getXrefViews()
{
return xrefViews;
}
@Override
public void run()
{
final long sttime = System.currentTimeMillis();
alignFrame.setProgressBar(MessageManager.formatMessage(
"status.searching_for_sequences_from", new Object[]
{ source }), sttime);
try
{
AlignmentI alignment = alignFrame.getViewport().getAlignment();
AlignmentI dataset = alignment.getDataset() == null ? alignment
: alignment.getDataset();
boolean dna = alignment.isNucleotide();
if (_odna != dna)
{
System.err
.println("Conflict: showProducts for alignment originally "
+ "thought to be " + (_odna ? "DNA" : "Protein")
+ " now searching for " + (dna ? "DNA" : "Protein")
+ " Context.");
}
AlignmentI xrefs = new CrossRef(sel, dataset)
.findXrefSequences(source, dna);
if (xrefs == null)
{
return;
}
/*
* try to look up chromosomal coordinates for nucleotide
* sequences (if not already retrieved)
*/
findGeneLoci(xrefs.getSequences());
/*
* get display scheme (if any) to apply to features
*/
FeatureSettingsModelI featureColourScheme = new SequenceFetcher()
.getFeatureColourScheme(source);
AlignmentI xrefsAlignment = makeCrossReferencesAlignment(dataset,
xrefs);
if (!dna)
{
xrefsAlignment = AlignmentUtils.makeCdsAlignment(
xrefsAlignment.getSequencesArray(), dataset, sel);
xrefsAlignment.alignAs(alignment);
}
/*
* If we are opening a splitframe, make a copy of this alignment (sharing the same dataset
* sequences). If we are DNA, drop introns and update mappings
*/
AlignmentI copyAlignment = null;
if (Cache.getDefault(Preferences.ENABLE_SPLIT_FRAME, true))
{
copyAlignment = copyAlignmentForSplitFrame(alignment, dataset, dna,
xrefs, xrefsAlignment);
if (copyAlignment == null)
{
return; // failed
}
}
/*
* build AlignFrame(s) according to available alignment data
*/
AlignFrame newFrame = new AlignFrame(xrefsAlignment,
AlignFrame.DEFAULT_WIDTH, AlignFrame.DEFAULT_HEIGHT);
if (Cache.getDefault("HIDE_INTRONS", true))
{
newFrame.hideFeatureColumns(SequenceOntologyI.EXON, false);
}
String newtitle = String.format("%s %s %s",
dna ? MessageManager.getString("label.proteins")
: MessageManager.getString("label.nucleotides"),
MessageManager.getString("label.for"), alignFrame.getTitle());
newFrame.setTitle(newtitle);
if (copyAlignment == null)
{
/*
* split frame display is turned off in preferences file
*/
Desktop.addInternalFrame(newFrame, newtitle,
AlignFrame.DEFAULT_WIDTH, AlignFrame.DEFAULT_HEIGHT);
xrefViews.add(newFrame.alignPanel);
return; // via finally clause
}
AlignFrame copyThis = new AlignFrame(copyAlignment,
AlignFrame.DEFAULT_WIDTH, AlignFrame.DEFAULT_HEIGHT);
copyThis.setTitle(alignFrame.getTitle());
boolean showSequenceFeatures = alignFrame.getViewport()
.isShowSequenceFeatures();
newFrame.setShowSeqFeatures(showSequenceFeatures);
copyThis.setShowSeqFeatures(showSequenceFeatures);
FeatureRenderer myFeatureStyling = alignFrame.alignPanel
.getSeqPanel().seqCanvas.getFeatureRenderer();
/*
* copy feature rendering settings to split frame
*/
FeatureRenderer fr1 = newFrame.alignPanel.getSeqPanel().seqCanvas
.getFeatureRenderer();
fr1.transferSettings(myFeatureStyling);
fr1.findAllFeatures(true);
FeatureRenderer fr2 = copyThis.alignPanel.getSeqPanel().seqCanvas
.getFeatureRenderer();
fr2.transferSettings(myFeatureStyling);
fr2.findAllFeatures(true);
/*
* apply 'database source' feature configuration
* if any was found
*/
// TODO is this the feature colouring for the original
// alignment or the fetched xrefs? either could be Ensembl
newFrame.getViewport().applyFeaturesStyle(featureColourScheme);
copyThis.getViewport().applyFeaturesStyle(featureColourScheme);
SplitFrame sf = new SplitFrame(dna ? copyThis : newFrame,
dna ? newFrame : copyThis);
newFrame.setVisible(true);
copyThis.setVisible(true);
String linkedTitle = MessageManager
.getString("label.linked_view_title");
Desktop.addInternalFrame(sf, linkedTitle, -1, -1);
sf.adjustInitialLayout();
// finally add the top, then bottom frame to the view list
xrefViews.add(dna ? copyThis.alignPanel : newFrame.alignPanel);
xrefViews.add(!dna ? copyThis.alignPanel : newFrame.alignPanel);
} catch (OutOfMemoryError e)
{
new OOMWarning("whilst fetching crossreferences", e);
} catch (Throwable e)
{
Cache.log.error("Error when finding crossreferences", e);
} finally
{
alignFrame.setProgressBar(MessageManager.formatMessage(
"status.finished_searching_for_sequences_from", new Object[]
{ source }), sttime);
}
}
/**
* Tries to add chromosomal coordinates to any nucleotide sequence which does
* not already have them. Coordinates are retrieved from Ensembl given an
* Ensembl identifier, either on the sequence itself or on a peptide sequence
* it has a reference to.
