2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
21 package jalview.ext.ensembl;
23 import jalview.analysis.AlignmentUtils;
24 import jalview.analysis.Dna;
25 import jalview.bin.Cache;
26 import jalview.datamodel.Alignment;
27 import jalview.datamodel.AlignmentI;
28 import jalview.datamodel.DBRefEntry;
29 import jalview.datamodel.DBRefSource;
30 import jalview.datamodel.Mapping;
31 import jalview.datamodel.Sequence;
32 import jalview.datamodel.SequenceFeature;
33 import jalview.datamodel.SequenceI;
34 import jalview.datamodel.features.SequenceFeatures;
35 import jalview.exceptions.JalviewException;
36 import jalview.io.gff.Gff3Helper;
37 import jalview.io.gff.SequenceOntologyFactory;
38 import jalview.io.gff.SequenceOntologyI;
39 import jalview.util.Comparison;
40 import jalview.util.DBRefUtils;
41 import jalview.util.IntRangeComparator;
42 import jalview.util.MapList;
44 import java.io.BufferedReader;
45 import java.io.IOException;
46 import java.net.MalformedURLException;
48 import java.util.ArrayList;
49 import java.util.Arrays;
50 import java.util.Collections;
51 import java.util.List;
53 import org.json.simple.JSONObject;
54 import org.json.simple.parser.JSONParser;
55 import org.json.simple.parser.ParseException;
58 * Base class for Ensembl sequence fetchers
60 * @see http://rest.ensembl.org/documentation/info/sequence_id
63 public abstract class EnsemblSeqProxy extends EnsemblRestClient
65 protected static final String DESCRIPTION = "description";
68 * enum for 'type' parameter to the /sequence REST service
70 public enum EnsemblSeqType
73 * type=genomic to fetch full dna including introns
78 * type=cdna to fetch coding dna including UTRs
83 * type=cds to fetch coding dna excluding UTRs
88 * type=protein to fetch peptide product sequence
93 * the value of the 'type' parameter to fetch this version of
98 EnsemblSeqType(String t)
103 public String getType()
111 * Default constructor (to use rest.ensembl.org)
113 public EnsemblSeqProxy()
119 * Constructor given the target domain to fetch data from
121 public EnsemblSeqProxy(String d)
127 * Makes the sequence queries to Ensembl's REST service and returns an
128 * alignment consisting of the returned sequences.
131 public AlignmentI getSequenceRecords(String query) throws Exception
133 // TODO use a String... query vararg instead?
135 // danger: accession separator used as a regex here, a string elsewhere
136 // in this case it is ok (it is just a space), but (e.g.) '\' would not be
137 List<String> allIds = Arrays
138 .asList(query.split(getAccessionSeparator()));
139 AlignmentI alignment = null;
143 * execute queries, if necessary in batches of the
144 * maximum allowed number of ids
146 int maxQueryCount = getMaximumQueryCount();
147 for (int v = 0, vSize = allIds.size(); v < vSize; v += maxQueryCount)
149 int p = Math.min(vSize, v + maxQueryCount);
150 List<String> ids = allIds.subList(v, p);
153 alignment = fetchSequences(ids, alignment);
154 } catch (Throwable r)
157 String msg = "Aborting ID retrieval after " + v
158 + " chunks. Unexpected problem (" + r.getLocalizedMessage()
160 System.err.println(msg);
166 if (alignment == null)
172 * fetch and transfer genomic sequence features,
173 * fetch protein product and add as cross-reference
175 for (String accId : allIds)
177 addFeaturesAndProduct(accId, alignment);
180 for (SequenceI seq : alignment.getSequences())
182 getCrossReferences(seq);
189 * Fetches Ensembl features using the /overlap REST endpoint, and adds them to
190 * the sequence in the alignment. Also fetches the protein product, maps it
191 * from the CDS features of the sequence, and saves it as a cross-reference of
197 protected void addFeaturesAndProduct(String accId, AlignmentI alignment)
199 if (alignment == null)
207 * get 'dummy' genomic sequence with gene, transcript,
208 * exon, cds and variation features
210 SequenceI genomicSequence = null;
211 EnsemblFeatures gffFetcher = new EnsemblFeatures(getDomain());
212 EnsemblFeatureType[] features = getFeaturesToFetch();
213 AlignmentI geneFeatures = gffFetcher.getSequenceRecords(accId,
215 if (geneFeatures != null && geneFeatures.getHeight() > 0)
217 genomicSequence = geneFeatures.getSequenceAt(0);
219 if (genomicSequence != null)
222 * transfer features to the query sequence
224 SequenceI querySeq = alignment.findName(accId, true);
225 if (transferFeatures(accId, genomicSequence, querySeq))
229 * fetch and map protein product, and add it as a cross-reference
230 * of the retrieved sequence
232 addProteinProduct(querySeq);
235 } catch (IOException e)
238 "Error transferring Ensembl features: " + e.getMessage());
243 * Returns those sequence feature types to fetch from Ensembl. We may want
244 * features either because they are of interest to the user, or as means to
245 * identify the locations of the sequence on the genomic sequence (CDS
246 * features identify CDS, exon features identify cDNA etc).
