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 java.util.Locale;
25 import java.util.ArrayList;
26 import java.util.Arrays;
27 import java.util.Enumeration;
28 import java.util.HashMap;
29 import java.util.Hashtable;
30 import java.util.List;
32 import java.util.StringTokenizer;
33 import java.util.Vector;
34 import java.util.regex.Matcher;
35 import java.util.regex.Pattern;
37 import jalview.analysis.AlignSeq;
38 import jalview.api.FeatureSettingsModelI;
39 import jalview.bin.Cache;
40 import jalview.datamodel.AlignmentI;
41 import jalview.datamodel.DBRefEntry;
42 import jalview.datamodel.DBRefSource;
43 import jalview.datamodel.Mapping;
44 import jalview.datamodel.SequenceI;
45 import jalview.gui.CutAndPasteTransfer;
46 import jalview.gui.Desktop;
47 import jalview.gui.FeatureSettings;
48 import jalview.gui.IProgressIndicator;
49 import jalview.gui.OOMWarning;
50 import jalview.util.DBRefUtils;
51 import jalview.util.MessageManager;
52 import jalview.ws.seqfetcher.DbSourceProxy;
53 import uk.ac.ebi.picr.model.UPEntry;
54 import uk.ac.ebi.www.picr.AccessionMappingService.AccessionMapperServiceLocator;
57 * Implements a runnable for validating a sequence against external databases
58 * and then propagating references and features onto the sequence(s)
63 public class DBRefFetcher implements Runnable
65 private static final String NEWLINE = System.lineSeparator();
67 public static final String TRIM_RETRIEVED_SEQUENCES = "TRIM_FETCHED_DATASET_SEQS";
69 public interface FetchFinishedListenerI
76 IProgressIndicator progressWindow;
78 CutAndPasteTransfer output = new CutAndPasteTransfer();
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 = jalview.gui.SequenceFetcher.getSequenceFetcherSingleton();
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 if (waitTillFinished)
235 new Thread(this).start();
240 * The sequence will be added to a vector of sequences belonging to key which
241 * could be either seq name or dbref id
248 void addSeqId(SequenceI seq, String key)
250 key = key.toUpperCase(Locale.ROOT);
252 Vector<SequenceI> seqs;
253 if (seqRefs.containsKey(key))
255 seqs = seqRefs.get(key);
257 if (seqs != null && !seqs.contains(seq))
259 seqs.addElement(seq);
261 else if (seqs == null)
263 seqs = new Vector<>();
264 seqs.addElement(seq);
270 seqs = new Vector<>();
271 seqs.addElement(seq);
274 seqRefs.put(key, seqs);
283 if (dbSources == null)
285 throw new Error(MessageManager
286 .getString("error.implementation_error_must_init_dbsources"));
288 long startTime = System.currentTimeMillis();
289 if (progressWindow != null)
291 progressWindow.setProgressBar(
292 MessageManager.getString("status.fetching_db_refs"),
297 if (Cache.getDefault("DBREFFETCH_USEPICR", false))
299 picrClient = new AccessionMapperServiceLocator()
300 .getAccessionMapperPort();
302 } catch (Exception e)
304 System.err.println("Couldn't locate PICR service instance.\n");
308 Vector<SequenceI> sdataset = new Vector<>(
309 Arrays.asList(dataset));
310 List<String> warningMessages = new ArrayList<>();
312 // clear any old feature display settings recorded from past sessions
313 featureDisplaySettings = null;
316 while (sdataset.size() > 0 && db < dbSources.length)
318 int maxqlen = 1; // default number of queries made at one time
319 System.out.println("Verifying against " + dbSources[db].getDbName());
321 // iterate through db for each remaining un-verified sequence
322 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
323 sdataset.copyInto(currSeqs);// seqs that are to be validated against
325 Vector<String> queries = new Vector<>(); // generated queries curSeq
326 seqRefs = new Hashtable<>();
330 DbSourceProxy dbsource = dbSources[db];
331 // for moment, we dumbly iterate over all retrieval sources for a
332 // particular database
333 // TODO: introduce multithread multisource queries and logic to remove a
