2 * Jalview - A Sequence Alignment Editor and Viewer (Version 2.4)
\r
3 * Copyright (C) 2008 AM Waterhouse, J Procter, G Barton, M Clamp, S Searle
\r
5 * This program is free software; you can redistribute it and/or
\r
6 * modify it under the terms of the GNU General Public License
\r
7 * as published by the Free Software Foundation; either version 2
\r
8 * of the License, or (at your option) any later version.
\r
10 * This program is distributed in the hope that it will be useful,
\r
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
\r
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
\r
13 * GNU General Public License for more details.
\r
15 * You should have received a copy of the GNU General Public License
\r
16 * along with this program; if not, write to the Free Software
\r
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
\r
24 import org.biojava.dasobert.dasregistry.DasSource;
\r
25 import org.exolab.castor.mapping.*;
\r
26 import org.exolab.castor.xml.*;
\r
27 import jalview.analysis.*;
\r
28 import jalview.datamodel.*;
\r
29 import jalview.datamodel.Mapping;
\r
30 import jalview.gui.*;
\r
31 import jalview.ws.dbsources.Uniprot;
\r
32 import jalview.ws.ebi.EBIFetchClient;
\r
35 * Implements a runnable for validating a sequence against external databases
\r
36 * and then propagating references and features onto the sequence(s)
\r
39 * @version $Revision$
\r
41 public class DBRefFetcher implements Runnable
\r
43 SequenceI[] dataset;
\r
45 IProgressIndicator af;
\r
47 CutAndPasteTransfer output = new CutAndPasteTransfer();
\r
49 StringBuffer sbuffer = new StringBuffer();
\r
51 boolean running = false;
\r
53 // /This will be a collection of Vectors of sequenceI refs.
\r
54 // The key will be the seq name or accession id of the seq
\r
59 SequenceFetcher sfetcher;
\r
61 public DBRefFetcher()
\r
66 * Creates a new SequenceFeatureFetcher object and fetches from the
\r
67 * currently selected set of databases.
\r
70 * fetch references for these sequences
\r
72 * the parent alignframe for progress bar monitoring.
\r
74 public DBRefFetcher(SequenceI[] seqs, AlignFrame af)
\r
76 this(seqs, af, null);
\r
79 * Creates a new SequenceFeatureFetcher object and fetches from the
\r
80 * currently selected set of databases.
\r
83 * fetch references for these sequences
\r
85 * the parent alignframe for progress bar monitoring.
\r
86 * @param sources array of database source strings to query references from
\r
88 public DBRefFetcher(SequenceI[] seqs, AlignFrame af, String[] sources)
\r
91 SequenceI[] ds = new SequenceI[seqs.length];
\r
92 for (int i = 0; i < seqs.length; i++)
\r
94 if (seqs[i].getDatasetSequence() != null)
\r
95 ds[i] = seqs[i].getDatasetSequence();
\r
100 // TODO Jalview 2.5 lots of this code should be in the gui package!
\r
101 sfetcher = jalview.gui.SequenceFetcher.getSequenceFetcherSingleton(af);
\r
104 // af.featureSettings_actionPerformed(null);
\r
105 String[] defdb=null,otherdb = sfetcher.getDbInstances(jalview.ws.dbsources.DasSequenceSource.class);
\r
106 Vector selsources = new Vector(), dasselsrc= (af.featureSettings!=null) ? af.featureSettings.getSelectedSources()
\r
107 : new jalview.gui.DasSourceBrowser().getSelectedSources();
\r
108 Enumeration en = dasselsrc.elements();
\r
109 while (en.hasMoreElements())
\r
111 DasSource src = (DasSource) en.nextElement();
\r
112 selsources.addElement(src.getNickname());
\r
115 for (int o=0;otherdb!=null && o<otherdb.length;o++)
\r
117 if (!selsources.contains(otherdb[o]))
\r
124 // select appropriate databases based on alignFrame context.