250 protected abstract EnsemblFeatureType[] getFeaturesToFetch();
253 * Fetches and maps the protein product, and adds it as a cross-reference of
254 * the retrieved sequence
256 protected void addProteinProduct(SequenceI querySeq)
258 String accId = querySeq.getName();
261 AlignmentI protein = new EnsemblProtein(getDomain())
262 .getSequenceRecords(accId);
263 if (protein == null || protein.getHeight() == 0)
265 System.out.println("No protein product found for " + accId);
268 SequenceI proteinSeq = protein.getSequenceAt(0);
271 * need dataset sequences (to be the subject of mappings)
273 proteinSeq.createDatasetSequence();
274 querySeq.createDatasetSequence();
276 MapList mapList = AlignmentUtils.mapCdsToProtein(querySeq,
280 // clunky: ensure Uniprot xref if we have one is on mapped sequence
281 SequenceI ds = proteinSeq.getDatasetSequence();
282 // TODO: Verify ensp primary ref is on proteinSeq.getDatasetSequence()
283 Mapping map = new Mapping(ds, mapList);
284 DBRefEntry dbr = new DBRefEntry(getDbSource(),
285 getEnsemblDataVersion(), proteinSeq.getName(), map);
286 querySeq.getDatasetSequence().addDBRef(dbr);
287 DBRefEntry[] uprots = DBRefUtils.selectRefs(ds.getDBRefs(),
289 { DBRefSource.UNIPROT });
290 DBRefEntry[] upxrefs = DBRefUtils.selectRefs(querySeq.getDBRefs(),
292 { DBRefSource.UNIPROT });
295 for (DBRefEntry up : uprots)
297 // locate local uniprot ref and map
298 List<DBRefEntry> upx = DBRefUtils.searchRefs(upxrefs,
299 up.getAccessionId());
308 "Implementation issue - multiple uniprot acc on product sequence.");
313 upxref = new DBRefEntry(DBRefSource.UNIPROT,
314 getEnsemblDataVersion(), up.getAccessionId());
317 Mapping newMap = new Mapping(ds, mapList);
318 upxref.setVersion(getEnsemblDataVersion());
319 upxref.setMap(newMap);
322 // add the new uniprot ref
323 querySeq.getDatasetSequence().addDBRef(upxref);
330 * copy exon features to protein, compute peptide variants from dna
331 * variants and add as features on the protein sequence ta-da
333 // JAL-3187 render on the fly instead
334 // AlignmentUtils.computeProteinFeatures(querySeq, proteinSeq, mapList);
336 } catch (Exception e)
339 .println(String.format("Error retrieving protein for %s: %s",
340 accId, e.getMessage()));
345 * Get database xrefs from Ensembl, and attach them to the sequence
349 protected void getCrossReferences(SequenceI seq)
351 while (seq.getDatasetSequence() != null)
353 seq = seq.getDatasetSequence();
356 EnsemblXref xrefFetcher = new EnsemblXref(getDomain(), getDbSource(),
357 getEnsemblDataVersion());
358 List<DBRefEntry> xrefs = xrefFetcher.getCrossReferences(seq.getName());
359 for (DBRefEntry xref : xrefs)
365 * and add a reference to itself
367 DBRefEntry self = new DBRefEntry(getDbSource(), getEnsemblDataVersion(),
373 * Fetches sequences for the list of accession ids and adds them to the
374 * alignment. Returns the extended (or created) alignment.