*
*
* Example (human):
* - fetch EMBLCDS cross-references for Uniprot entry P30419
* - the EMBL sequences do not have xrefs to Ensembl
* - the Uniprot entry has xrefs to
* ENSP00000258960, ENSP00000468424, ENST00000258960, ENST00000592782
* - either of the transcript ids can be used to retrieve gene loci e.g.
* http://rest.ensembl.org/map/cds/ENST00000592782/1..100000
* Example (invertebrate):
* - fetch EMBLCDS cross-references for Uniprot entry Q43517 (FER1_SOLLC)
* - the Uniprot entry has an xref to ENSEMBLPLANTS Solyc10g044520.1.1
* - can retrieve gene loci with
* http://rest.ensemblgenomes.org/map/cds/Solyc10g044520.1.1/1..100000
*
*
* @param sequences
*/
public static void findGeneLoci(List sequences)
{
Map retrievedLoci = new HashMap<>();
for (SequenceI seq : sequences)
{
findGeneLoci(seq, retrievedLoci);
}
}
/**
* Tres to find chromosomal coordinates for the sequence, by searching its
* direct and indirect cross-references for Ensembl. If the loci have already
* been retrieved, just reads them out of the map of retrievedLoci; this is
* the case of an alternative transcript for the same protein. Otherwise calls
* a REST service to retrieve the loci, and if successful, adds them to the
* sequence and to the retrievedLoci.
*
* @param seq
* @param retrievedLoci
*/
static void findGeneLoci(SequenceI seq,
Map retrievedLoci)
{
/*
* don't replace any existing chromosomal coordinates
*/
if (seq == null || seq.isProtein() || seq.getGeneLoci() != null
|| seq.getDBRefs() == null)
{
return;
}
Set ensemblDivisions = new EnsemblInfo().getDivisions();
/*
* first look for direct dbrefs from sequence to Ensembl
*/
String[] divisionsArray = ensemblDivisions
.toArray(new String[ensemblDivisions.size()]);
DBRefEntry[] seqRefs = seq.getDBRefs();
DBRefEntry[] directEnsemblRefs = DBRefUtils.selectRefs(seqRefs,
divisionsArray);
if (directEnsemblRefs != null)
{
for (DBRefEntry ensemblRef : directEnsemblRefs)
{
if (fetchGeneLoci(seq, ensemblRef, retrievedLoci))
{
return;
}
}
}
/*
* else look for indirect dbrefs from sequence to Ensembl
*/
for (DBRefEntry dbref : seq.getDBRefs())
{
if (dbref.getMap() != null && dbref.getMap().getTo() != null)
{
DBRefEntry[] dbrefs = dbref.getMap().getTo().getDBRefs();
DBRefEntry[] indirectEnsemblRefs = DBRefUtils.selectRefs(dbrefs,
divisionsArray);
if (indirectEnsemblRefs != null)
{
for (DBRefEntry ensemblRef : indirectEnsemblRefs)
{
if (fetchGeneLoci(seq, ensemblRef, retrievedLoci))
{
return;
}
}
}
}
}
}
/**
* Retrieves chromosomal coordinates for the Ensembl (or EnsemblGenomes)
* identifier in dbref. If successful, and the sequence length matches gene
* loci length, then add it to the sequence, and to the retrievedLoci map.
* Answers true if successful, else false.
*
* @param seq
* @param dbref
* @param retrievedLoci
* @return
*/
static boolean fetchGeneLoci(SequenceI seq, DBRefEntry dbref,
Map retrievedLoci)
{
String accession = dbref.getAccessionId();
String division = dbref.getSource();
/*
* hack: ignore cross-references to Ensembl protein ids
* (or use map/translation perhaps?)
* todo: is there an equivalent in EnsemblGenomes?
