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.
23 import jalview.analysis.AlignSeq;
24 import jalview.api.FeatureSettingsModelI;
25 import jalview.bin.Cache;
26 import jalview.datamodel.AlignmentI;
27 import jalview.datamodel.DBRefEntry;
28 import jalview.datamodel.DBRefSource;
29 import jalview.datamodel.Mapping;
30 import jalview.datamodel.SequenceI;
31 import jalview.gui.CutAndPasteTransfer;
32 import jalview.gui.Desktop;
33 import jalview.gui.FeatureSettings;
34 import jalview.gui.IProgressIndicator;
35 import jalview.gui.OOMWarning;
36 import jalview.gui.Preferences;
37 import jalview.util.DBRefUtils;
38 import jalview.util.MessageManager;
39 import jalview.ws.seqfetcher.DbSourceProxy;
41 import java.util.ArrayList;
42 import java.util.Arrays;
43 import java.util.Enumeration;
44 import java.util.HashMap;
45 import java.util.Hashtable;
46 import java.util.List;
48 import java.util.StringTokenizer;
49 import java.util.Vector;
51 import uk.ac.ebi.picr.model.UPEntry;
52 import uk.ac.ebi.www.picr.AccessionMappingService.AccessionMapperServiceLocator;
55 * Implements a runnable for validating a sequence against external databases
56 * and then propagating references and features onto the sequence(s)
61 public class DBRefFetcher implements Runnable
63 private static final String NEWLINE = System.lineSeparator();
65 public static final String TRIM_RETRIEVED_SEQUENCES = "TRIM_FETCHED_DATASET_SEQS";
67 public interface FetchFinishedListenerI
74 IProgressIndicator progressWindow;
76 CutAndPasteTransfer output = new CutAndPasteTransfer();
78 boolean running = false;
81 * picr client instance
83 uk.ac.ebi.www.picr.AccessionMappingService.AccessionMapperInterface picrClient = null;
85 // This will be a collection of Vectors of sequenceI refs.
86 // The key will be the seq name or accession id of the seq
87 Hashtable<String, Vector<SequenceI>> seqRefs;
89 DbSourceProxy[] dbSources;
91 SequenceFetcher sfetcher;
93 private List<FetchFinishedListenerI> listeners;
95 private SequenceI[] alseqs;
98 * when true - retrieved sequences will be trimmed to cover longest derived
101 private boolean trimDsSeqs = true;
104 * Creates a new DBRefFetcher object and fetches from the currently selected
105 * set of databases, if this is null then it fetches based on feature settings
108 * fetch references for these SequenceI array
109 * @param progressIndicatorFrame
110 * the frame for progress bar monitoring
112 * array of DbSourceProxy to query references form
113 * @param featureSettings
114 * FeatureSettings to get alternative DbSourceProxy from
115 * @param isNucleotide
116 * indicates if the array of SequenceI are Nucleotides or not
118 public DBRefFetcher(SequenceI[] seqs,
119 IProgressIndicator progressIndicatorFrame,
120 DbSourceProxy[] sources, FeatureSettings featureSettings,
121 boolean isNucleotide)
123 listeners = new ArrayList<>();
124 this.progressWindow = progressIndicatorFrame;
125 alseqs = new SequenceI[seqs.length];
126 SequenceI[] ds = new SequenceI[seqs.length];
127 for (int i = 0; i < seqs.length; i++)
130 if (seqs[i].getDatasetSequence() != null)
132 ds[i] = seqs[i].getDatasetSequence();
140 // TODO Jalview 2.5 lots of this code should be in the gui package!
141 sfetcher = SequenceFetcher.getInstance();
142 // set default behaviour for transferring excess sequence data to the
144 trimDsSeqs = Cache.getDefault(TRIM_RETRIEVED_SEQUENCES, true);
147 setDatabaseSources(featureSettings, isNucleotide);
151 // we assume the caller knows what they're doing and ensured that all the
152 // db source names are valid
158 * Helper method to configure the list of database sources to query
160 * @param featureSettings
161 * @param forNucleotide
163 void setDatabaseSources(FeatureSettings featureSettings,
164 boolean forNucleotide)
166 // af.featureSettings_actionPerformed(null);
167 String[] defdb = null;
168 List<DbSourceProxy> selsources = new ArrayList<>();
169 // select appropriate databases based on alignFrame context.