\r
125 if (af.getViewport().getAlignment().isNucleotide())
\r
127 defdb = DBRefSource.DNACODINGDBS;
\r
131 defdb = DBRefSource.PROTEINDBS;
\r
133 // append the selected sequence sources to the default dbs
\r
134 dbSources = new String[defdb.length+osel];
\r
135 System.arraycopy(defdb, 0, dbSources, 0, defdb.length);
\r
136 for (int o=0,op=defdb.length; otherdb!=null && o<otherdb.length; o++)
\r
138 if (otherdb[o]!=null)
\r
140 dbSources[op++] = otherdb[o];
\r
144 // we assume the caller knows what they're doing and ensured that all the db source names are valid
\r
145 dbSources = sources;
\r
149 * retrieve all the das sequence sources and add them to the list of db sources to retrieve from
\r
151 public void appendAllDasSources()
\r
153 if (dbSources == null)
\r
155 dbSources = new String[] {};
\r
157 // append additional sources
\r
158 String[] otherdb = sfetcher.getDbInstances(jalview.ws.dbsources.DasSequenceSource.class);
\r
159 if (otherdb!=null && otherdb.length>0)
\r
161 String[] newsrc = new String[dbSources.length+otherdb.length];
\r
162 System.arraycopy(dbSources, 0, newsrc,0,dbSources.length);
\r
163 System.arraycopy(otherdb, 0, newsrc,dbSources.length, otherdb.length);
\r
164 dbSources = newsrc;
\r
168 * start the fetcher thread
\r
170 * @param waitTillFinished
\r
171 * true to block until the fetcher has finished
\r
173 public void fetchDBRefs(boolean waitTillFinished)
\r
175 Thread thread = new Thread(this);
\r
179 if (waitTillFinished)
\r
186 } catch (Exception ex)
\r
194 * The sequence will be added to a vector of sequences belonging to key which
\r
195 * could be either seq name or dbref id
\r
202 void addSeqId(SequenceI seq, String key)
\r
204 key = key.toUpperCase();
\r
207 if (seqRefs.containsKey(key))
\r
209 seqs = (Vector) seqRefs.get(key);
\r
211 if (seqs != null && !seqs.contains(seq))
\r
213 seqs.addElement(seq);
\r
215 else if (seqs == null)
\r
217 seqs = new Vector();
\r
218 seqs.addElement(seq);
\r
224 seqs = new Vector();
\r
225 seqs.addElement(seq);
\r
228 seqRefs.put(key, seqs);
\r
236 if (dbSources == null)
\r
238 throw new Error("Implementation error. Must initialise dbSources");
\r
240 long startTime = System.currentTimeMillis();
\r
241 af.setProgressBar("Fetching db refs", startTime);
\r
244 Vector sdataset = new Vector();
\r
245 for (int s = 0; s < dataset.length; s++)
\r
247 sdataset.addElement(dataset[s]);
\r
249 while (sdataset.size() > 0 && db < dbSources.length)
\r
251 int maxqlen = 1; // default number of queries made to at one time
\r
252 System.err.println("Verifying against " + dbSources[db]);
\r
253 jalview.ws.seqfetcher.DbSourceProxy dbsource = sfetcher
\r
254 .getSourceProxy(dbSources[db]);
\r
255 if (dbsource == null)
\r
257 System.err.println("No proxy for " + dbSources[db]);
\r
261 if (dbsource.getDbSourceProperties()
\r
262 .containsKey(DBRefSource.MULTIACC))
\r
264 maxqlen = ((Integer) dbsource.getDbSourceProperties().get(
\r
265 DBRefSource.MULTIACC)).intValue();
\r
271 // iterate through db for each remaining un-verified sequence
\r
272 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
\r
273 sdataset.copyInto(currSeqs);// seqs that are to be validated against
\r
275 Vector queries = new Vector(); // generated queries curSeq
\r
276 seqRefs = new Hashtable();
\r
280 while (queries.size() > 0 || seqIndex < currSeqs.length)
\r
282 if (queries.size() > 0)
\r
284 // Still queries to make for current seqIndex
\r
285 StringBuffer queryString = new StringBuffer("");
\r
286 int numq=0,nqSize = (maxqlen > queries.size()) ? queries.size()
\r
289 while (queries.size()>0 && numq < nqSize)
\r
291 String query = (String) queries.elementAt(0);
\r
292 if (dbsource.isValidReference(query))
\r
294 queryString.append((numq == 0) ? "" : dbsource
\r
295 .getAccessionSeparator());
\r
296 queryString.append(query);
\r
299 // remove the extracted query string
\r
300 queries.removeElementAt(0);
\r
302 // make the queries and process the response
\r
303 AlignmentI retrieved = null;
\r
306 if (jalview.bin.Cache.log.isDebugEnabled())
\r
308 jalview.bin.Cache.log.debug("Querying "+dbsource.getDbName()+" with : '"+queryString.toString()+"'");
\r
310 retrieved = dbsource.getSequenceRecords(queryString.toString());
\r
311 } catch (Exception ex)
\r
313 ex.printStackTrace();
\r
314 } catch (OutOfMemoryError err)
\r
316 new OOMWarning("retrieving database references ("
\r
317 + queryString.toString() + ")", err);
\r
319 if (retrieved != null)
\r
321 transferReferences(sdataset, dbSources[db], retrieved);
\r
326 // make some more strings for use as queries
\r
327 for (int i = 0; (seqIndex < dataset.length) && (i < 50); seqIndex++, i++)
\r
329 SequenceI sequence = dataset[seqIndex];
\r
330 DBRefEntry[] uprefs = jalview.util.DBRefUtils.selectRefs(
\r
331 sequence.getDBRef(), new String[]
\r
332 { dbSources[db] }); // jalview.datamodel.DBRefSource.UNIPROT
\r
334 // check for existing dbrefs to use
\r
335 if (uprefs != null && uprefs.length>0)
\r
337 for (int j = 0; j < uprefs.length; j++)
\r
339 addSeqId(sequence, uprefs[j].getAccessionId());
\r
341 .addElement(uprefs[j].getAccessionId()
\r
347 // generate queries from sequence ID string
\r
348 StringTokenizer st = new StringTokenizer(sequence.getName(),
\r
350 while (st.hasMoreTokens())
\r
352 String token = st.nextToken();
\r
353 addSeqId(sequence, token);
\r
354 queries.addElement(token.toUpperCase());
\r
360 // advance to next database
\r
362 } // all databases have been queries.
