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.exolab.castor.mapping.*;
\r
25 import org.exolab.castor.xml.*;
\r
26 import jalview.analysis.*;
\r
27 import jalview.datamodel.*;
\r
28 import jalview.datamodel.Mapping;
\r
29 import jalview.gui.*;
\r
30 import jalview.ws.dbsources.Uniprot;
\r
31 import jalview.ws.ebi.EBIFetchClient;
\r
34 * Implements a runnable for validating a sequence
\r
35 * against external databases and then propagating
\r
36 * 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.
\r
69 * fetch references for these sequences
\r
71 * the parent alignframe for progress bar monitoring.
\r
73 public DBRefFetcher(SequenceI[] seqs, AlignFrame af)
\r
76 SequenceI[] ds = new SequenceI[seqs.length];
\r
77 for (int i = 0; i < seqs.length; i++)
\r
79 if (seqs[i].getDatasetSequence() != null)
\r
80 ds[i] = seqs[i].getDatasetSequence();
\r
85 // TODO Jalview 2.5 lots of this code should be in the gui package!
\r
86 sfetcher = jalview.gui.SequenceFetcher.getSequenceFetcherSingleton(af);
\r
87 // select appropriate databases based on alignFrame context.
\r
88 if (af.getViewport().getAlignment().isNucleotide())
\r
90 dbSources = DBRefSource.DNACODINGDBS;
\r
94 dbSources = DBRefSource.PROTEINDBS;
\r
99 * start the fetcher thread
\r
101 * @param waitTillFinished
\r
102 * true to block until the fetcher has finished
\r
104 public void fetchDBRefs(boolean waitTillFinished)
\r
106 Thread thread = new Thread(this);
\r
110 if (waitTillFinished)
\r
117 } catch (Exception ex)
\r
125 * The sequence will be added to a vector of sequences belonging to key which
\r
126 * could be either seq name or dbref id
\r
133 void addSeqId(SequenceI seq, String key)
\r
135 key = key.toUpperCase();
\r
138 if (seqRefs.containsKey(key))
\r
140 seqs = (Vector) seqRefs.get(key);
\r
142 if (seqs != null && !seqs.contains(seq))
\r
144 seqs.addElement(seq);
\r
146 else if (seqs == null)
\r
148 seqs = new Vector();
\r
149 seqs.addElement(seq);
\r
155 seqs = new Vector();
\r
156 seqs.addElement(seq);
\r
159 seqRefs.put(key, seqs);
\r
167 if (dbSources == null)
\r
169 throw new Error("Implementation error. Must initialise dbSources");
\r
171 long startTime = System.currentTimeMillis();
\r
172 af.setProgressBar("Fetching db refs", startTime);
\r
175 Vector sdataset = new Vector();
\r
176 for (int s = 0; s < dataset.length; s++)
\r
178 sdataset.addElement(dataset[s]);
\r
180 while (sdataset.size() > 0 && db < dbSources.length)
\r
182 int maxqlen = 1; // default number of queries made to at one time
\r
183 System.err.println("Verifying against " + dbSources[db]);
\r
184 jalview.ws.seqfetcher.DbSourceProxy dbsource = sfetcher
\r
185 .getSourceProxy(dbSources[db]);
\r
186 if (dbsource == null)
\r
188 System.err.println("No proxy for " + dbSources[db]);
\r
192 if (dbsource.getDbSourceProperties()
\r
193 .containsKey(DBRefSource.MULTIACC))
\r
195 maxqlen = ((Integer) dbsource.getDbSourceProperties().get(
\r
196 DBRefSource.MULTIACC)).intValue();
\r
200 // iterate through db for each remaining un-verified sequence
\r
201 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
\r
202 sdataset.copyInto(currSeqs);// seqs that are to be validated against
\r
204 Vector queries = new Vector(); // generated queries curSeq
\r
205 seqRefs = new Hashtable();
\r
209 while (queries.size() > 0 || seqIndex < currSeqs.length)
\r
211 if (queries.size() > 0)
\r
213 // Still queries to make for current seqIndex
\r
214 StringBuffer queryString = new StringBuffer("");
\r
215 int nqSize = (maxqlen > queries.size()) ? queries.size()
\r
217 for (int nq = 0, numq = 0; nq < nqSize; nq++)
\r
219 String query = (String) queries.elementAt(nq);
\r
220 if (dbsource.isValidReference(query))
\r
222 queryString.append((nq == 0) ? "" : dbsource
\r
223 .getAccessionSeparator());
\r
224 queryString.append(query);
\r
228 for (int nq = 0; nq < nqSize; nq++)
\r
230 queries.removeElementAt(0);
\r