172 defdb = DBRefSource.DNACODINGDBS;
176 defdb = DBRefSource.PROTEINDBS;
178 List<DbSourceProxy> srces = new ArrayList<>();
179 for (String ddb : defdb)
181 List<DbSourceProxy> srcesfordb = sfetcher.getSourceProxy(ddb);
182 if (srcesfordb != null)
184 for (DbSourceProxy src : srcesfordb)
186 if (!srces.contains(src))
188 srces.addAll(srcesfordb);
193 // append the PDB data source, since it is 'special', catering for both
194 // nucleotide and protein
195 // srces.addAll(sfetcher.getSourceProxy(DBRefSource.PDB));
197 srces.addAll(selsources);
198 dbSources = srces.toArray(new DbSourceProxy[srces.size()]);
202 * Constructor with only sequences provided
206 public DBRefFetcher(SequenceI[] sequences)
208 this(sequences, null, null, null, false);
212 * Add a listener to be notified when sequence fetching is complete
216 public void addListener(FetchFinishedListenerI l)
222 * start the fetcher thread
224 * @param waitTillFinished
225 * true to block until the fetcher has finished
227 public void fetchDBRefs(boolean waitTillFinished)
229 // TODO can we not simply write
230 // if (waitTillFinished) { run(); } else { new Thread(this).start(); }
232 Thread thread = new Thread(this);
236 if (waitTillFinished)
243 } catch (Exception ex)
251 * The sequence will be added to a vector of sequences belonging to key which
252 * could be either seq name or dbref id
259 void addSeqId(SequenceI seq, String key)
261 key = key.toUpperCase();
263 Vector<SequenceI> seqs;
264 if (seqRefs.containsKey(key))
266 seqs = seqRefs.get(key);
268 if (seqs != null && !seqs.contains(seq))
270 seqs.addElement(seq);
272 else if (seqs == null)
274 seqs = new Vector<>();
275 seqs.addElement(seq);
281 seqs = new Vector<>();
282 seqs.addElement(seq);
285 seqRefs.put(key, seqs);
294 if (dbSources == null)
296 throw new Error(MessageManager
297 .getString("error.implementation_error_must_init_dbsources"));
300 long startTime = System.currentTimeMillis();
301 if (progressWindow != null)
303 progressWindow.setProgressBar(
304 MessageManager.getString("status.fetching_db_refs"),
309 if (Cache.getDefault(Preferences.DBREFFETCH_USEPICR, false))
311 picrClient = new AccessionMapperServiceLocator()
312 .getAccessionMapperPort();
314 } catch (Exception e)
316 System.err.println("Couldn't locate PICR service instance.\n");
320 Vector<SequenceI> sdataset = new Vector<>(
321 Arrays.asList(dataset));
322 List<String> warningMessages = new ArrayList<>();
324 // clear any old feature display settings recorded from past sessions
325 featureDisplaySettings = null;
328 while (sdataset.size() > 0 && db < dbSources.length)
330 int maxqlen = 1; // default number of queries made at one time
331 System.out.println("Verifying against " + dbSources[db].getDbName());
333 // iterate through db for each remaining un-verified sequence
334 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
335 sdataset.copyInto(currSeqs);// seqs that are to be validated against
337 Vector<String> queries = new Vector<>(); // generated queries curSeq
338 seqRefs = new Hashtable<>();
342 DbSourceProxy dbsource = dbSources[db];
343 // for moment, we dumbly iterate over all retrieval sources for a
344 // particular database
345 // TODO: introduce multithread multisource queries and logic to remove a
346 // query from other sources if any source for a database returns a
348 maxqlen = dbsource.getMaximumQueryCount();
350 while (queries.size() > 0 || seqIndex < currSeqs.length)
352 if (queries.size() > 0)
354 // Still queries to make for current seqIndex
355 StringBuffer queryString = new StringBuffer("");
