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.util.DBRefUtils;
37 import jalview.util.MessageManager;
38 import jalview.ws.seqfetcher.DbSourceProxy;
40 import java.util.ArrayList;
41 import java.util.Arrays;
42 import java.util.Enumeration;
43 import java.util.HashMap;
44 import java.util.Hashtable;
45 import java.util.List;
47 import java.util.StringTokenizer;
48 import java.util.Vector;
50 import uk.ac.ebi.picr.model.UPEntry;
51 import uk.ac.ebi.www.picr.AccessionMappingService.AccessionMapperServiceLocator;
54 * Implements a runnable for validating a sequence against external databases
55 * and then propagating references and features onto the sequence(s)
60 public class DBRefFetcher implements Runnable
62 private static final String NEWLINE = System.lineSeparator();
64 public static final String TRIM_RETRIEVED_SEQUENCES = "TRIM_FETCHED_DATASET_SEQS";
66 public interface FetchFinishedListenerI
73 IProgressIndicator progressWindow;
75 CutAndPasteTransfer output = new CutAndPasteTransfer();
77 boolean running = false;
80 * picr client instance
82 uk.ac.ebi.www.picr.AccessionMappingService.AccessionMapperInterface picrClient = null;
84 // This will be a collection of Vectors of sequenceI refs.
85 // The key will be the seq name or accession id of the seq
86 Hashtable<String, Vector<SequenceI>> seqRefs;
88 DbSourceProxy[] dbSources;
90 SequenceFetcher sfetcher;
92 private List<FetchFinishedListenerI> listeners;
94 private SequenceI[] alseqs;
97 * when true - retrieved sequences will be trimmed to cover longest derived
100 private boolean trimDsSeqs = true;
103 * Creates a new DBRefFetcher object and fetches from the currently selected
104 * set of databases, if this is null then it fetches based on feature settings
107 * fetch references for these SequenceI array
108 * @param progressIndicatorFrame
109 * the frame for progress bar monitoring
111 * array of DbSourceProxy to query references form
112 * @param featureSettings
113 * FeatureSettings to get alternative DbSourceProxy from
114 * @param isNucleotide
115 * indicates if the array of SequenceI are Nucleotides or not
117 public DBRefFetcher(SequenceI[] seqs,
118 IProgressIndicator progressIndicatorFrame,
119 DbSourceProxy[] sources, FeatureSettings featureSettings,
120 boolean isNucleotide)
122 listeners = new ArrayList<>();
123 this.progressWindow = progressIndicatorFrame;
124 alseqs = new SequenceI[seqs.length];
125 SequenceI[] ds = new SequenceI[seqs.length];
126 for (int i = 0; i < seqs.length; i++)
129 if (seqs[i].getDatasetSequence() != null)
131 ds[i] = seqs[i].getDatasetSequence();
139 // TODO Jalview 2.5 lots of this code should be in the gui package!
140 sfetcher = jalview.gui.SequenceFetcher.getSequenceFetcherSingleton();
141 // set default behaviour for transferring excess sequence data to the
143 trimDsSeqs = Cache.getDefault(TRIM_RETRIEVED_SEQUENCES, true);
146 setDatabaseSources(featureSettings, isNucleotide);
150 // we assume the caller knows what they're doing and ensured that all the
151 // db source names are valid
157 * Helper method to configure the list of database sources to query
159 * @param featureSettings
160 * @param forNucleotide
162 void setDatabaseSources(FeatureSettings featureSettings,
163 boolean forNucleotide)
165 // af.featureSettings_actionPerformed(null);
166 String[] defdb = null;
167 List<DbSourceProxy> selsources = new ArrayList<>();
168 // select appropriate databases based on alignFrame context.