\r
363 if (sbuffer.length() > 0)
\r
366 .setText("Your sequences have been verified against known sequence databases. Some of the ids have been\n"
\r
367 + "altered, most likely the start/end residue will have been updated.\n"
\r
368 + "Save your alignment to maintain the updated id.\n\n"
\r
369 + sbuffer.toString());
\r
370 Desktop.addInternalFrame(output, "Sequence names updated ", 600, 300);
\r
371 // The above is the dataset, we must now find out the index
\r
372 // of the viewed sequence
\r
376 af.setProgressBar("DBRef search completed", startTime);
\r
377 // promptBeforeBlast();
\r
384 * Verify local sequences in seqRefs against the retrieved sequence database
\r
388 void transferReferences(Vector sdataset, String dbSource,
\r
389 AlignmentI retrievedAl) // File
\r
393 if (retrievedAl == null || retrievedAl.getHeight() == 0)
\r
397 SequenceI[] retrieved = retrievedAl.getSequencesArray();
\r
398 SequenceI sequence = null;
\r
400 // Vector entries = new Uniprot().getUniprotEntries(file);
\r
402 int i, iSize = retrieved.length; // entries == null ? 0 : entries.size();
\r
403 // UniprotEntry entry;
\r
404 for (i = 0; i < iSize; i++)
\r
406 SequenceI entry = retrieved[i]; // (UniprotEntry) entries.elementAt(i);
\r
408 // Work out which sequences this sequence matches,
\r
409 // taking into account all accessionIds and names in the file
\r
410 Vector sequenceMatches = new Vector();
\r
411 // look for corresponding accession ids
\r
412 DBRefEntry[] entryRefs = jalview.util.DBRefUtils.selectRefs(entry
\r
413 .getDBRef(), new String[]
\r
415 for (int j = 0; j < entryRefs.length; j++)
\r
417 String accessionId = entryRefs[j].getAccessionId(); // .getAccession().elementAt(j).toString();
\r
418 // match up on accessionId
\r
419 if (seqRefs.containsKey(accessionId.toUpperCase()))
\r
421 Vector seqs = (Vector) seqRefs.get(accessionId);
\r
422 for (int jj = 0; jj < seqs.size(); jj++)
\r
424 sequence = (SequenceI) seqs.elementAt(jj);
\r
425 if (!sequenceMatches.contains(sequence))
\r
427 sequenceMatches.addElement(sequence);
\r
432 if (sequenceMatches.size() == 0)
\r
434 // failed to match directly on accessionId==query so just compare all
\r
435 // sequences to entry
\r
436 Enumeration e = seqRefs.keys();
\r
437 while (e.hasMoreElements())
\r
439 Vector sqs = (Vector) seqRefs.get(e.nextElement());
\r
440 if (sqs != null && sqs.size() > 0)
\r
442 Enumeration sqe = sqs.elements();
\r
443 while (sqe.hasMoreElements())
\r
445 sequenceMatches.addElement(sqe.nextElement());
\r
450 // look for corresponding names
\r
451 // this is uniprot specific ?
\r
452 // could be useful to extend this so we try to find any 'significant'
\r
453 // information in common between two sequence objects.
