seqMappings = MappingUtils
.findMappingsForSequence(ds, mappings);
if (!seqMappings.isEmpty())
{
/*
* We assume here that only one protein mapping is expected per dna
* sequence. Mapping to multiple protein sequences is conceivable but
* undefined. Splitting a mapping to one protein sequence across
* multiple mappings is possible but pathological. Need closer
* constraints on the contents of AlignedCodonFrame.
*/
AlignedCodonFrame newMapping = new AlignedCodonFrame();
final SequenceI exonSequence = makeExonSequence(ds,
seqMappings.get(0), newMapping);
exonSequences.add(exonSequence);
newMappings.add(newMapping);
}
}
AlignmentI al = new Alignment(
exonSequences.toArray(new SequenceI[exonSequences.size()]));
al.setDataset(null);
/*
* Replace the old mappings with the new ones
*/
mappings.clear();
mappings.addAll(newMappings);
return al;
}
/**
* Helper method to make an exon-only sequence and populate its mapping to
* protein
*
* For example, if ggCCaTTcGAg has mappings [3, 4, 6, 7, 9, 10] to protein
* then generate a sequence CCTTGA with mapping [1, 6] to the same protein
* residues
*
* @param dnaSeq
* a dna dataset sequence
* @param mapping
* the current mapping of the sequence to protein
* @param newMapping
* the new mapping to populate, from the exon-only sequence
* @return
*/
protected static SequenceI makeExonSequence(SequenceI dnaSeq,
AlignedCodonFrame acf, AlignedCodonFrame newMapping)
{
Mapping mapping = acf.getMappingForSequence(dnaSeq);
final char[] dna = dnaSeq.getSequence();
StringBuilder newSequence = new StringBuilder(dnaSeq.getLength());
/*
* Get the codon regions as { [2, 5], [7, 12], [14, 14] etc }
*/
List exonRanges = mapping.getMap().getFromRanges();
for (int[] range : exonRanges)
{
for (int pos = range[0]; pos <= range[1]; pos++)
{
newSequence.append(dna[pos - 1]);
}
}
SequenceI exon = new Sequence(dnaSeq.getName(), newSequence.toString());
/*
* Locate any xrefs to CDS database on the protein product and attach to the
* CDS sequence. Also add as a sub-token of the sequence name.
*/
// default to "CDS" if we can't locate an actual gene id
String cdsAccId = FeatureProperties.getCodingFeature(DBRefSource.EMBL);
DBRefEntry[] cdsRefs = DBRefUtils.selectRefs(
mapping.getTo().getDBRef(), DBRefSource.CODINGDBS);
if (cdsRefs != null)
{
for (DBRefEntry cdsRef : cdsRefs)
{
exon.addDBRef(new DBRefEntry(cdsRef));
cdsAccId = cdsRef.getAccessionId();
}
}
exon.setName(exon.getName() + "|" + cdsAccId);
exon.createDatasetSequence();
/*
* Build new mappings - from the same protein regions, but now to contiguous
* exons
*/
List exonRange = new ArrayList();
exonRange.add(new int[]
{ 1, newSequence.length() });
MapList map = new MapList(exonRange, mapping.getMap().getToRanges(), 3, 1);
newMapping.addMap(exon.getDatasetSequence(), mapping.getTo(), map);
return exon;
}
}