232 // make the queries and process the response
\r
233 AlignmentI retrieved = null;
\r
236 retrieved = dbsource.getSequenceRecords(queryString.toString());
\r
237 } catch (Exception ex)
\r
239 ex.printStackTrace();
\r
241 catch (OutOfMemoryError err)
\r
243 new OOMWarning("retrieving database references ("+queryString.toString()+")", err);
\r
245 if (retrieved != null)
\r
247 transferReferences(sdataset, dbSources[db], retrieved);
\r
252 // make some more strings for use as queries
\r
253 for (int i = 0; (seqIndex < dataset.length) && (i < 50); seqIndex++, i++)
\r
255 SequenceI sequence = dataset[seqIndex];
\r
256 DBRefEntry[] uprefs = jalview.util.DBRefUtils.selectRefs(
\r
257 sequence.getDBRef(), new String[]
\r
258 { dbSources[db] }); // jalview.datamodel.DBRefSource.UNIPROT
\r
260 // check for existing dbrefs to use
\r
261 if (uprefs != null)
\r
263 for (int j = 0; j < uprefs.length; j++)
\r
265 addSeqId(sequence, uprefs[j].getAccessionId());
\r
267 .addElement(uprefs[j].getAccessionId()
\r
273 // generate queries from sequence ID string
\r
274 StringTokenizer st = new StringTokenizer(sequence.getName(),
\r
276 while (st.hasMoreTokens())
\r
278 String token = st.nextToken();
\r
279 addSeqId(sequence, token);
\r
280 queries.addElement(token.toUpperCase());
\r
286 // advance to next database
\r
288 } // all databases have been queries.
\r
289 if (sbuffer.length() > 0)
\r
292 .setText("Your sequences have been verified against known sequence databases. Some of the ids have been\n"
\r
293 + "altered, most likely the start/end residue will have been updated.\n"
\r
294 + "Save your alignment to maintain the updated id.\n\n"
\r
295 + sbuffer.toString());
\r
296 Desktop.addInternalFrame(output, "Sequence names updated ", 600, 300);
\r
297 // The above is the dataset, we must now find out the index
\r
298 // of the viewed sequence
\r
302 af.setProgressBar("DBRef search completed", startTime);
\r
303 // promptBeforeBlast();
\r
310 * Verify local sequences in seqRefs against the retrieved sequence database
\r
314 void transferReferences(Vector sdataset, String dbSource,
\r
315 AlignmentI retrievedAl) // File
\r
319 if (retrievedAl == null || retrievedAl.getHeight() == 0)
\r
323 SequenceI[] retrieved = retrievedAl.getSequencesArray();
\r
324 SequenceI sequence = null;
\r
326 // Vector entries = new Uniprot().getUniprotEntries(file);
\r
328 int i, iSize = retrieved.length; // entries == null ? 0 : entries.size();
\r
329 // UniprotEntry entry;
\r
330 for (i = 0; i < iSize; i++)
\r
332 SequenceI entry = retrieved[i]; // (UniprotEntry) entries.elementAt(i);
\r
334 // Work out which sequences this sequence matches,
\r
335 // taking into account all accessionIds and names in the file
\r
336 Vector sequenceMatches = new Vector();
\r
337 // look for corresponding accession ids
\r
338 DBRefEntry[] entryRefs = jalview.util.DBRefUtils.selectRefs(entry
\r
339 .getDBRef(), new String[]
\r
341 for (int j = 0; j < entryRefs.length; j++)
\r
343 String accessionId = entryRefs[j].getAccessionId(); // .getAccession().elementAt(j).toString();
\r
344 // match up on accessionId
\r
345 if (seqRefs.containsKey(accessionId.toUpperCase()))
\r
347 Vector seqs = (Vector) seqRefs.get(accessionId);
\r
348 for (int jj = 0; jj < seqs.size(); jj++)
\r
350 sequence = (SequenceI) seqs.elementAt(jj);
\r
351 if (!sequenceMatches.contains(sequence))
\r
353 sequenceMatches.addElement(sequence);
\r
358 if (sequenceMatches.size()==0)
\r
360 // failed to match directly on accessionId==query so just compare all sequences to entry
\r
361 Enumeration e = seqRefs.keys();
\r
362 while (e.hasMoreElements())
\r
364 Vector sqs = (Vector) seqRefs.get(e.nextElement());
\r
365 if (sqs!=null && sqs.size()>0)
\r
367 Enumeration sqe = sqs.elements();
\r
368 while (sqe.hasMoreElements())
\r
370 sequenceMatches.addElement(sqe.nextElement());
\r
375 // look for corresponding names
\r
376 // this is uniprot specific ?
