package jalview.ext.ensembl;
+import jalview.analysis.AlignmentUtils;
import jalview.datamodel.Alignment;
import jalview.datamodel.AlignmentI;
import jalview.datamodel.DBRefEntry;
import jalview.exceptions.JalviewException;
import jalview.io.FastaFile;
import jalview.io.FileParse;
-import jalview.io.gff.SequenceOntology;
+import jalview.io.gff.SequenceOntologyFactory;
+import jalview.io.gff.SequenceOntologyI;
import jalview.util.DBRefUtils;
import jalview.util.MapList;
/**
* Base class for Ensembl sequence fetchers
*
+ * @see http://rest.ensembl.org/documentation/info/sequence_id
* @author gmcarstairs
*/
public abstract class EnsemblSeqProxy extends EnsemblRestClient
{
+ private static final List<String> CROSS_REFERENCES = Arrays
+ .asList(new String[] { "CCDS", "Uniprot/SWISSPROT",
+ "Uniprot/SPTREMBL" });
+
protected static final String CONSEQUENCE_TYPE = "consequence_type";
protected static final String PARENT = "Parent";
protected static final String ID = "ID";
+ protected static final String NAME = "Name";
+
+ protected static final String DESCRIPTION = "description";
+
/*
- * this needs special handling, as it isA sequence_variant in the
- * Sequence Ontology, but behaves in Ensembl as if it isA transcript
+ * enum for 'type' parameter to the /sequence REST service
*/
- protected static final String NMD_VARIANT = "NMD_transcript_variant";
-
public enum EnsemblSeqType
{
/**
- * type=genomic for the full dna including introns
+ * type=genomic to fetch full dna including introns
*/
GENOMIC("genomic"),
/**
- * type=cdna for transcribed dna including UTRs
+ * type=cdna to fetch dna including UTRs
*/
CDNA("cdna"),
/**
- * type=cds for coding dna excluding UTRs
+ * type=cds to fetch coding dna excluding UTRs
*/
CDS("cds"),
/**
- * type=protein for the peptide product sequence
+ * type=protein to fetch peptide product sequence
*/
PROTEIN("protein");
}
/**
- * Constructor
+ * Default constructor (to use rest.ensembl.org)
*/
public EnsemblSeqProxy()
{
+ super();
+ }
+
+ /**
+ * Constructor given the target domain to fetch data from
+ */
+ public EnsemblSeqProxy(String d)
+ {
+ super(d);
}
/**
* Makes the sequence queries to Ensembl's REST service and returns an
- * alignment consisting of the returned sequences. This overloaded method
- * allows the genomic sequence (with features) to be passed in if it has
- * already been retrieved, to avoid repeat calls to fetch it.
+ * alignment consisting of the returned sequences.
*/
- public AlignmentI getSequenceRecords(String query,
- SequenceI genomicSequence) throws Exception
+ @Override
+ public AlignmentI getSequenceRecords(String query) throws Exception
{
- long now = System.currentTimeMillis();
// TODO use a String... query vararg instead?