357 int nqSize = (maxqlen > queries.size()) ? queries.size()
360 while (queries.size() > 0 && numq < nqSize)
362 String query = queries.elementAt(0);
363 if (dbsource.isValidReference(query))
366 (numq == 0) ? "" : dbsource.getAccessionSeparator());
367 queryString.append(query);
370 // remove the extracted query string
371 queries.removeElementAt(0);
373 // make the queries and process the response
374 AlignmentI retrieved = null;
377 if (Cache.log.isDebugEnabled())
379 Cache.log.debug("Querying " + dbsource.getDbName()
380 + " with : '" + queryString.toString() + "'");
382 retrieved = dbsource.getSequenceRecords(queryString.toString());
383 } catch (Exception ex)
385 ex.printStackTrace();
386 } catch (OutOfMemoryError err)
388 new OOMWarning("retrieving database references ("
389 + queryString.toString() + ")", err);
391 if (retrieved != null)
393 transferReferences(sdataset, dbsource, retrieved,
394 trimDsSeqs, warningMessages);
399 // make some more strings for use as queries
400 for (int i = 0; (seqIndex < dataset.length)
401 && (i < 50); seqIndex++, i++)
403 SequenceI sequence = dataset[seqIndex];
404 List<DBRefEntry> uprefs = DBRefUtils
405 .selectRefs(sequence.getDBRefs(), new String[]
406 { dbsource.getDbSource() }); // jalview.datamodel.DBRefSource.UNIPROT
408 // check for existing dbrefs to use
409 if (uprefs != null && uprefs.size() > 0)
411 for (int j = 0, n = uprefs.size(); j < n; j++)
413 DBRefEntry upref = uprefs.get(j);
414 addSeqId(sequence, upref.getAccessionId());
416 upref.getAccessionId().toUpperCase());
421 // generate queries from sequence ID string
422 StringTokenizer st = new StringTokenizer(sequence.getName(),
424 while (st.hasMoreTokens())
426 String token = st.nextToken();
427 UPEntry[] presp = null;
428 if (picrClient != null)
430 // resolve the string against PICR to recover valid IDs
433 presp = picrClient.getUPIForAccession(token, null,
434 picrClient.getMappedDatabaseNames(), null,
436 } catch (Exception e)
439 "Exception with Picr for '" + token + "'\n");
443 if (presp != null && presp.length > 0)
445 for (int id = 0; id < presp.length; id++)
447 // construct sequences from response if sequences are
448 // present, and do a transferReferences
449 // otherwise transfer non sequence x-references directly.
452 "Validated ID against PICR... (for what its worth):"
454 addSeqId(sequence, token);
455 queries.addElement(token.toUpperCase());
460 // System.out.println("Not querying source with
461 // token="+token+"\n");
462 addSeqId(sequence, token);
463 queries.addElement(token.toUpperCase());
470 // advance to next database
472 } // all databases have been queried
473 if (!warningMessages.isEmpty())
475 StringBuilder sb = new StringBuilder(warningMessages.size() * 30);
476 sb.append(MessageManager
477 .getString("label.your_sequences_have_been_verified"));
478 for (String msg : warningMessages)
480 sb.append(msg).append(NEWLINE);
482 output.setText(sb.toString());
484 Desktop.addInternalFrame(output,
485 MessageManager.getString("label.sequences_updated"), 600,
487 // The above is the dataset, we must now find out the index
488 // of the viewed sequence
491 if (progressWindow != null)
493 progressWindow.setProgressBar(
494 MessageManager.getString("label.dbref_search_completed"),
498 for (FetchFinishedListenerI listener : listeners)
506 * Verify local sequences in seqRefs against the retrieved sequence database
507 * records. Returns true if any sequence was modified as a result (start/end
508 * changed and/or sequence enlarged), else false.
