MapList ml = AlignmentUtils.mapCdsToProtein(dna, peptide);
+    assertEquals(3, ml.getFromRatio());
+    assertEquals(1, ml.getToRatio());
+    assertEquals("[[1, 3]]",
+            Arrays.deepToString(ml.getToRanges().toArray()));
+    assertEquals("[[1, 4], [9, 13]]",
+            Arrays.deepToString(ml.getFromRanges().toArray()));
+
+    /*
+     * Case 2: CDS 3 times length of peptide + stop codon
+     * (note code does not currently check trailing codon is a stop codon)
+     */
+    dna = new Sequence("dna", "AACGacgtCTCCTTGA");
+    dna.createDatasetSequence();
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 1, 4, null));
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 9, 16, null));
+    ml = AlignmentUtils.mapCdsToProtein(dna, peptide);
+    assertEquals(3, ml.getFromRatio());
+    assertEquals(1, ml.getToRatio());
+    assertEquals("[[1, 3]]",
+            Arrays.deepToString(ml.getToRanges().toArray()));
+    assertEquals("[[1, 4], [9, 13]]",
+            Arrays.deepToString(ml.getFromRanges().toArray()));
+
+    /*
+     * Case 3: CDS not 3 times length of peptide - no mapping is made
+     */
+    dna = new Sequence("dna", "AACGacgtCTCCTTG");
+    dna.createDatasetSequence();
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 1, 4, null));
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 9, 15, null));
+    ml = AlignmentUtils.mapCdsToProtein(dna, peptide);
+    assertNull(ml);
+
+    /*
+     * Case 4: incomplete start codon corresponding to X in peptide
+     */
+    dna = new Sequence("dna", "ACGacgtCTCCTTGG");
+    dna.createDatasetSequence();
+    SequenceFeature sf = new SequenceFeature("CDS", "", 1, 3, null);
+    sf.setPhase("2"); // skip 2 positions (AC) to start of next codon (GCT)
+    dna.addSequenceFeature(sf);
+    dna.addSequenceFeature(new SequenceFeature("CDS", "", 8, 15, null));
+    peptide = new Sequence("pep", "XLQ");
+    ml = AlignmentUtils.mapCdsToProtein(dna, peptide);
+    assertEquals("[[2, 3]]",
+            Arrays.deepToString(ml.getToRanges().toArray()));
+    assertEquals("[[3, 3], [8, 12]]",
+            Arrays.deepToString(ml.getFromRanges().toArray()));
+  }
+
 }