// danger: accession separator used as a regex here, a string elsewhere
+ " chunks. Unexpected problem (" + r.getLocalizedMessage()
+ ")";
System.err.println(msg);
- if (alignment != null)
- {
- break; // return what we got
- }
- else
- {
- throw new JalviewException(msg, r);
- }
+ break;
}
}
+ if (alignment == null)
+ {
+ return null;
+ }
+
/*
- * fetch and transfer genomic sequence features
+ * fetch and transfer genomic sequence features,
+ * fetch protein product and add as cross-reference
*/
for (String accId : allIds)
{
- addFeaturesAndProduct(accId, alignment, genomicSequence);
+ addFeaturesAndProduct(accId, alignment);
+ }
+
+ for (SequenceI seq : alignment.getSequences())
+ {
+ getCrossReferences(seq);
}
- inProgress = false;
- System.out.println(getClass().getName() + " took "
- + (System.currentTimeMillis() - now) + "ms to fetch");
return alignment;
}
* @param accId
* @param alignment
*/
- protected void addFeaturesAndProduct(String accId, AlignmentI alignment,
- SequenceI genomicSequence)
+ protected void addFeaturesAndProduct(String accId, AlignmentI alignment)
{
+ if (alignment == null)
+ {
+ return;
+ }
+
try
{
/*
* get 'dummy' genomic sequence with exon, cds and variation features
*/
- if (genomicSequence == null)
+ SequenceI genomicSequence = null;
+ EnsemblFeatures gffFetcher = new EnsemblFeatures(getDomain());
+ EnsemblFeatureType[] features = getFeaturesToFetch();
+ AlignmentI geneFeatures = gffFetcher.getSequenceRecords(accId,
+ features);
+ if (geneFeatures.getHeight() > 0)
{
- EnsemblOverlap gffFetcher = new EnsemblOverlap();
- EnsemblFeatureType[] features = getFeaturesToFetch();
- AlignmentI geneFeatures = gffFetcher.getSequenceRecords(accId,
- features);
- if (geneFeatures.getHeight() > 0)
- {
- genomicSequence = geneFeatures.getSequenceAt(0);
- }
+ genomicSequence = geneFeatures.getSequenceAt(0);
}
if (genomicSequence != null)
{
String accId = querySeq.getName();
try
{
- AlignmentI protein = new EnsemblProtein().getSequenceRecords(accId);
+ AlignmentI protein = new EnsemblProtein(getDomain())
+ .getSequenceRecords(accId);
if (protein == null || protein.getHeight() == 0)
{
- System.out.println("Failed to retrieve protein for " + accId);
+ System.out.println("No protein product found for " + accId);
return;
}
SequenceI proteinSeq = protein.getSequenceAt(0);
proteinSeq.createDatasetSequence();
querySeq.createDatasetSequence();
- MapList mapList = mapCdsToProtein(querySeq, proteinSeq);
+ MapList mapList = AlignmentUtils.mapCdsToProtein(querySeq, proteinSeq);
if (mapList != null)
{
- Mapping map = new Mapping(proteinSeq.getDatasetSequence(), mapList);
+ // clunky: ensure Uniprot xref if we have one is on mapped sequence
+ SequenceI ds = proteinSeq.getDatasetSequence();
+ ds.setSourceDBRef(proteinSeq.getSourceDBRef());
+ Mapping map = new Mapping(ds, mapList);
DBRefEntry dbr = new DBRefEntry(getDbSource(), getDbVersion(),
accId, map);
querySeq.getDatasetSequence().addDBRef(dbr);
+
+ /*
+ * copy exon features to protein, compute peptide variants from dna
+ * variants and add as features on the protein sequence ta-da
+ */
+ AlignmentUtils.computeProteinFeatures(querySeq, proteinSeq, mapList);
}
} catch (Exception e)
{
}
/**
- * Returns a mapping from dna to protein by inspecting sequence features of
- * type "CDS" on the dna.
+ * Get database xrefs from Ensembl, and attach them to the sequence
*
- * @param dnaSeq
- * @param proteinSeq
- * @return
+ * @param seq
*/
- protected MapList mapCdsToProtein(SequenceI dnaSeq, SequenceI proteinSeq)
+ protected void getCrossReferences(SequenceI seq)
{
- SequenceFeature[] sfs = dnaSeq.getSequenceFeatures();
- if (sfs == null)
+ while (seq.getDatasetSequence() != null)
{
- return null;
+ seq = seq.getDatasetSequence();
}
- List<int[]> ranges = new ArrayList<int[]>(50);
- SequenceOntology so = SequenceOntology.getInstance();
-
- int mappedDnaLength = 0;
-
- /*
- * Map CDS columns of dna to peptide. No need to worry about reverse strand
- * dna here since the retrieved sequence is as transcribed (reverse
- * complement for reverse strand), i.e in the same sense as the peptide.