\r
377 // could be useful to extend this so we try to find any 'significant'
\r
378 // information in common between two sequence objects.
\r
380 * DBRefEntry[] entryRefs =
\r
381 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
\r
382 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
\r
383 * name = entry.getName().elementAt(j).toString(); if
\r
384 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
\r
385 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
\r
386 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
\r
387 * sequenceMatches.addElement(sequence); } } } }
\r
389 // sequenceMatches now contains the set of all sequences associated with
\r
390 // the returned db record
\r
391 String entrySeq = entry.getSequenceAsString().toUpperCase();
\r
392 for (int m = 0; m < sequenceMatches.size(); m++)
\r
394 sequence = (SequenceI) sequenceMatches.elementAt(m);
\r
395 // only update start and end positions and shift features if there are no existing references
\r
396 // TODO: test for legacy where uniprot or EMBL refs exist but no mappings are made (but content matches retrieved set)
\r
397 boolean updateRefFrame = sequence.getDBRef()==null || sequence.getDBRef().length==0;
\r
398 // verify sequence against the entry sequence
\r
400 String nonGapped = AlignSeq.extractGaps("-. ",
\r
401 sequence.getSequenceAsString()).toUpperCase();
\r
403 int absStart = entrySeq.indexOf(nonGapped);
\r
404 int mapStart = entry.getStart();
\r
405 jalview.datamodel.Mapping mp;
\r
407 if (absStart == -1)
\r
409 // Is local sequence contained in dataset sequence?
\r
410 absStart = nonGapped.indexOf(entrySeq);
\r
411 if (absStart == -1)
\r
412 { // verification failed.
\r
413 sbuffer.append(sequence.getName()
\r
414 + " SEQUENCE NOT %100 MATCH \n");
\r
418 sbuffer.append(sequence.getName() + " HAS " + absStart
\r
419 + " PREFIXED RESIDUES COMPARED TO " + dbSource+"\n");
\r
421 // + " - ANY SEQUENCE FEATURES"
\r
422 // + " HAVE BEEN ADJUSTED ACCORDINGLY \n");
\r
424 // create valid mapping between matching region of local sequence and
\r
425 // the mapped sequence
\r
426 mp = new Mapping(null, new int[]
\r
427 { sequence.getStart()+absStart, sequence.getStart()+absStart+entrySeq.length()-1 }, new int[]
\r
428 { entry.getStart(),
\r
429 entry.getStart() + entrySeq.length() - 1 }, 1, 1);
\r
430 updateRefFrame=false; // mapping is based on current start/end so don't modify start and end
\r
434 // update start and end of local sequence to place it in entry's
\r
435 // reference frame.
\r
436 // apply identity map map from whole of local sequence to matching
\r
437 // region of database
\r
439 mp = null; // Mapping.getIdentityMap();
\r
440 // new Mapping(null,
\r
441 // new int[] { absStart+sequence.getStart(),
\r
442 // absStart+sequence.getStart()+entrySeq.length()-1},
\r
443 // new int[] { entry.getStart(), entry.getEnd() }, 1, 1);
\r
444 // relocate local features for updated start
\r
445 if (updateRefFrame && sequence.getSequenceFeatures() != null)
\r
447 SequenceFeature[] sf = sequence.getSequenceFeatures();
\r
448 int start = sequence.getStart();
\r
449 int end = sequence.getEnd();
\r
450 int startShift = 1-absStart-start; // how much the features are to be shifted by
\r
451 for (int sfi = 0; sfi < sf.length; sfi++)
\r
453 if (sf[sfi].getBegin() >= start && sf[sfi].getEnd() <= end)
\r
455 // shift feature along by absstart
\r
456 sf[sfi].setBegin(sf[sfi].getBegin() + startShift);
\r
457 sf[sfi].setEnd(sf[sfi].getEnd() + startShift);
\r
463 System.out.println("Adding dbrefs to " + sequence.getName()
\r
464 + " from " + dbSource + " sequence : " + entry.getName());
\r
465 sequence.transferAnnotation(entry, mp);
\r
466 // unknownSequences.remove(sequence);
\r
467 int absEnd = absStart + nonGapped.length();
\r
469 if (updateRefFrame)
\r
471 // finally, update local sequence reference frame if we're allowed
\r
472 sequence.setStart(absStart);
\r
473 sequence.setEnd(absEnd);
\r
475 // and remove it from the rest
\r
476 // TODO: decide if we should remove annotated sequence from set
\r
477 sdataset.remove(sequence);
\r