171 defdb = DBRefSource.DNACODINGDBS;
175 defdb = DBRefSource.PROTEINDBS;
177 List<DbSourceProxy> srces = new ArrayList<>();
178 for (String ddb : defdb)
180 List<DbSourceProxy> srcesfordb = sfetcher.getSourceProxy(ddb);
181 if (srcesfordb != null)
183 for (DbSourceProxy src : srcesfordb)
185 if (!srces.contains(src))
187 srces.addAll(srcesfordb);
192 // append the PDB data source, since it is 'special', catering for both
193 // nucleotide and protein
194 // srces.addAll(sfetcher.getSourceProxy(DBRefSource.PDB));
196 srces.addAll(selsources);
197 dbSources = srces.toArray(new DbSourceProxy[srces.size()]);
201 * Constructor with only sequences provided
205 public DBRefFetcher(SequenceI[] sequences)
207 this(sequences, null, null, null, false);
211 * Add a listener to be notified when sequence fetching is complete
215 public void addListener(FetchFinishedListenerI l)
221 * start the fetcher thread
223 * @param waitTillFinished
224 * true to block until the fetcher has finished
226 public void fetchDBRefs(boolean waitTillFinished)
228 // TODO can we not simply write
229 // if (waitTillFinished) { run(); } else { new Thread(this).start(); }
231 Thread thread = new Thread(this);
235 if (waitTillFinished)
242 } catch (Exception ex)
250 * The sequence will be added to a vector of sequences belonging to key which
251 * could be either seq name or dbref id
258 void addSeqId(SequenceI seq, String key)
260 key = key.toUpperCase();
262 Vector<SequenceI> seqs;
263 if (seqRefs.containsKey(key))
265 seqs = seqRefs.get(key);
267 if (seqs != null && !seqs.contains(seq))
269 seqs.addElement(seq);
271 else if (seqs == null)
273 seqs = new Vector<>();
274 seqs.addElement(seq);
280 seqs = new Vector<>();
281 seqs.addElement(seq);
284 seqRefs.put(key, seqs);
293 if (dbSources == null)
295 throw new Error(MessageManager
296 .getString("error.implementation_error_must_init_dbsources"));
299 long startTime = System.currentTimeMillis();
300 if (progressWindow != null)
302 progressWindow.setProgressBar(
303 MessageManager.getString("status.fetching_db_refs"),
308 if (Cache.getDefault("DBREFFETCH_USEPICR", false))
310 picrClient = new AccessionMapperServiceLocator()
311 .getAccessionMapperPort();
313 } catch (Exception e)
315 System.err.println("Couldn't locate PICR service instance.\n");
319 Vector<SequenceI> sdataset = new Vector<>(
320 Arrays.asList(dataset));
321 List<String> warningMessages = new ArrayList<>();
323 // clear any old feature display settings recorded from past sessions
324 featureDisplaySettings = null;
327 while (sdataset.size() > 0 && db < dbSources.length)
329 int maxqlen = 1; // default number of queries made at one time
330 System.out.println("Verifying against " + dbSources[db].getDbName());
332 // iterate through db for each remaining un-verified sequence
333 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
334 sdataset.copyInto(currSeqs);// seqs that are to be validated against
336 Vector<String> queries = new Vector<>(); // generated queries curSeq
337 seqRefs = new Hashtable<>();
341 DbSourceProxy dbsource = dbSources[db];
342 // for moment, we dumbly iterate over all retrieval sources for a
343 // particular database
344 // TODO: introduce multithread multisource queries and logic to remove a
345 // query from other sources if any source for a database returns a
347 maxqlen = dbsource.getMaximumQueryCount();
349 while (queries.size() > 0 || seqIndex < currSeqs.length)
351 if (queries.size() > 0)
353 // Still queries to make for current seqIndex
354 StringBuffer queryString = new StringBuffer("");
356 int nqSize = (maxqlen > queries.size()) ? queries.size()
359 while (queries.size() > 0 && numq < nqSize)
361 String query = queries.elementAt(0);
362 if (dbsource.isValidReference(query))
365 (numq == 0) ? "" : dbsource.getAccessionSeparator());
366 queryString.append(query);
369 // remove the extracted query string
370 queries.removeElementAt(0);
372 // make the queries and process the response