511 * dataset sequences we are retrieving for
513 * database source we are retrieving from
515 * retrieved sequences as alignment
516 * @param trimDatasetSeqs
517 * if true, sequences will not be enlarged to match longer retrieved
518 * sequences, only their start/end adjusted
519 * @param warningMessages
520 * a list of messages to add to
522 boolean transferReferences(Vector<SequenceI> sdataset,
523 DbSourceProxy dbSourceProxy,
524 AlignmentI retrievedAl, boolean trimDatasetSeqs,
525 List<String> warningMessages)
527 // System.out.println("trimming ? " + trimDatasetSeqs);
528 if (retrievedAl == null || retrievedAl.getHeight() == 0)
533 String dbSource = dbSourceProxy.getDbName();
534 boolean modified = false;
535 SequenceI[] retrieved = recoverDbSequences(
536 retrievedAl.getSequencesArray());
537 SequenceI sequence = null;
539 for (SequenceI retrievedSeq : retrieved)
541 // Work out which sequences this sequence matches,
542 // taking into account all accessionIds and names in the file
543 Vector<SequenceI> sequenceMatches = new Vector<>();
544 // look for corresponding accession ids
545 List<DBRefEntry> entryRefs = DBRefUtils
546 .selectRefs(retrievedSeq.getDBRefs(), new String[]
548 if (entryRefs == null)
551 .println("Dud dbSource string ? no entryrefs selected for "
552 + dbSource + " on " + retrievedSeq.getName());
555 for (int j = 0, n = entryRefs.size(); j < n; j++)
557 DBRefEntry ref = entryRefs.get(j);
558 String accessionId = ref.getAccessionId();
559 // match up on accessionId
560 if (seqRefs.containsKey(accessionId.toUpperCase()))
562 Vector<SequenceI> seqs = seqRefs.get(accessionId);
563 for (int jj = 0; jj < seqs.size(); jj++)
565 sequence = seqs.elementAt(jj);
566 if (!sequenceMatches.contains(sequence))
568 sequenceMatches.addElement(sequence);
573 if (sequenceMatches.isEmpty())
575 // failed to match directly on accessionId==query so just compare all
576 // sequences to entry
577 Enumeration<String> e = seqRefs.keys();
578 while (e.hasMoreElements())
580 Vector<SequenceI> sqs = seqRefs.get(e.nextElement());
581 if (sqs != null && sqs.size() > 0)
583 Enumeration<SequenceI> sqe = sqs.elements();
584 while (sqe.hasMoreElements())
586 sequenceMatches.addElement(sqe.nextElement());
591 // look for corresponding names
592 // this is uniprot specific ?
593 // could be useful to extend this so we try to find any 'significant'
594 // information in common between two sequence objects.
596 * List<DBRefEntry> entryRefs =
597 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
598 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
599 * name = entry.getName().elementAt(j).toString(); if
600 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
601 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
602 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
603 * sequenceMatches.addElement(sequence); } } } }
605 if (sequenceMatches.size() > 0)
607 addFeatureSettings(dbSourceProxy);
609 // sequenceMatches now contains the set of all sequences associated with
610 // the returned db record
611 final String retrievedSeqString = retrievedSeq.getSequenceAsString();
612 String entrySeq = retrievedSeqString.toUpperCase();
613 for (int m = 0; m < sequenceMatches.size(); m++)
615 sequence = sequenceMatches.elementAt(m);
616 // only update start and end positions and shift features if there are
617 // no existing references
618 // TODO: test for legacy where uniprot or EMBL refs exist but no
619 // mappings are made (but content matches retrieved set)
620 boolean updateRefFrame = sequence.getDBRefs() == null
621 || sequence.getDBRefs().size() == 0;
623 // verify sequence against the entry sequence
626 final int sequenceStart = sequence.getStart();
628 boolean remoteEnclosesLocal = false;
629 String nonGapped = AlignSeq
630 .extractGaps("-. ", sequence.getSequenceAsString())
632 int absStart = entrySeq.indexOf(nonGapped);
635 // couldn't find local sequence in sequence from database, so check if
636 // the database sequence is a subsequence of local sequence
637 absStart = nonGapped.indexOf(entrySeq);
640 // verification failed. couldn't find any relationship between
641 // entrySeq and local sequence
642 // messages suppressed as many-to-many matches are confusing
643 // String msg = sequence.getName()
644 // + " Sequence not 100% match with "
645 // + retrievedSeq.getName();
646 // addWarningMessage(warningMessages, msg);
650 * retrieved sequence is a proper subsequence of local sequence
652 String msg = sequence.getName() + " has " + absStart
653 + " prefixed residues compared to "
654 + retrievedSeq.getName();
655 addWarningMessage(warningMessages, msg);
658 * So create a mapping to the external entry from the matching region of
659 * the local sequence, and leave local start/end untouched.