334 // query from other sources if any source for a database returns a
336 maxqlen = dbsource.getMaximumQueryCount();
338 while (queries.size() > 0 || seqIndex < currSeqs.length)
340 if (queries.size() > 0)
342 // Still queries to make for current seqIndex
343 StringBuffer queryString = new StringBuffer("");
345 int nqSize = (maxqlen > queries.size()) ? queries.size()
348 while (queries.size() > 0 && numq < nqSize)
350 String query = queries.elementAt(0);
351 if (dbsource.isValidReference(query))
354 (numq == 0) ? "" : dbsource.getAccessionSeparator());
355 queryString.append(query);
358 // remove the extracted query string
359 queries.removeElementAt(0);
361 // make the queries and process the response
362 AlignmentI retrieved = null;
365 if (Cache.log.isDebugEnabled())
367 Cache.log.debug("Querying " + dbsource.getDbName()
368 + " with : '" + queryString.toString() + "'");
370 retrieved = dbsource.getSequenceRecords(queryString.toString());
371 } catch (Exception ex)
373 ex.printStackTrace();
374 } catch (OutOfMemoryError err)
376 new OOMWarning("retrieving database references ("
377 + queryString.toString() + ")", err);
379 if (retrieved != null)
381 transferReferences(sdataset, dbsource, retrieved,
382 trimDsSeqs, warningMessages);
387 // make some more strings for use as queries
388 for (int i = 0; (seqIndex < dataset.length)
389 && (i < 50); seqIndex++, i++)
391 SequenceI sequence = dataset[seqIndex];
392 List<DBRefEntry> uprefs = DBRefUtils
393 .selectRefs(sequence.getDBRefs(), new String[]
394 { dbsource.getDbSource() }); // jalview.datamodel.DBRefSource.UNIPROT
396 // check for existing dbrefs to use
397 if (uprefs != null && uprefs.size() > 0)
399 for (int j = 0, n = uprefs.size(); j < n; j++)
401 DBRefEntry upref = uprefs.get(j);
402 addSeqId(sequence, upref.getAccessionId());
404 upref.getAccessionId().toUpperCase(Locale.ROOT));
409 Pattern possibleIds = Pattern.compile("[A-Za-z0-9_]+");
410 // generate queries from sequence ID string
411 Matcher tokens = possibleIds.matcher(sequence.getName());
413 while (tokens.find(p))
415 String token = tokens.group();
417 UPEntry[] presp = null;
418 if (picrClient != null)
420 // resolve the string against PICR to recover valid IDs
423 presp = picrClient.getUPIForAccession(token, null,
424 picrClient.getMappedDatabaseNames(), null,
426 } catch (Exception e)
429 "Exception with Picr for '" + token + "'\n");
433 if (presp != null && presp.length > 0)
435 for (int id = 0; id < presp.length; id++)
437 // construct sequences from response if sequences are
438 // present, and do a transferReferences
439 // otherwise transfer non sequence x-references directly.
442 "Validated ID against PICR... (for what its worth):"
444 addSeqId(sequence, token);
445 queries.addElement(token.toUpperCase(Locale.ROOT));
450 // System.out.println("Not querying source with
451 // token="+token+"\n");
452 addSeqId(sequence, token);
453 queries.addElement(token.toUpperCase(Locale.ROOT));
460 // advance to next database
462 } // all databases have been queried
463 if (!warningMessages.isEmpty())
465 StringBuilder sb = new StringBuilder(warningMessages.size() * 30);
466 sb.append(MessageManager
467 .getString("label.your_sequences_have_been_verified"));
468 for (String msg : warningMessages)
470 sb.append(msg).append(NEWLINE);
472 output.setText(sb.toString());
474 Desktop.addInternalFrame(output,
475 MessageManager.getString("label.sequences_updated"), 600,
477 // The above is the dataset, we must now find out the index
478 // of the viewed sequence
481 if (progressWindow != null)
483 progressWindow.setProgressBar(
484 MessageManager.getString("label.dbref_search_completed"),
488 for (FetchFinishedListenerI listener : listeners)
495 * Verify local sequences in seqRefs against the retrieved sequence database
496 * records. Returns true if any sequence was modified as a result (start/end
497 * changed and/or sequence enlarged), else false.