373 AlignmentI retrieved = null;
376 if (Cache.log.isDebugEnabled())
378 Cache.log.debug("Querying " + dbsource.getDbName()
379 + " with : '" + queryString.toString() + "'");
381 retrieved = dbsource.getSequenceRecords(queryString.toString());
382 } catch (Exception ex)
384 ex.printStackTrace();
385 } catch (OutOfMemoryError err)
387 new OOMWarning("retrieving database references ("
388 + queryString.toString() + ")", err);
390 if (retrieved != null)
392 transferReferences(sdataset, dbsource, retrieved,
393 trimDsSeqs, warningMessages);
398 // make some more strings for use as queries
399 for (int i = 0; (seqIndex < dataset.length)
400 && (i < 50); seqIndex++, i++)
402 SequenceI sequence = dataset[seqIndex];
403 List<DBRefEntry> uprefs = DBRefUtils
404 .selectRefs(sequence.getDBRefs(), new String[]
405 { dbsource.getDbSource() }); // jalview.datamodel.DBRefSource.UNIPROT
407 // check for existing dbrefs to use
408 if (uprefs != null && uprefs.size() > 0)
410 for (int j = 0, n = uprefs.size(); j < n; j++)
412 DBRefEntry upref = uprefs.get(j);
413 addSeqId(sequence, upref.getAccessionId());
415 upref.getAccessionId().toUpperCase());
420 // generate queries from sequence ID string
421 StringTokenizer st = new StringTokenizer(sequence.getName(),
423 while (st.hasMoreTokens())
425 String token = st.nextToken();
426 UPEntry[] presp = null;
427 if (picrClient != null)
429 // resolve the string against PICR to recover valid IDs
432 presp = picrClient.getUPIForAccession(token, null,
433 picrClient.getMappedDatabaseNames(), null,
435 } catch (Exception e)
438 "Exception with Picr for '" + token + "'\n");
442 if (presp != null && presp.length > 0)
444 for (int id = 0; id < presp.length; id++)
446 // construct sequences from response if sequences are
447 // present, and do a transferReferences
448 // otherwise transfer non sequence x-references directly.
451 "Validated ID against PICR... (for what its worth):"
453 addSeqId(sequence, token);
454 queries.addElement(token.toUpperCase());
459 // System.out.println("Not querying source with
460 // token="+token+"\n");
461 addSeqId(sequence, token);
462 queries.addElement(token.toUpperCase());
469 // advance to next database
471 } // all databases have been queried
472 if (!warningMessages.isEmpty())
474 StringBuilder sb = new StringBuilder(warningMessages.size() * 30);
475 sb.append(MessageManager
476 .getString("label.your_sequences_have_been_verified"));
477 for (String msg : warningMessages)
479 sb.append(msg).append(NEWLINE);
481 output.setText(sb.toString());
483 Desktop.addInternalFrame(output,
484 MessageManager.getString("label.sequences_updated"), 600,
486 // The above is the dataset, we must now find out the index
487 // of the viewed sequence
490 if (progressWindow != null)
492 progressWindow.setProgressBar(
493 MessageManager.getString("label.dbref_search_completed"),
497 for (FetchFinishedListenerI listener : listeners)
505 * Verify local sequences in seqRefs against the retrieved sequence database
506 * records. Returns true if any sequence was modified as a result (start/end
507 * changed and/or sequence enlarged), else false.
510 * dataset sequences we are retrieving for
512 * database source we are retrieving from
514 * retrieved sequences as alignment
515 * @param trimDatasetSeqs
516 * if true, sequences will not be enlarged to match longer retrieved
517 * sequences, only their start/end adjusted
518 * @param warningMessages
519 * a list of messages to add to
521 boolean transferReferences(Vector<SequenceI> sdataset,
522 DbSourceProxy dbSourceProxy,
523 AlignmentI retrievedAl, boolean trimDatasetSeqs,
524 List<String> warningMessages)
526 // System.out.println("trimming ? " + trimDatasetSeqs);
527 if (retrievedAl == null || retrievedAl.getHeight() == 0)
532 String dbSource = dbSourceProxy.getDbName();
533 boolean modified = false;
534 SequenceI[] retrieved = recoverDbSequences(
535 retrievedAl.getSequencesArray());