661 mp = new Mapping(null,
663 { sequenceStart + absStart,
664 sequenceStart + absStart + entrySeq.length() - 1 },
666 { retrievedSeq.getStart(),
667 retrievedSeq.getStart() + entrySeq.length() - 1 },
669 updateRefFrame = false;
674 * local sequence is a subsequence of (or matches) retrieved sequence
676 remoteEnclosesLocal = true;
682 * relocate existing sequence features by offset
684 int startShift = absStart - sequenceStart + 1;
687 modified |= sequence.getFeatures().shiftFeatures(1,
693 System.out.println("Adding dbrefs to " + sequence.getName()
694 + " from " + dbSource + " sequence : "
695 + retrievedSeq.getName());
696 sequence.transferAnnotation(retrievedSeq, mp);
698 absStart += retrievedSeq.getStart();
699 int absEnd = absStart + nonGapped.length() - 1;
700 if (!trimDatasetSeqs)
703 * update start position and/or expand to longer retrieved sequence
705 if (!retrievedSeqString.equals(sequence.getSequenceAsString())
706 && remoteEnclosesLocal)
708 sequence.setSequence(retrievedSeqString);
710 addWarningMessage(warningMessages,
711 "Sequence for " + sequence.getName() + " expanded from "
712 + retrievedSeq.getName());
714 if (sequence.getStart() != retrievedSeq.getStart())
716 sequence.setStart(retrievedSeq.getStart());
718 if (absStart != sequenceStart)
720 addWarningMessage(warningMessages,
721 "Start/end position for " + sequence.getName()
722 + " updated from " + retrievedSeq.getName());
728 // finally, update local sequence reference frame if we're allowed
731 // just fix start/end
732 if (sequence.getStart() != absStart
733 || sequence.getEnd() != absEnd)
735 sequence.setStart(absStart);
736 sequence.setEnd(absEnd);
738 addWarningMessage(warningMessages,
739 "Start/end for " + sequence.getName()
740 + " updated from " + retrievedSeq.getName());
743 // search for alignment sequences to update coordinate frame for
744 for (int alsq = 0; alsq < alseqs.length; alsq++)
746 if (alseqs[alsq].getDatasetSequence() == sequence)
748 String ngAlsq = AlignSeq
750 alseqs[alsq].getSequenceAsString())
752 int oldstrt = alseqs[alsq].getStart();
753 alseqs[alsq].setStart(sequence.getSequenceAsString()
754 .toUpperCase().indexOf(ngAlsq) + sequence.getStart());
755 if (oldstrt != alseqs[alsq].getStart())
758 ngAlsq.length() + alseqs[alsq].getStart() - 1);
763 // TODO: search for all other references to this dataset sequence, and
765 // TODO: update all AlCodonMappings which involve this alignment
766 // sequence (e.g. Q30167 cdna translation from exon2 product (vamsas
769 // and remove it from the rest
770 // TODO: decide if we should remove annotated sequence from set
771 sdataset.remove(sequence);
777 Map<String, FeatureSettingsModelI> featureDisplaySettings = null;
779 private void addFeatureSettings(DbSourceProxy dbSourceProxy)
781 FeatureSettingsModelI fsettings = dbSourceProxy
782 .getFeatureColourScheme();
783 if (fsettings != null)
785 if (featureDisplaySettings == null)
787 featureDisplaySettings = new HashMap<>();
789 featureDisplaySettings.put(dbSourceProxy.getDbName(), fsettings);
795 * @return any feature settings associated with sources that have provided sequences
797 public List<FeatureSettingsModelI>getFeatureSettingsModels()
799 return featureDisplaySettings == null
800 ? Arrays.asList(new FeatureSettingsModelI[0])
801 : Arrays.asList(featureDisplaySettings.values()
802 .toArray(new FeatureSettingsModelI[1]));
805 * Adds the message to the list unless it already contains it
810 void addWarningMessage(List<String> messageList, String msg)
812 if (!messageList.contains(msg))
814 messageList.add(msg);
819 * loop thru and collect additional sequences in Map.
821 * @param sequencesArray
824 private SequenceI[] recoverDbSequences(SequenceI[] sequencesArray)
827 if (sequencesArray == null || (n = sequencesArray.length) == 0)
828 return sequencesArray;
829 ArrayList<SequenceI> nseq = new ArrayList<>();
830 for (int i = 0;i < n; i++)
832 nseq.add(sequencesArray[i]);
833 List<DBRefEntry> dbr = sequencesArray[i].getDBRefs();
837 for (int r = 0, rn = dbr.size(); r < rn; r++)
839 if ((map = dbr.get(r).getMap()) != null)
841 if (map.getTo() != null && !nseq.contains(map.getTo()))
843 nseq.add(map.getTo());
849 // BH 2019.01.25 question here if this is the right logic. Return the original if nothing found?
852 return nseq.toArray(new SequenceI[nseq.size()]);
854 return sequencesArray;