- */
- boolean fivePrimeIncomplete = false;
- for (SequenceFeature sf : sfs)
+ EnsemblXref xrefFetcher = new EnsemblXref(getDomain());
+ List<DBRefEntry> xrefs = xrefFetcher.getCrossReferences(seq.getName(),
+ getCrossReferenceDatabases());
+ for (DBRefEntry xref : xrefs)
{
+ seq.addDBRef(xref);
/*
- * process a CDS feature (or a sub-type of CDS)
+ * Save any Uniprot xref to be the reference for SIFTS mapping
*/
- if (so.isA(sf.getType(), SequenceOntology.CDS))
+ if (DBRefSource.UNIPROT.equals(xref.getSource()))
{
- int phase = 0;
- try {
- phase = Integer.parseInt(sf.getPhase());
- } catch (NumberFormatException e)
- {
- // ignore
- }
- /*
- * phase > 0 on first codon means 5' incomplete - skip to the start
- * of the next codon; example ENST00000496384
- */
- int begin = sf.getBegin();
- int end = sf.getEnd();
- if (ranges.isEmpty() && phase > 0)
- {
- fivePrimeIncomplete = true;
- begin += phase;
- if (begin > end)
- {
- continue; // shouldn't happen?
- }
- }
- ranges.add(new int[] { begin, end });
- mappedDnaLength += Math.abs(end - begin) + 1;
+ seq.setSourceDBRef(xref);
}
}
- int proteinLength = proteinSeq.getLength();
- List<int[]> proteinRange = new ArrayList<int[]>();
- int proteinStart = 1;
- if (fivePrimeIncomplete && proteinSeq.getCharAt(0) == 'X')
- {
- proteinStart = 2;
- proteinLength--;
- }
- proteinRange.add(new int[] { proteinStart, proteinLength });
+ }
- /*
- * dna length should map to protein (or protein plus stop codon)
- */
- int codesForResidues = mappedDnaLength / 3;
- if (codesForResidues == proteinLength
- || codesForResidues == (proteinLength + 1))
- {
- return new MapList(ranges, proteinRange, 3, 1);
- }
- return null;
+ /**
+ * Returns a list of database names to be used when fetching cross-references.
+ * Specifically, the names are used to filter data returned by the Ensembl
+ * xrefs REST service on the value in field 'dbname'.
+ *
+ * @return
+ */
+ protected List<String> getCrossReferenceDatabases()
+ {
+ return CROSS_REFERENCES;
}
/**
"Only retrieved %d sequences for %d query strings", fr
.getSeqs().size(), ids.size()));
}
+
+ if (fr.getSeqs().size() == 1 && fr.getSeqs().get(0).getLength() == 0)
+ {
+ /*
+ * POST request has returned an empty FASTA file e.g. for invalid id
+ */
+ throw new IOException("No data returned for " + ids);
+ }
+
if (fr.getSeqs().size() > 0)
{
AlignmentI seqal = new Alignment(
@Override
protected URL getUrl(List<String> ids) throws MalformedURLException
{
- // ids are not used - they go in the POST body instead
+ /*
+ * a single id is included in the URL path
+ * multiple ids go in the POST body instead
+ */
StringBuffer urlstring = new StringBuffer(128);
- urlstring.append(SEQUENCE_ID_URL);
-
+ urlstring.append(getDomain() + "/sequence/id");
+ if (ids.size() == 1)
+ {
+ urlstring.append("/").append(ids.get(0));
+ }
// @see https://github.com/Ensembl/ensembl-rest/wiki/Output-formats
urlstring.append("?type=").append(getSourceEnsemblType().getType());
urlstring.append(("&Accept=text/x-fasta"));
}
@Override
- protected String getRequestMimeType()
+ protected String getRequestMimeType(boolean multipleIds)
{
- return "application/json";
+ return multipleIds ? "application/json" : "text/x-fasta";
}
@Override
* the start position of the sequence we are mapping to
* @return
*/
- protected MapList getGenomicRanges(SequenceI sourceSequence,
