+/*
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ *
+ * This file is part of Jalview.
+ *
+ * Jalview is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *
+ * Jalview is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ * PURPOSE. See the GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
+ */
package jalview.ext.ensembl;
+import jalview.analysis.AlignmentUtils;
+import jalview.analysis.Dna;
+import jalview.bin.Cache;
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.Comparison;
import jalview.util.DBRefUtils;
import jalview.util.MapList;
+import jalview.util.RangeComparator;
import java.io.IOException;
import java.net.MalformedURLException;
/**
* Base class for Ensembl sequence fetchers
*
+ * @see http://rest.ensembl.org/documentation/info/sequence_id
* @author gmcarstairs
*/
public abstract class EnsemblSeqProxy extends EnsemblRestClient
{
- protected static final String CONSEQUENCE_TYPE = "consequence_type";
+ private static final String ALLELES = "alleles";
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 coding 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");
}
/**
- * A comparator to sort ranges into ascending start position order
+ * Default constructor (to use rest.ensembl.org)
*/
- private class RangeSorter implements Comparator<int[]>
+ public EnsemblSeqProxy()
{
- boolean forwards;
-
- RangeSorter(boolean forward)
- {
- forwards = forward;
- }
-
- @Override
- public int compare(int[] o1, int[] o2)
- {
- return (forwards ? 1 : -1) * Integer.compare(o1[0], o2[0]);
- }
-
+ super();
}
/**
- * Constructor
+ * Constructor given the target domain to fetch data from
*/
- public EnsemblSeqProxy()
+ 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);
- }
+ r.printStackTrace();
+ 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);
- DBRefEntry dbr = new DBRefEntry(getDbSource(), getDbVersion(),
- accId, map);
+ // clunky: ensure Uniprot xref if we have one is on mapped sequence
+ SequenceI ds = proteinSeq.getDatasetSequence();
+ // TODO: Verify ensp primary ref is on proteinSeq.getDatasetSequence()
+ Mapping map = new Mapping(ds, mapList);
+ DBRefEntry dbr = new DBRefEntry(getDbSource(),
+ getEnsemblDataVersion(), proteinSeq.getName(), map);
querySeq.getDatasetSequence().addDBRef(dbr);
+ DBRefEntry[] uprots = DBRefUtils.selectRefs(ds.getDBRefs(),
+ new String[] { DBRefSource.UNIPROT });
+ DBRefEntry[] upxrefs = DBRefUtils.selectRefs(querySeq.getDBRefs(),
+ new String[] { DBRefSource.UNIPROT });
+ if (uprots != null)
+ {
+ for (DBRefEntry up : uprots)
+ {
+ // locate local uniprot ref and map
+ List<DBRefEntry> upx = DBRefUtils.searchRefs(upxrefs,
+ up.getAccessionId());
+ DBRefEntry upxref;
+ if (upx.size() != 0)
+ {
+ upxref = upx.get(0);
+
+ if (upx.size() > 1)
+ {
+ Cache.log
+ .warn("Implementation issue - multiple uniprot acc on product sequence.");
+ }
+ }
+ else
+ {
+ upxref = new DBRefEntry(DBRefSource.UNIPROT,
+ getEnsemblDataVersion(), up.getAccessionId());
+ }
+
+ Mapping newMap = new Mapping(ds, mapList);
+ upxref.setVersion(getEnsemblDataVersion());
+ upxref.setMap(newMap);
+ if (upx.size() == 0)
+ {
+ // add the new uniprot ref
+ querySeq.getDatasetSequence().addDBRef(upxref);
+ }
+
+ }
+ }
+
+ /*
+ * 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)
- {
- /*
- * process a CDS feature (or a sub-type of CDS)
- */
- if (so.isA(sf.getType(), SequenceOntology.CDS))
- {
- 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;
- }
- }
- int proteinLength = proteinSeq.getLength();
- List<int[]> proteinRange = new ArrayList<int[]>();
- int proteinStart = 1;
- if (fivePrimeIncomplete && proteinSeq.getCharAt(0) == 'X')
+ EnsemblXref xrefFetcher = new EnsemblXref(getDomain(), getDbSource(),
+ getEnsemblDataVersion());
+ List<DBRefEntry> xrefs = xrefFetcher.getCrossReferences(seq.getName());
+ for (DBRefEntry xref : xrefs)
{
- proteinStart = 2;
- proteinLength--;
+ seq.addDBRef(xref);
}