379 * @throws JalviewException
380 * @throws IOException
382 protected AlignmentI fetchSequences(List<String> ids,
383 AlignmentI alignment) throws JalviewException, IOException
385 if (!isEnsemblAvailable())
388 throw new JalviewException("ENSEMBL Rest API not available.");
390 BufferedReader br = getSequenceReader(ids);
396 List<SequenceI> seqs = parseSequenceJson(br);
400 throw new IOException("No data returned for " + ids);
403 if (seqs.size() != ids.size())
405 System.out.println(String.format(
406 "Only retrieved %d sequences for %d query strings",
407 seqs.size(), ids.size()));
412 AlignmentI seqal = new Alignment(
413 seqs.toArray(new SequenceI[seqs.size()]));
414 for (SequenceI seq : seqs)
416 if (seq.getDescription() == null)
418 seq.setDescription(getDbName());
420 String name = seq.getName();
421 if (ids.contains(name)
422 || ids.contains(name.replace("ENSP", "ENST")))
424 // TODO JAL-3077 use true accession version in dbref
425 DBRefEntry dbref = DBRefUtils.parseToDbRef(seq, getDbSource(),
426 getEnsemblDataVersion(), name);
430 if (alignment == null)
436 alignment.append(seqal);
443 * Parses a JSON response for a single sequence ID query
446 * @return a single jalview.datamodel.Sequence
447 * @see http://rest.ensembl.org/documentation/info/sequence_id
449 protected List<SequenceI> parseSequenceJson(BufferedReader br)
451 JSONParser jp = new JSONParser();
452 List<SequenceI> result = new ArrayList<>();
456 * for now, assumes only one sequence returned; refactor if needed
457 * in future to handle a JSONArray with more than one
459 final JSONObject val = (JSONObject) jp.parse(br);
460 Object s = val.get("desc");
461 String desc = s == null ? null : s.toString();
463 String id = s == null ? null : s.toString();
465 String seq = s == null ? null : s.toString();
466 Sequence sequence = new Sequence(id, seq);
469 sequence.setDescription(desc);
471 // todo JAL-3077 make a DBRefEntry with true accession version
472 // s = val.get("version");
473 // String version = s == null ? "0" : s.toString();
474 // DBRefEntry dbref = new DBRefEntry(getDbSource(), version, id);
475 // sequence.addDBRef(dbref);
476 result.add(sequence);
477 } catch (ParseException | IOException e)
479 System.err.println("Error processing JSON response: " + e.toString());
486 * Returns the URL for the REST call
489 * @throws MalformedURLException
492 protected URL getUrl(List<String> ids) throws MalformedURLException
495 * a single id is included in the URL path
496 * multiple ids go in the POST body instead
498 StringBuffer urlstring = new StringBuffer(128);
499 urlstring.append(getDomain() + "/sequence/id");
502 urlstring.append("/").append(ids.get(0));
504 // @see https://github.com/Ensembl/ensembl-rest/wiki/Output-formats
505 urlstring.append("?type=").append(getSourceEnsemblType().getType());
506 urlstring.append(("&Accept=application/json"));
507 urlstring.append(("&Content-Type=application/json"));
509 String objectType = getObjectType();
510 if (objectType != null)
512 urlstring.append("&").append(OBJECT_TYPE).append("=")
516 URL url = new URL(urlstring.toString());
521 * Override this method to specify object_type request parameter
525 protected String getObjectType()
531 * A sequence/id POST request currently allows up to 50 queries
533 * @see http://rest.ensembl.org/documentation/info/sequence_id_post
536 public int getMaximumQueryCount()
542 protected boolean useGetRequest()
549 * @return the configured sequence return type for this source
551 protected abstract EnsemblSeqType getSourceEnsemblType();
554 * Returns a list of [start, end] genomic ranges corresponding to the sequence
557 * The correspondence between the frames of reference is made by locating
558 * those features on the genomic sequence which identify the retrieved
559 * sequence. Specifically
561 * <li>genomic sequence is identified by "transcript" features with
562 * ID=transcript:transcriptId</li>
563 * <li>cdna sequence is identified by "exon" features with
564 * Parent=transcript:transcriptId</li>
565 * <li>cds sequence is identified by "CDS" features with
566 * Parent=transcript:transcriptId</li>
569 * The returned ranges are sorted to run forwards (for positive strand) or
570 * backwards (for negative strand). Aborts and returns null if both positive
571 * and negative strand are found (this should not normally happen).
