X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fdatamodel%2FAlignedCodonFrame.java;fp=src%2Fjalview%2Fdatamodel%2FAlignedCodonFrame.java;h=f5154ec1da1d151cebd4e240ff86d20914c90c1e;hb=d043ce47fc710d3eb2629ba926a8a7417bd67d8c;hp=fffa13759a9c1cc7c9b9048ce11257d272502e71;hpb=49db0dff1da16c3355b43a41498c1fc93ef47e91;p=jalview.git

diff --git a/src/jalview/datamodel/AlignedCodonFrame.java b/src/jalview/datamodel/AlignedCodonFrame.java
index fffa137..f5154ec 100644
--- a/src/jalview/datamodel/AlignedCodonFrame.java
+++ b/src/jalview/datamodel/AlignedCodonFrame.java
@@ -20,13 +20,13 @@
  */
 package jalview.datamodel;
 
-import jalview.util.MapList;
-import jalview.util.MappingUtils;
-
 import java.util.AbstractList;
 import java.util.ArrayList;
 import java.util.List;
 
+import jalview.util.MapList;
+import jalview.util.MappingUtils;
+
 /**
  * Stores mapping between the columns of a protein alignment and a DNA alignment
  * and a list of individual codon to amino acid mappings between sequences.
@@ -107,6 +107,151 @@ public class AlignedCodonFrame
     {
       return mapping;
     }
+
+    /**
+     * Returns true if the mapping covers the full length of the given sequence.
+     * This allows us to distinguish the CDS that codes for a protein from
+     * another overlapping CDS in the parent dna sequence.
+     * 
+     * @param seq
+     * @return
+     */
+    public boolean covers(SequenceI seq)
+    {
+      return covers(seq, false, false);
+    }
+
+    /**
+     * 
+     * @param seq
+     * @param localCover
+     *          - when true - compare extent of seq's dataset sequence rather
+     *          than the local extent
+     * @param either
+     *          - when true coverage is required for either seq or the mapped
+     *          sequence
+     * @return true if mapping covers full length of given sequence (or the
+     *         other if either==true)
+     */
+    public boolean covers(SequenceI seq, boolean localCover, boolean either)
+    {
+      List<int[]> mappedRanges = null, otherRanges = null;
+      MapList mapList = mapping.getMap();
+      int mstart = seq.getStart(), mend = seq.getEnd(), ostart, oend;
+      ;
+      if (fromSeq == seq || fromSeq == seq.getDatasetSequence())
+      {
+        if (localCover && fromSeq != seq)
+        {
+          mstart = fromSeq.getStart();
+          mend = fromSeq.getEnd();
+        }
+        mappedRanges = mapList.getFromRanges();
+        otherRanges = mapList.getToRanges();
+        ostart = mapping.to.getStart();
+        oend = mapping.to.getEnd();
+      }
+      else if (mapping.to == seq || mapping.to == seq.getDatasetSequence())
+      {
+        if (localCover && mapping.to != seq)
+        {
+          mstart = mapping.to.getStart();
+          mend = mapping.to.getEnd();
+        }
+        mappedRanges = mapList.getToRanges();
+        otherRanges = mapList.getFromRanges();
+        ostart = fromSeq.getStart();
+        oend = fromSeq.getEnd();
+      }
+      else
+      {
+        return false;
+      }
+
+      /*
+       * check that each mapped range lies within the sequence range
+       * (necessary for circular CDS - example EMBL:J03321:AAA91567)
+       * and mapped length covers (at least) sequence length
+       */
+      int length = countRange(mappedRanges, mstart, mend);
+
+      if (length != -1)
+      {
+        // add 3 to mapped length to allow for a mapped stop codon
+        if (length + 3 >= (mend - mstart + 1))
+        {
+          return true;
+        }
+      }
+      if (either)
+      {
+        // also check coverage of the other range
+        length = countRange(otherRanges, ostart, oend);
+        if (length != -1)
+        {
+          if (length + 1 >= (oend - ostart + 1))
+          {
+            return true;
+          }
+        }
+      }
+      return false;
+    }
+
+    private int countRange(List<int[]> mappedRanges, int mstart, int mend)
+    {
+      int length = 0;
+      for (int[] range : mappedRanges)
+      {
+        int from = Math.min(range[0], range[1]);
+        int to = Math.max(range[0], range[1]);
+        if (from < mstart || to > mend)
+        {
+          return -1;
+        }
+        length += (to - from + 1);
+      }
+      return length;
+    }
+
+    /**
+     * Adds any regions mapped to or from position {@code pos} in sequence
+     * {@code seq} to the given search results Note: recommend first using the
+     * .covers(,true,true) to ensure mapping covers both sequences
+     * 
+     * @param seq
+     * @param pos
+     * @param sr
+     */
+    public void markMappedRegion(SequenceI seq, int pos, SearchResultsI sr)
+    {
+      int[] codon = null;
+      SequenceI mappedSeq = null;
+      SequenceI ds = seq.getDatasetSequence();
+      if (ds == null)
+      {
+        ds = seq;
+      }
+
+      if (this.fromSeq == seq || this.fromSeq == ds)
+      {
+        codon = this.mapping.map.locateInTo(pos, pos);
+        mappedSeq = this.mapping.to;
+      }
+      else if (this.mapping.to == seq || this.mapping.to == ds)
+      {
+        codon = this.mapping.map.locateInFrom(pos, pos);
+        mappedSeq = this.fromSeq;
+      }
+
+      if (codon != null)
+      {
+        for (int i = 0; i < codon.length; i += 2)
+        {
+          sr.addResult(mappedSeq, codon[i], codon[i + 1]);
+        }
+      }
+    }
   }
 
   private List<SequenceToSequenceMapping> mappings;
@@ -261,9 +406,12 @@ public class AlignedCodonFrame
   }
 