- proteinRange.add(new int[] { proteinStart, proteinLength });
/*
- * dna length should map to protein (or protein plus stop codon)
+ * and add a reference to itself
*/
- int codesForResidues = mappedDnaLength / 3;
- if (codesForResidues == proteinLength
- || codesForResidues == (proteinLength + 1))
- {
- return new MapList(ranges, proteinRange, 3, 1);
- }
- return null;
+ DBRefEntry self = new DBRefEntry(getDbSource(),
+ getEnsemblDataVersion(), seq.getName());
+ seq.addDBRef(self);
}
/**
throw new JalviewException("ENSEMBL Rest API not available.");
}
FileParse fp = getSequenceReader(ids);
+ if (fp == null)
+ {
+ return alignment;
+ }
+
FastaFile fr = new FastaFile(fp);
if (fr.hasWarningMessage())
{
"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(
- fr.getSeqsAsArray());
- for (SequenceI sq:seqal.getSequences())
+ AlignmentI seqal = new Alignment(fr.getSeqsAsArray());
+ for (SequenceI sq : seqal.getSequences())
{
if (sq.getDescription() == null)
{
if (ids.contains(name)
|| ids.contains(name.replace("ENSP", "ENST")))
{
- DBRefUtils.parseToDbRef(sq, DBRefSource.ENSEMBL, "0", name);
+ DBRefEntry dbref = DBRefUtils.parseToDbRef(sq, getDbSource(),
+ getEnsemblDataVersion(), name);
+ sq.addDBRef(dbref);
}
}
if (alignment == null)
@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();
int mappedLength = 0;
int direction = 1; // forward
boolean directionSet = false;
-
+
for (SequenceFeature sf : sfs)
{
/*
*/
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;
- }
- direction = strand;
- directionSet = true;
-
- /*
- * add to CDS ranges, semi-sorted forwards/backwards
- */
- if (strand < 0)
- {
- regions.add(0, new int[] { sf.getEnd(), sf.getBegin() });
- }
- else
- {
+ return null;
+ }
+ direction = strand;
+ directionSet = true;
+
+ /*
+ * add to CDS ranges, semi-sorted forwards/backwards
+ */
+ if (strand < 0)
+ {
+ regions.add(0, new int[] { sf.getEnd(), sf.getBegin() });
+ }
+ else
+ {
regions.add(new int[] { sf.getBegin(), sf.getEnd() });
}
mappedLength += Math.abs(sf.getEnd() - sf.getBegin() + 1);
}
}
}
-
+
if (regions.isEmpty())
{
System.out.println("Failed to identify target sequence for " + accId
* a final sort is needed since Ensembl returns CDS sorted within source
* (havana / ensembl_havana)
*/
- Collections.sort(regions, new RangeSorter(direction == 1));
-
- List<int[]> to = new ArrayList<int[]>();
- to.add(new int[] { start, start + mappedLength - 1 });
-
+ Collections.sort(regions, new RangeComparator(direction == 1));
+
+ List<int[]> to = Arrays.asList(new int[] { start,
+ start + mappedLength - 1 });
+
return new MapList(regions, to, 1, 1);
}
* @param mapping
* mapping from the sequence feature's coordinates to the target
* sequence
+ * @param forwardStrand
*/
protected void transferFeature(SequenceFeature sf,
- SequenceI targetSequence, MapList mapping)
+ SequenceI targetSequence, MapList mapping, boolean forwardStrand)
{
int start = sf.getBegin();
int end = sf.getEnd();
int[] mappedRange = mapping.locateInTo(start, end);
-
+
if (mappedRange != null)
{
SequenceFeature copy = new SequenceFeature(sf);
copy.setBegin(Math.min(mappedRange[0], mappedRange[1]));
copy.setEnd(Math.max(mappedRange[0], mappedRange[1]));
+ if (".".equals(copy.getFeatureGroup()))
+ {
+ copy.setFeatureGroup(getDbSource());
+ }
targetSequence.addSequenceFeature(copy);
/*
- * for sequence_variant, make an additional feature with consequence
+ * for sequence_variant on reverse strand, have to convert the allele
+ * values to their complements
*/
- if (SequenceOntology.getInstance().isSequenceVariant(sf.getType()))
+ if (!forwardStrand
+ && 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);
- }
+ reverseComplementAlleles(copy);
+ }
+ }
+ }
+
+ /**
+ * Change the 'alleles' value of a feature by converting to complementary
+ * bases, and also update the feature description to match
+ *
+ * @param sf
+ */
+ static void reverseComplementAlleles(SequenceFeature sf)
+ {