573 * @param sourceSequence
576 * the start position of the sequence we are mapping to
579 protected MapList getGenomicRangesFromFeatures(SequenceI sourceSequence,
580 String accId, int start)
582 List<SequenceFeature> sfs = getIdentifyingFeatures(sourceSequence,
590 * generously initial size for number of cds regions
591 * (worst case titin Q8WZ42 has c. 313 exons)
593 List<int[]> regions = new ArrayList<>(100);
594 int mappedLength = 0;
595 int direction = 1; // forward
596 boolean directionSet = false;
598 for (SequenceFeature sf : sfs)
600 int strand = sf.getStrand();
601 strand = strand == 0 ? 1 : strand; // treat unknown as forward
603 if (directionSet && strand != direction)
605 // abort - mix of forward and backward
607 .println("Error: forward and backward strand for " + accId);
614 * add to CDS ranges, semi-sorted forwards/backwards
618 regions.add(0, new int[] { sf.getEnd(), sf.getBegin() });
622 regions.add(new int[] { sf.getBegin(), sf.getEnd() });
624 mappedLength += Math.abs(sf.getEnd() - sf.getBegin() + 1);
627 if (regions.isEmpty())
629 System.out.println("Failed to identify target sequence for " + accId
630 + " from genomic features");
635 * a final sort is needed since Ensembl returns CDS sorted within source
636 * (havana / ensembl_havana)
638 Collections.sort(regions, direction == 1 ? IntRangeComparator.ASCENDING
639 : IntRangeComparator.DESCENDING);
641 List<int[]> to = Arrays
643 { start, start + mappedLength - 1 });
645 return new MapList(regions, to, 1, 1);
649 * Answers a list of sequence features that mark positions of the genomic
650 * sequence feature which are within the sequence being retrieved. For
651 * example, an 'exon' feature whose parent is the target transcript marks the
652 * cdna positions of the transcript. For a gene sequence, this is trivially
653 * just the 'gene' feature with matching gene id.
659 protected abstract List<SequenceFeature> getIdentifyingFeatures(
660 SequenceI seq, String accId);
663 * Transfers the sequence feature to the target sequence, locating its start
664 * and end range based on the mapping. Features which do not overlap the
665 * target sequence are ignored.
668 * @param targetSequence
670 * mapping from the sequence feature's coordinates to the target
672 * @param forwardStrand
674 protected void transferFeature(SequenceFeature sf,
675 SequenceI targetSequence, MapList mapping, boolean forwardStrand)
677 int start = sf.getBegin();
678 int end = sf.getEnd();
679 int[] mappedRange = mapping.locateInTo(start, end);
681 if (mappedRange != null)
683 String group = sf.getFeatureGroup();
684 if (".".equals(group))
686 group = getDbSource();
688 int newBegin = Math.min(mappedRange[0], mappedRange[1]);
689 int newEnd = Math.max(mappedRange[0], mappedRange[1]);
690 SequenceFeature copy = new SequenceFeature(sf, newBegin, newEnd,
691 group, sf.getScore());
692 targetSequence.addSequenceFeature(copy);
695 * for sequence_variant on reverse strand, have to convert the allele
696 * values to their complements
698 if (!forwardStrand && SequenceOntologyFactory.getInstance()
699 .isA(sf.getType(), SequenceOntologyI.SEQUENCE_VARIANT))
701 reverseComplementAlleles(copy);
707 * Change the 'alleles' value of a feature by converting to complementary
708 * bases, and also update the feature description to match
712 static void reverseComplementAlleles(SequenceFeature sf)
714 final String alleles = (String) sf.getValue(Gff3Helper.ALLELES);
719 StringBuilder complement = new StringBuilder(alleles.length());
720 for (String allele : alleles.split(","))
722 reverseComplementAllele(complement, allele);
724 String comp = complement.toString();
725 sf.setValue(Gff3Helper.ALLELES, comp);
726 sf.setDescription(comp);
730 * Makes the 'reverse complement' of the given allele and appends it to the
731 * buffer, after a comma separator if not the first
736 static void reverseComplementAllele(StringBuilder complement,
739 if (complement.length() > 0)
741 complement.append(",");
745 * some 'alleles' are actually descriptive terms
746 * e.g. HGMD_MUTATION, PhenCode_variation
747 * - we don't want to 'reverse complement' these
749 if (!Comparison.isNucleotideSequence(allele, true))
751 complement.append(allele);