536 SequenceI sequence = null;
538 for (SequenceI retrievedSeq : retrieved)
540 // Work out which sequences this sequence matches,
541 // taking into account all accessionIds and names in the file
542 Vector<SequenceI> sequenceMatches = new Vector<>();
543 // look for corresponding accession ids
544 List<DBRefEntry> entryRefs = DBRefUtils
545 .selectRefs(retrievedSeq.getDBRefs(), new String[]
547 if (entryRefs == null)
550 .println("Dud dbSource string ? no entryrefs selected for "
551 + dbSource + " on " + retrievedSeq.getName());
554 for (int j = 0, n = entryRefs.size(); j < n; j++)
556 DBRefEntry ref = entryRefs.get(j);
557 String accessionId = ref.getAccessionId();
558 // match up on accessionId
559 if (seqRefs.containsKey(accessionId.toUpperCase()))
561 Vector<SequenceI> seqs = seqRefs.get(accessionId);
562 for (int jj = 0; jj < seqs.size(); jj++)
564 sequence = seqs.elementAt(jj);
565 if (!sequenceMatches.contains(sequence))
567 sequenceMatches.addElement(sequence);
572 if (sequenceMatches.isEmpty())
574 // failed to match directly on accessionId==query so just compare all
575 // sequences to entry
576 Enumeration<String> e = seqRefs.keys();
577 while (e.hasMoreElements())
579 Vector<SequenceI> sqs = seqRefs.get(e.nextElement());
580 if (sqs != null && sqs.size() > 0)
582 Enumeration<SequenceI> sqe = sqs.elements();
583 while (sqe.hasMoreElements())
585 sequenceMatches.addElement(sqe.nextElement());
590 // look for corresponding names
591 // this is uniprot specific ?
592 // could be useful to extend this so we try to find any 'significant'
593 // information in common between two sequence objects.
595 * List<DBRefEntry> entryRefs =
596 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
597 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
598 * name = entry.getName().elementAt(j).toString(); if
599 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
600 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
601 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
602 * sequenceMatches.addElement(sequence); } } } }
604 if (sequenceMatches.size() > 0)
606 addFeatureSettings(dbSourceProxy);
608 // sequenceMatches now contains the set of all sequences associated with
609 // the returned db record
610 final String retrievedSeqString = retrievedSeq.getSequenceAsString();
611 String entrySeq = retrievedSeqString.toUpperCase();
612 for (int m = 0; m < sequenceMatches.size(); m++)
614 sequence = sequenceMatches.elementAt(m);
615 // only update start and end positions and shift features if there are
616 // no existing references
617 // TODO: test for legacy where uniprot or EMBL refs exist but no
618 // mappings are made (but content matches retrieved set)
619 boolean updateRefFrame = sequence.getDBRefs() == null
620 || sequence.getDBRefs().size() == 0;
622 // verify sequence against the entry sequence
625 final int sequenceStart = sequence.getStart();
627 boolean remoteEnclosesLocal = false;
628 String nonGapped = AlignSeq
629 .extractGaps("-. ", sequence.getSequenceAsString())
631 int absStart = entrySeq.indexOf(nonGapped);
634 // couldn't find local sequence in sequence from database, so check if
635 // the database sequence is a subsequence of local sequence
636 absStart = nonGapped.indexOf(entrySeq);
639 // verification failed. couldn't find any relationship between
640 // entrySeq and local sequence
641 // messages suppressed as many-to-many matches are confusing
642 // String msg = sequence.getName()
643 // + " Sequence not 100% match with "
644 // + retrievedSeq.getName();
645 // addWarningMessage(warningMessages, msg);
649 * retrieved sequence is a proper subsequence of local sequence
651 String msg = sequence.getName() + " has " + absStart
652 + " prefixed residues compared to "
653 + retrievedSeq.getName();
654 addWarningMessage(warningMessages, msg);
657 * So create a mapping to the external entry from the matching region of
658 * the local sequence, and leave local start/end untouched.