+ final String alleles = (String) sf.getValue(ALLELES);
+ if (alleles == null)
+ {
+ return;
+ }
+ StringBuilder complement = new StringBuilder(alleles.length());
+ for (String allele : alleles.split(","))
+ {
+ reverseComplementAllele(complement, allele);
+ }
+ String comp = complement.toString();
+ sf.setValue(ALLELES, comp);
+ sf.setDescription(comp);
+
+ /*
+ * replace value of "alleles=" in sf.ATTRIBUTES as well
+ * so 'output as GFF' shows reverse complement alleles
+ */
+ String atts = sf.getAttributes();
+ if (atts != null)
+ {
+ atts = atts.replace(ALLELES + "=" + alleles, ALLELES + "=" + comp);
+ sf.setAttributes(atts);
+ }
+ }
+
+ /**
+ * Makes the 'reverse complement' of the given allele and appends it to the
+ * buffer, after a comma separator if not the first
+ *
+ * @param complement
+ * @param allele
+ */
+ static void reverseComplementAllele(StringBuilder complement,
+ String allele)
+ {
+ if (complement.length() > 0)
+ {
+ complement.append(",");
+ }
+
+ /*
+ * some 'alleles' are actually descriptive terms
+ * e.g. HGMD_MUTATION, PhenCode_variation
+ * - we don't want to 'reverse complement' these
+ */
+ if (!Comparison.isNucleotideSequence(allele, true))
+ {
+ complement.append(allele);
+ }
+ else
+ {
+ for (int i = allele.length() - 1; i >= 0; i--)
+ {
+ complement.append(Dna.getComplement(allele.charAt(i)));
}
}
}
return false;
}
+ // long start = System.currentTimeMillis();
SequenceFeature[] sfs = sourceSequence.getSequenceFeatures();
- MapList mapping = getGenomicRanges(sourceSequence, accessionId,
- targetSequence.getStart());
+ 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;
}
/**
final boolean forwardStrand = mapping.isFromForwardStrand();
/*
- * sort features by start position (descending if reverse strand)
- * before transferring (in forwards order) to the target sequence
+ * sort features by start position (which corresponds to end
+ * position descending if reverse strand) so as to add them in
+ * 'forwards' order to the target sequence
*/
- Arrays.sort(features, new Comparator<SequenceFeature>()
- {
- @Override
- public int compare(SequenceFeature o1, SequenceFeature o2)
- {
- int c = Integer.compare(o1.getBegin(), o2.getBegin());
- return forwardStrand ? c : -c;
- }
- });
+ sortFeatures(features, forwardStrand);
boolean transferred = false;
for (SequenceFeature sf : features)
{
if (retainFeature(sf, parentId))
{
- transferFeature(sf, targetSequence, mapping);
+ transferFeature(sf, targetSequence, mapping, forwardStrand);
transferred = true;
}
}
}
/**
+ * Sort features by start position ascending (if on forward strand), or end
+ * position descending (if on reverse strand)
+ *
+ * @param features
+ * @param forwardStrand
+ */
+ protected static void sortFeatures(SequenceFeature[] features,
+ final boolean forwardStrand)
+ {
+ Arrays.sort(features, new Comparator<SequenceFeature>()
+ {
+ @Override
+ public int compare(SequenceFeature o1, SequenceFeature o2)
+ {
+ if (forwardStrand)
+ {
+ return Integer.compare(o1.getBegin(), o2.getBegin());
+ }
+ else
+ {
+ return Integer.compare(o2.getEnd(), o1.getEnd());
+ }
+ }
+ });
+ }
+
+ /**
* Answers true if the feature type is one we want to keep for the sequence.
* Some features are only retrieved in order to identify the sequence range,
* and may then be discarded as redundant information (e.g. "CDS" feature for
+ " 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
String type, String parentId)
{
List<SequenceFeature> result = new ArrayList<SequenceFeature>();
-
+
SequenceFeature[] sfs = sequence.getSequenceFeatures();
- if (sfs != null) {
- SequenceOntology so = SequenceOntology.getInstance();
- for (SequenceFeature sf :sfs) {
+ if (sfs != null)
+ {
+ SequenceOntologyI so = SequenceOntologyFactory.getInstance();
+ for (SequenceFeature sf : sfs)
+ {
if (so.isA(sf.getType(), type))
{
String parent = (String) sf.getValue(PARENT);
*/
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);
}
}