755 for (int i = allele.length() - 1; i >= 0; i--)
757 complement.append(Dna.getComplement(allele.charAt(i)));
763 * Transfers features from sourceSequence to targetSequence
766 * @param sourceSequence
767 * @param targetSequence
768 * @return true if any features were transferred, else false
770 protected boolean transferFeatures(String accessionId,
771 SequenceI sourceSequence, SequenceI targetSequence)
773 if (sourceSequence == null || targetSequence == null)
778 // long start = System.currentTimeMillis();
779 List<SequenceFeature> sfs = sourceSequence.getFeatures()
780 .getPositionalFeatures();
781 MapList mapping = getGenomicRangesFromFeatures(sourceSequence,
782 accessionId, targetSequence.getStart());
788 boolean result = transferFeatures(sfs, targetSequence, mapping,
790 // System.out.println("transferFeatures (" + (sfs.size()) + " --> "
791 // + targetSequence.getFeatures().getFeatureCount(true) + ") to "
792 // + targetSequence.getName() + " took "
793 // + (System.currentTimeMillis() - start) + "ms");
798 * Transfer features to the target sequence. The start/end positions are
799 * converted using the mapping. Features which do not overlap are ignored.
800 * Features whose parent is not the specified identifier are also ignored.
803 * @param targetSequence
808 protected boolean transferFeatures(List<SequenceFeature> sfs,
809 SequenceI targetSequence, MapList mapping, String parentId)
811 final boolean forwardStrand = mapping.isFromForwardStrand();
814 * sort features by start position (which corresponds to end
815 * position descending if reverse strand) so as to add them in
816 * 'forwards' order to the target sequence
818 SequenceFeatures.sortFeatures(sfs, forwardStrand);
820 boolean transferred = false;
821 for (SequenceFeature sf : sfs)
823 if (retainFeature(sf, parentId))
825 transferFeature(sf, targetSequence, mapping, forwardStrand);
833 * Answers true if the feature type is one we want to keep for the sequence.
834 * Some features are only retrieved in order to identify the sequence range,
835 * and may then be discarded as redundant information (e.g. "CDS" feature for
838 @SuppressWarnings("unused")
839 protected boolean retainFeature(SequenceFeature sf, String accessionId)
841 return true; // override as required
845 * Answers true if the feature has a Parent which refers to the given
846 * accession id, or if the feature has no parent. Answers false if the
847 * feature's Parent is for a different accession id.
853 protected boolean featureMayBelong(SequenceFeature sf, String identifier)
855 String parent = (String) sf.getValue(PARENT);
857 && !parent.equalsIgnoreCase(identifier))
859 // this genomic feature belongs to a different transcript
866 * Answers a short description of the sequence fetcher
869 public String getDescription()
871 return "Ensembl " + getSourceEnsemblType().getType()
872 + " sequence with variant features";
876 * Returns a (possibly empty) list of features on the sequence which have the
877 * specified sequence ontology term (or a sub-type of it), and the given
878 * identifier as parent
885 protected List<SequenceFeature> findFeatures(SequenceI sequence,
886 String term, String parentId)
888 List<SequenceFeature> result = new ArrayList<>();
890 List<SequenceFeature> sfs = sequence.getFeatures()
891 .getFeaturesByOntology(term);
892 for (SequenceFeature sf : sfs)
894 String parent = (String) sf.getValue(PARENT);
895 if (parent != null && parent.equalsIgnoreCase(parentId))
905 * Answers true if the feature type is either 'NMD_transcript_variant' or
906 * 'transcript' (or one of its sub-types in the Sequence Ontology). This is
907 * because NMD_transcript_variant behaves like 'transcript' in Ensembl
908 * although strictly speaking it is not (it is a sub-type of
911 * (This test was needed when fetching transcript features as GFF. As we are
912 * now fetching as JSON, all features have type 'transcript' so the check for
913 * NMD_transcript_variant is redundant. Left in for any future case arising.)
918 public static boolean isTranscript(String featureType)
920 return SequenceOntologyI.NMD_TRANSCRIPT_VARIANT.equals(featureType)
921 || SequenceOntologyFactory.getInstance().isA(featureType,
922 SequenceOntologyI.TRANSCRIPT);