\r
455 * DBRefEntry[] entryRefs =
\r
456 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
\r
457 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
\r
458 * name = entry.getName().elementAt(j).toString(); if
\r
459 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
\r
460 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
\r
461 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
\r
462 * sequenceMatches.addElement(sequence); } } } }
\r
464 // sequenceMatches now contains the set of all sequences associated with
\r
465 // the returned db record
\r
466 String entrySeq = entry.getSequenceAsString().toUpperCase();
\r
467 for (int m = 0; m < sequenceMatches.size(); m++)
\r
469 sequence = (SequenceI) sequenceMatches.elementAt(m);
\r
470 // only update start and end positions and shift features if there are
\r
471 // no existing references
\r
472 // TODO: test for legacy where uniprot or EMBL refs exist but no
\r
473 // mappings are made (but content matches retrieved set)
\r
474 boolean updateRefFrame = sequence.getDBRef() == null
\r
475 || sequence.getDBRef().length == 0;
\r
476 // verify sequence against the entry sequence
\r
478 String nonGapped = AlignSeq.extractGaps("-. ",
\r
479 sequence.getSequenceAsString()).toUpperCase();
\r
481 int absStart = entrySeq.indexOf(nonGapped);
\r
482 int mapStart = entry.getStart();
\r
483 jalview.datamodel.Mapping mp;
\r
485 if (absStart == -1)
\r
487 // Is local sequence contained in dataset sequence?
\r
488 absStart = nonGapped.indexOf(entrySeq);
\r
489 if (absStart == -1)
\r
490 { // verification failed.
\r
491 sbuffer.append(sequence.getName()
\r
492 + " SEQUENCE NOT %100 MATCH \n");
\r
496 sbuffer.append(sequence.getName() + " HAS " + absStart
\r
497 + " PREFIXED RESIDUES COMPARED TO " + dbSource + "\n");
\r
499 // + " - ANY SEQUENCE FEATURES"
\r
500 // + " HAVE BEEN ADJUSTED ACCORDINGLY \n");
\r
502 // create valid mapping between matching region of local sequence and
\r
503 // the mapped sequence
\r
504 mp = new Mapping(null, new int[]
\r
505 { sequence.getStart() + absStart,
\r
506 sequence.getStart() + absStart + entrySeq.length() - 1 },
\r
508 { entry.getStart(),
\r
509 entry.getStart() + entrySeq.length() - 1 }, 1, 1);
\r
510 updateRefFrame = false; // mapping is based on current start/end so
\r
511 // don't modify start and end
\r
515 // update start and end of local sequence to place it in entry's
\r
516 // reference frame.
\r
517 // apply identity map map from whole of local sequence to matching
\r
518 // region of database
\r
520 mp = null; // Mapping.getIdentityMap();
\r
521 // new Mapping(null,
\r
522 // new int[] { absStart+sequence.getStart(),
\r
523 // absStart+sequence.getStart()+entrySeq.length()-1},
\r
524 // new int[] { entry.getStart(), entry.getEnd() }, 1, 1);
\r
525 // relocate local features for updated start
\r
526 if (updateRefFrame && sequence.getSequenceFeatures() != null)
\r
528 SequenceFeature[] sf = sequence.getSequenceFeatures();
\r
529 int start = sequence.getStart();
\r
530 int end = sequence.getEnd();
\r
531 int startShift = 1 - absStart - start; // how much the features are
\r
532 // to be shifted by
\r
533 for (int sfi = 0; sfi < sf.length; sfi++)
\r
535 if (sf[sfi].getBegin() >= start && sf[sfi].getEnd() <= end)
\r
537 // shift feature along by absstart
\r
538 sf[sfi].setBegin(sf[sfi].getBegin() + startShift);
\r
539 sf[sfi].setEnd(sf[sfi].getEnd() + startShift);
\r
545 System.out.println("Adding dbrefs to " + sequence.getName()
\r
546 + " from " + dbSource + " sequence : " + entry.getName());
\r
547 sequence.transferAnnotation(entry, mp);
\r
548 // unknownSequences.remove(sequence);
\r
549 int absEnd = absStart + nonGapped.length();
\r
551 if (updateRefFrame)
\r
553 // finally, update local sequence reference frame if we're allowed
\r
554 sequence.setStart(absStart);
\r
555 sequence.setEnd(absEnd);
\r
557 // and remove it from the rest
\r
558 // TODO: decide if we should remove annotated sequence from set
\r
559 sdataset.remove(sequence);
\r
560 // TODO: should we make a note of sequences that have received new DB ids, so we can query all enabled DAS servers for them ?
\r