   /**
+   * Return the corresponding aligned or dataset dna sequence for given amino
+   * acid sequence, or null if not found. returns the sequence from the first
+   * mapping found that involves the protein sequence.
    * 
-   * @param sequenceRef
-   * @return null or corresponding aaSeq entry for dnaSeq entry
+   * @param aaSeqRef
+   * @return
    */
   public SequenceI getDnaForAaSeq(SequenceI aaSeqRef)
   {
@@ -293,7 +441,8 @@ public class AlignedCodonFrame
 
   /**
    * Add search results for regions in other sequences that translate or are
-   * translated from a particular position in seq
+   * translated from a particular position in seq (which may be an aligned or
+   * dataset sequence)
    * 
    * @param seq
    * @param index
@@ -304,69 +453,18 @@ public class AlignedCodonFrame
   public void markMappedRegion(SequenceI seq, int index,
           SearchResultsI results)
   {
-    int[] codon;
     SequenceI ds = seq.getDatasetSequence();
-    for (SequenceToSequenceMapping ssm : mappings)
+    if (ds == null)
     {
-      if (ssm.fromSeq == seq || ssm.fromSeq == ds)
-      {
-        codon = ssm.mapping.map.locateInTo(index, index);
-        if (codon != null)
-        {
-          for (int i = 0; i < codon.length; i += 2)
-          {
-            results.addResult(ssm.mapping.to, codon[i], codon[i + 1]);
-          }
-        }
-      }
-      else if (ssm.mapping.to == seq || ssm.mapping.to == ds)
-      {
-        {
-          codon = ssm.mapping.map.locateInFrom(index, index);
-          if (codon != null)
-          {
-            for (int i = 0; i < codon.length; i += 2)
-            {
-              results.addResult(ssm.fromSeq, codon[i], codon[i + 1]);
-            }
-          }
-        }
-      }
+      ds = seq;
     }
-  }
-
-  /**
-   * Returns the DNA codon positions (base 1) for the given position (base 1) in
-   * a mapped protein sequence, or null if no mapping is found.
-   * 
-   * Intended for use in aligning cDNA to match aligned protein. Only the first
-   * mapping found is returned, so not suitable for use if multiple protein
-   * sequences are mapped to the same cDNA (but aligning cDNA as protein is
-   * ill-defined for this case anyway).
-   * 
-   * @param seq
-   *          the DNA dataset sequence
-   * @param aaPos
-   *          residue position (base 1) in a protein sequence
-   * @return
-   */
-  public int[] getDnaPosition(SequenceI seq, int aaPos)
-  {
-    /*
-     * Adapted from markMappedRegion().
-     */
-    MapList ml = null;
-    int i = 0;
     for (SequenceToSequenceMapping ssm : mappings)
     {
-      if (ssm.fromSeq == seq)
+      if (ssm.covers(seq, true, true))
       {
-        ml = getdnaToProt()[i];
-        break;
+        ssm.markMappedRegion(ds, index, results);
       }
-      i++;
     }
-    return ml == null ? null : ml.locateInFrom(aaPos, aaPos);
   }
 
   /**
@@ -767,7 +865,7 @@ public class AlignedCodonFrame
    * Two AlignedCodonFrame objects are equal if they hold the same ordered list
    * of mappings
    * 
-   * @see SequenceToSequenceMapping#
+   * @see SequenceToSequenceMapping#equals
    */
   @Override
   public boolean equals(Object obj)
@@ -783,4 +881,55 @@ public class AlignedCodonFrame
   {
     return mappings;
   }
+
+  /**
+   * Returns the first mapping found which is between the two given sequences,
+   * and covers the full extent of both.
+   * 
+   * @param seq1
+   * @param seq2
+   * @return
+   */
+  public SequenceToSequenceMapping getCoveringMapping(SequenceI seq1,
+          SequenceI seq2)
+  {
+    for (SequenceToSequenceMapping mapping : mappings)
+    {
+      if (mapping.covers(seq2) && mapping.covers(seq1))
+      {
+        return mapping;
+      }
+    }
+    return null;
+  }
+
+  /**
+   * Returns the first mapping found which is between the given dataset sequence
+   * and another, is a triplet mapping (3:1 or 1:3), and covers the full extent
+   * of both sequences involved
+   * 
+   * @param seq
+   * @return
+   */
+  public SequenceToSequenceMapping getCoveringCodonMapping(SequenceI seq)
+  {
+    for (SequenceToSequenceMapping mapping : mappings)
+    {
+      if (mapping.getMapping().getMap().isTripletMap()
+              && mapping.covers(seq))
+      {
+        if (mapping.fromSeq == seq
+                && mapping.covers(mapping.getMapping().getTo()))
+        {
+          return mapping;
+        }
+        else if (mapping.getMapping().getTo() == seq
+                && mapping.covers(mapping.fromSeq))
+        {
+          return mapping;
+        }
+      }
+    }
+    return null;
+  }
 }