*/
if (accession.startsWith("ENSP"))
{
return false;
}
EnsemblMap mapper = new EnsemblMap();
/*
* try CDS mapping first
*/
GeneLociI geneLoci = mapper.getCdsMapping(division, accession, 1,
seq.getLength());
if (geneLoci != null)
{
MapList map = geneLoci.getMapping();
int mappedFromLength = MappingUtils.getLength(map.getFromRanges());
if (mappedFromLength == seq.getLength())
{
seq.setGeneLoci(geneLoci.getSpeciesId(), geneLoci.getAssemblyId(),
geneLoci.getChromosomeId(), map);
retrievedLoci.put(dbref, geneLoci);
return true;
}
}
/*
* else try CDNA mapping
*/
geneLoci = mapper.getCdnaMapping(division, accession, 1,
seq.getLength());
if (geneLoci != null)
{
MapList map = geneLoci.getMapping();
int mappedFromLength = MappingUtils.getLength(map.getFromRanges());
if (mappedFromLength == seq.getLength())
{
seq.setGeneLoci(geneLoci.getSpeciesId(), geneLoci.getAssemblyId(),
geneLoci.getChromosomeId(), map);
retrievedLoci.put(dbref, geneLoci);
return true;
}
}
return false;
}
/**
* @param alignment
* @param dataset
* @param dna
* @param xrefs
* @param xrefsAlignment
* @return
*/
protected AlignmentI copyAlignmentForSplitFrame(AlignmentI alignment,
AlignmentI dataset, boolean dna, AlignmentI xrefs,
AlignmentI xrefsAlignment)
{
AlignmentI copyAlignment;
boolean copyAlignmentIsAligned = false;
if (dna)
{
copyAlignment = AlignmentUtils.makeCdsAlignment(sel, dataset,
xrefsAlignment.getSequencesArray());
if (copyAlignment.getHeight() == 0)
{
JvOptionPane.showMessageDialog(alignFrame,
MessageManager.getString("label.cant_map_cds"),
MessageManager.getString("label.operation_failed"),
JvOptionPane.OK_OPTION);
System.err.println("Failed to make CDS alignment");
return null;
}
/*
* pending getting Embl transcripts to 'align',
* we are only doing this for Ensembl
*/
// TODO proper criteria for 'can align as cdna'
if (DBRefSource.ENSEMBL.equalsIgnoreCase(source)
|| AlignmentUtils.looksLikeEnsembl(alignment))
{
copyAlignment.alignAs(alignment);
copyAlignmentIsAligned = true;
}
}
else
{
copyAlignment = AlignmentUtils.makeCopyAlignment(sel,
xrefs.getSequencesArray(), dataset);
}
copyAlignment
.setGapCharacter(alignFrame.viewport.getGapCharacter());
StructureSelectionManager ssm = StructureSelectionManager
.getStructureSelectionManager(Desktop.instance);
/*
* register any new mappings for sequence mouseover etc
* (will not duplicate any previously registered mappings)
*/
ssm.registerMappings(dataset.getCodonFrames());
if (copyAlignment.getHeight() <= 0)
{
System.err.println(
"No Sequences generated for xRef type " + source);
return null;
}
/*
* align protein to dna
*/
if (dna && copyAlignmentIsAligned)
{
xrefsAlignment.alignAs(copyAlignment);
}
else
{
/*
* align cdna to protein - currently only if
* fetching and aligning Ensembl transcripts!
*/
// TODO: generalise for other sources of locus/transcript/cds data
if (dna && DBRefSource.ENSEMBL.equalsIgnoreCase(source))
{
copyAlignment.alignAs(xrefsAlignment);
}
}
return copyAlignment;
}
/**
* Makes an alignment containing the given sequences, and adds them to the
* given dataset, which is also set as the dataset for the new alignment
*
* TODO: refactor to DatasetI method
*
* @param dataset
* @param seqs
* @return
*/
protected AlignmentI makeCrossReferencesAlignment(AlignmentI dataset,
AlignmentI seqs)
{
SequenceI[] sprods = new SequenceI[seqs.getHeight()];
for (int s = 0; s < sprods.length; s++)
{
sprods[s] = (seqs.getSequenceAt(s)).deriveSequence();
if (dataset.getSequences() == null || !dataset.getSequences()
.contains(sprods[s].getDatasetSequence()))
{
dataset.addSequence(sprods[s].getDatasetSequence());
}
sprods[s].updatePDBIds();
}
Alignment al = new Alignment(sprods);
al.setDataset(dataset);
return al;
}
/**
* Constructor
*
* @param af
* @param seqs
* @param fromDna
* @param dbSource
*/
CrossRefAction(AlignFrame af, SequenceI[] seqs, boolean fromDna,
String dbSource)
{
this.alignFrame = af;
this.sel = seqs;
this._odna = fromDna;
this.source = dbSource;
}
public static CrossRefAction getHandlerFor(final SequenceI[] sel,
final boolean fromDna, final String source,
final AlignFrame alignFrame)
{
return new CrossRefAction(alignFrame, sel, fromDna, source);
}
}