660 mp = new Mapping(null,
662 { sequenceStart + absStart,
663 sequenceStart + absStart + entrySeq.length() - 1 },
665 { retrievedSeq.getStart(),
666 retrievedSeq.getStart() + entrySeq.length() - 1 },
668 updateRefFrame = false;
673 * local sequence is a subsequence of (or matches) retrieved sequence
675 remoteEnclosesLocal = true;
681 * relocate existing sequence features by offset
683 int startShift = absStart - sequenceStart + 1;
686 modified |= sequence.getFeatures().shiftFeatures(1,
692 System.out.println("Adding dbrefs to " + sequence.getName()
693 + " from " + dbSource + " sequence : "
694 + retrievedSeq.getName());
695 sequence.transferAnnotation(retrievedSeq, mp);
697 absStart += retrievedSeq.getStart();
698 int absEnd = absStart + nonGapped.length() - 1;
699 if (!trimDatasetSeqs)
702 * update start position and/or expand to longer retrieved sequence
704 if (!retrievedSeqString.equals(sequence.getSequenceAsString())
705 && remoteEnclosesLocal)
707 sequence.setSequence(retrievedSeqString);
709 addWarningMessage(warningMessages,
710 "Sequence for " + sequence.getName() + " expanded from "
711 + retrievedSeq.getName());
713 if (sequence.getStart() != retrievedSeq.getStart())
715 sequence.setStart(retrievedSeq.getStart());
717 if (absStart != sequenceStart)
719 addWarningMessage(warningMessages,
720 "Start/end position for " + sequence.getName()
721 + " updated from " + retrievedSeq.getName());
727 // finally, update local sequence reference frame if we're allowed
730 // just fix start/end
731 if (sequence.getStart() != absStart
732 || sequence.getEnd() != absEnd)
734 sequence.setStart(absStart);
735 sequence.setEnd(absEnd);
737 addWarningMessage(warningMessages,
738 "Start/end for " + sequence.getName()
739 + " updated from " + retrievedSeq.getName());
742 // search for alignment sequences to update coordinate frame for
743 for (int alsq = 0; alsq < alseqs.length; alsq++)
745 if (alseqs[alsq].getDatasetSequence() == sequence)
747 String ngAlsq = AlignSeq
749 alseqs[alsq].getSequenceAsString())
751 int oldstrt = alseqs[alsq].getStart();
752 alseqs[alsq].setStart(sequence.getSequenceAsString()
753 .toUpperCase().indexOf(ngAlsq) + sequence.getStart());
754 if (oldstrt != alseqs[alsq].getStart())
757 ngAlsq.length() + alseqs[alsq].getStart() - 1);
762 // TODO: search for all other references to this dataset sequence, and
764 // TODO: update all AlCodonMappings which involve this alignment
765 // sequence (e.g. Q30167 cdna translation from exon2 product (vamsas
768 // and remove it from the rest
769 // TODO: decide if we should remove annotated sequence from set
770 sdataset.remove(sequence);
776 Map<String, FeatureSettingsModelI> featureDisplaySettings = null;
778 private void addFeatureSettings(DbSourceProxy dbSourceProxy)
780 FeatureSettingsModelI fsettings = dbSourceProxy
781 .getFeatureColourScheme();
782 if (fsettings != null)
784 if (featureDisplaySettings == null)
786 featureDisplaySettings = new HashMap<>();
788 featureDisplaySettings.put(dbSourceProxy.getDbName(), fsettings);
794 * @return any feature settings associated with sources that have provided sequences
796 public List<FeatureSettingsModelI>getFeatureSettingsModels()
798 return featureDisplaySettings == null
799 ? Arrays.asList(new FeatureSettingsModelI[0])
800 : Arrays.asList(featureDisplaySettings.values()
801 .toArray(new FeatureSettingsModelI[1]));
804 * Adds the message to the list unless it already contains it
809 void addWarningMessage(List<String> messageList, String msg)
811 if (!messageList.contains(msg))
813 messageList.add(msg);
818 * loop thru and collect additional sequences in Map.
820 * @param sequencesArray
823 private SequenceI[] recoverDbSequences(SequenceI[] sequencesArray)
826 if (sequencesArray == null || (n = sequencesArray.length) == 0)
827 return sequencesArray;
828 ArrayList<SequenceI> nseq = new ArrayList<>();
829 for (int i = 0;i < n; i++)
831 nseq.add(sequencesArray[i]);
832 List<DBRefEntry> dbr = sequencesArray[i].getDBRefs();
836 for (int r = 0, rn = dbr.size(); r < rn; r++)
838 if ((map = dbr.get(r).getMap()) != null)
840 if (map.getTo() != null && !nseq.contains(map.getTo()))
842 nseq.add(map.getTo());
848 // BH 2019.01.25 question here if this is the right logic. Return the original if nothing found?
851 return nseq.toArray(new SequenceI[nseq.size()]);
853 return sequencesArray;