500 * dataset sequences we are retrieving for
502 * database source we are retrieving from
504 * retrieved sequences as alignment
505 * @param trimDatasetSeqs
506 * if true, sequences will not be enlarged to match longer retrieved
507 * sequences, only their start/end adjusted
508 * @param warningMessages
509 * a list of messages to add to
511 boolean transferReferences(Vector<SequenceI> sdataset,
512 DbSourceProxy dbSourceProxy,
513 AlignmentI retrievedAl, boolean trimDatasetSeqs,
514 List<String> warningMessages)
516 // System.out.println("trimming ? " + trimDatasetSeqs);
517 if (retrievedAl == null || retrievedAl.getHeight() == 0)
522 String dbSource = dbSourceProxy.getDbName();
523 boolean modified = false;
524 SequenceI[] retrieved = recoverDbSequences(
525 retrievedAl.getSequencesArray());
526 SequenceI sequence = null;
528 for (SequenceI retrievedSeq : retrieved)
530 // Work out which sequences this sequence matches,
531 // taking into account all accessionIds and names in the file
532 Vector<SequenceI> sequenceMatches = new Vector<>();
533 // look for corresponding accession ids
534 List<DBRefEntry> entryRefs = DBRefUtils
535 .selectRefs(retrievedSeq.getDBRefs(), new String[]
537 if (entryRefs == null)
540 .println("Dud dbSource string ? no entryrefs selected for "
541 + dbSource + " on " + retrievedSeq.getName());
544 for (int j = 0, n = entryRefs.size(); j < n; j++)
546 DBRefEntry ref = entryRefs.get(j);
547 String accessionId = ref.getAccessionId();
548 // match up on accessionId
549 if (seqRefs.containsKey(accessionId.toUpperCase(Locale.ROOT)))
551 Vector<SequenceI> seqs = seqRefs.get(accessionId);
552 for (int jj = 0; jj < seqs.size(); jj++)
554 sequence = seqs.elementAt(jj);
555 if (!sequenceMatches.contains(sequence))
557 sequenceMatches.addElement(sequence);
562 if (sequenceMatches.isEmpty())
564 // failed to match directly on accessionId==query so just compare all
565 // sequences to entry
566 Enumeration<String> e = seqRefs.keys();
567 while (e.hasMoreElements())
569 Vector<SequenceI> sqs = seqRefs.get(e.nextElement());
570 if (sqs != null && sqs.size() > 0)
572 Enumeration<SequenceI> sqe = sqs.elements();
573 while (sqe.hasMoreElements())
575 sequenceMatches.addElement(sqe.nextElement());
580 // look for corresponding names
581 // this is uniprot specific ?
582 // could be useful to extend this so we try to find any 'significant'
583 // information in common between two sequence objects.
585 * List<DBRefEntry> entryRefs =
586 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
587 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
588 * name = entry.getName().elementAt(j).toString(); if
589 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
590 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
591 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
592 * sequenceMatches.addElement(sequence); } } } }
594 if (sequenceMatches.size() > 0)
596 addFeatureSettings(dbSourceProxy);
598 // sequenceMatches now contains the set of all sequences associated with
599 // the returned db record
600 final String retrievedSeqString = retrievedSeq.getSequenceAsString();
601 String entrySeq = retrievedSeqString.toUpperCase(Locale.ROOT);
602 for (int m = 0; m < sequenceMatches.size(); m++)
604 sequence = sequenceMatches.elementAt(m);
605 // only update start and end positions and shift features if there are
606 // no existing references
607 // TODO: test for legacy where uniprot or EMBL refs exist but no
608 // mappings are made (but content matches retrieved set)
609 boolean updateRefFrame = sequence.getDBRefs() == null
610 || sequence.getDBRefs().size() == 0;
612 // verify sequence against the entry sequence
615 final int sequenceStart = sequence.getStart();
617 boolean remoteEnclosesLocal = false;
618 String nonGapped = AlignSeq
619 .extractGaps("-. ", sequence.getSequenceAsString())
620 .toUpperCase(Locale.ROOT);
621 int absStart = entrySeq.indexOf(nonGapped);
624 // couldn't find local sequence in sequence from database, so check if
625 // the database sequence is a subsequence of local sequence
626 absStart = nonGapped.indexOf(entrySeq);
629 // verification failed. couldn't find any relationship between
630 // entrySeq and local sequence
631 // messages suppressed as many-to-many matches are confusing
632 // String msg = sequence.getName()
633 // + " Sequence not 100% match with "
634 // + retrievedSeq.getName();
635 // addWarningMessage(warningMessages, msg);
639 * retrieved sequence is a proper subsequence of local sequence
641 String msg = sequence.getName() + " has " + absStart
642 + " prefixed residues compared to "
643 + retrievedSeq.getName();
644 addWarningMessage(warningMessages, msg);
647 * So create a mapping to the external entry from the matching region of
648 * the local sequence, and leave local start/end untouched.