+ protected MapList getGenomicRangesFromFeatures(SequenceI sourceSequence,
String accId, int start)
{
SequenceFeature[] sfs = sourceSequence.getSequenceFeatures();
*/
if (identifiesSequence(sf, accId))
{
- int strand = sf.getStrand();
-
- if (directionSet && strand != direction)
- {
- // abort - mix of forward and backward
+ int strand = sf.getStrand();
+ strand = strand == 0 ? 1 : strand; // treat unknown as forward
+
+ if (directionSet && strand != direction)
+ {
+ // abort - mix of forward and backward
System.err.println("Error: forward and backward strand for "
+ accId);
return null;
*/
Collections.sort(regions, new RangeSorter(direction == 1));
- List<int[]> to = new ArrayList<int[]>();
- to.add(new int[] { start, start + mappedLength - 1 });
+ List<int[]> to = Arrays.asList(new int[] { start,
+ start + mappedLength - 1 });
return new MapList(regions, to, 1, 1);
}
/*
* for sequence_variant, make an additional feature with consequence
*/
- if (SequenceOntology.getInstance().isSequenceVariant(sf.getType()))
- {
- String consequence = (String) sf.getValue(CONSEQUENCE_TYPE);
- if (consequence != null)
- {
- SequenceFeature sf2 = new SequenceFeature("consequence",
- consequence, copy.getBegin(), copy.getEnd(), 0f,
- null);
- targetSequence.addSequenceFeature(sf2);
- }
- }
+ // if (SequenceOntologyFactory.getInstance().isA(sf.getType(),
+ // SequenceOntologyI.SEQUENCE_VARIANT))
+ // {
+ // String consequence = (String) sf.getValue(CONSEQUENCE_TYPE);
+ // if (consequence != null)
+ // {
+ // SequenceFeature sf2 = new SequenceFeature("consequence",
+ // consequence, copy.getBegin(), copy.getEnd(), 0f,
+ // null);
+ // targetSequence.addSequenceFeature(sf2);
+ // }
+ // }
}
}
return false;
}
+ // long start = System.currentTimeMillis();
SequenceFeature[] sfs = sourceSequence.getSequenceFeatures();
- MapList mapping = getGenomicRanges(sourceSequence, accessionId,
+ MapList mapping = getGenomicRangesFromFeatures(sourceSequence, accessionId,
targetSequence.getStart());
if (mapping == null)
{
return false;
}
- return transferFeatures(sfs, targetSequence, mapping, accessionId);
+ boolean result = transferFeatures(sfs, targetSequence, mapping,
+ accessionId);
+ // System.out.println("transferFeatures (" + (sfs.length) + " --> "
+ // + targetSequence.getSequenceFeatures().length + ") to "
+ // + targetSequence.getName()
+ // + " took " + (System.currentTimeMillis() - start) + "ms");
+ return result;
}
/**
+ " sequence with variant features";
}
- @Override
- public AlignmentI getSequenceRecords(String identifier) throws Exception
- {
- return getSequenceRecords(identifier, null);
- }
-
/**
* Returns a (possibly empty) list of features on the sequence which have the
* specified sequence ontology type (or a sub-type of it), and the given
SequenceFeature[] sfs = sequence.getSequenceFeatures();
if (sfs != null) {
- SequenceOntology so = SequenceOntology.getInstance();
+ SequenceOntologyI so = SequenceOntologyFactory.getInstance();
for (SequenceFeature sf :sfs) {
if (so.isA(sf.getType(), type))
{
*/
public static boolean isTranscript(String featureType)
{
- return NMD_VARIANT.equals(featureType)
- || SequenceOntology.getInstance().isA(featureType, SequenceOntology.TRANSCRIPT);
+ return SequenceOntologyI.NMD_TRANSCRIPT_VARIANT.equals(featureType)
+ || SequenceOntologyFactory.getInstance().isA(featureType,
+ SequenceOntologyI.TRANSCRIPT);
}
}