650 mp = new Mapping(null,
652 { sequenceStart + absStart,
653 sequenceStart + absStart + entrySeq.length() - 1 },
655 { retrievedSeq.getStart(),
656 retrievedSeq.getStart() + entrySeq.length() - 1 },
658 updateRefFrame = false;
663 * local sequence is a subsequence of (or matches) retrieved sequence
665 remoteEnclosesLocal = true;
671 * relocate existing sequence features by offset
673 int startShift = absStart - sequenceStart + 1;
676 modified |= sequence.getFeatures().shiftFeatures(1,
682 System.out.println("Adding dbrefs to " + sequence.getName()
683 + " from " + dbSource + " sequence : "
684 + retrievedSeq.getName());
685 sequence.transferAnnotation(retrievedSeq, mp);
687 absStart += retrievedSeq.getStart();
688 int absEnd = absStart + nonGapped.length() - 1;
689 if (!trimDatasetSeqs)
692 * update start position and/or expand to longer retrieved sequence
694 if (!retrievedSeqString.equals(sequence.getSequenceAsString())
695 && remoteEnclosesLocal)
697 sequence.setSequence(retrievedSeqString);
699 addWarningMessage(warningMessages,
700 "Sequence for " + sequence.getName() + " expanded from "
701 + retrievedSeq.getName());
703 if (sequence.getStart() != retrievedSeq.getStart())
705 sequence.setStart(retrievedSeq.getStart());
707 if (absStart != sequenceStart)
709 addWarningMessage(warningMessages,
710 "Start/end position for " + sequence.getName()
711 + " updated from " + retrievedSeq.getName());
717 // finally, update local sequence reference frame if we're allowed
720 // just fix start/end
721 if (sequence.getStart() != absStart
722 || sequence.getEnd() != absEnd)
724 sequence.setStart(absStart);
725 sequence.setEnd(absEnd);
727 addWarningMessage(warningMessages,
728 "Start/end for " + sequence.getName()
729 + " updated from " + retrievedSeq.getName());
732 // search for alignment sequences to update coordinate frame for
733 for (int alsq = 0; alsq < alseqs.length; alsq++)
735 if (alseqs[alsq].getDatasetSequence() == sequence)
737 String ngAlsq = AlignSeq
739 alseqs[alsq].getSequenceAsString())
740 .toUpperCase(Locale.ROOT);
741 int oldstrt = alseqs[alsq].getStart();
742 alseqs[alsq].setStart(sequence.getSequenceAsString()
743 .toUpperCase(Locale.ROOT).indexOf(ngAlsq) + sequence.getStart());
744 if (oldstrt != alseqs[alsq].getStart())
747 ngAlsq.length() + alseqs[alsq].getStart() - 1);
752 // TODO: search for all other references to this dataset sequence, and
754 // TODO: update all AlCodonMappings which involve this alignment
755 // sequence (e.g. Q30167 cdna translation from exon2 product (vamsas
758 // and remove it from the rest
759 // TODO: decide if we should remove annotated sequence from set
760 sdataset.remove(sequence);
766 Map<String, FeatureSettingsModelI> featureDisplaySettings = null;
768 private void addFeatureSettings(DbSourceProxy dbSourceProxy)
770 FeatureSettingsModelI fsettings = dbSourceProxy
771 .getFeatureColourScheme();
772 if (fsettings != null)
774 if (featureDisplaySettings == null)
776 featureDisplaySettings = new HashMap<>();
778 featureDisplaySettings.put(dbSourceProxy.getDbName(), fsettings);
784 * @return any feature settings associated with sources that have provided sequences
786 public List<FeatureSettingsModelI>getFeatureSettingsModels()
788 return featureDisplaySettings == null
789 ? Arrays.asList(new FeatureSettingsModelI[0])
790 : Arrays.asList(featureDisplaySettings.values()
791 .toArray(new FeatureSettingsModelI[1]));
794 * Adds the message to the list unless it already contains it
799 void addWarningMessage(List<String> messageList, String msg)
801 if (!messageList.contains(msg))
803 messageList.add(msg);
808 * loop thru and collect additional sequences in Map.
810 * @param sequencesArray
813 private SequenceI[] recoverDbSequences(SequenceI[] sequencesArray)
816 if (sequencesArray == null || (n = sequencesArray.length) == 0)
817 return sequencesArray;
818 ArrayList<SequenceI> nseq = new ArrayList<>();
819 for (int i = 0;i < n; i++)
821 nseq.add(sequencesArray[i]);
822 List<DBRefEntry> dbr = sequencesArray[i].getDBRefs();
826 for (int r = 0, rn = dbr.size(); r < rn; r++)
828 if ((map = dbr.get(r).getMap()) != null)
830 if (map.getTo() != null && !nseq.contains(map.getTo()))
832 nseq.add(map.getTo());
838 // BH 2019.01.25 question here if this is the right logic. Return the original if nothing found?
841 return nseq.toArray(new SequenceI[nseq.size()]);
843 return sequencesArray;