2 * Jalview - A Sequence Alignment Editor and Viewer
\r
3 * Copyright (C) 2007 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
37 * @version $Revision$
\r
39 public class DBRefFetcher implements Runnable
\r
41 SequenceI[] dataset;
\r
45 CutAndPasteTransfer output = new CutAndPasteTransfer();
\r
47 StringBuffer sbuffer = new StringBuffer();
\r
49 boolean running = false;
\r
51 // /This will be a collection of Vectors of sequenceI refs.
\r
52 // The key will be the seq name or accession id of the seq
\r
57 SequenceFetcher sfetcher;
\r
59 public DBRefFetcher()
\r
64 * Creates a new SequenceFeatureFetcher object.
\r
67 * fetch references for these sequences
\r
69 * the parent alignframe for progress bar monitoring.
\r
71 public DBRefFetcher(SequenceI[] seqs, AlignFrame af)
\r
74 SequenceI[] ds = new SequenceI[seqs.length];
\r
75 for (int i = 0; i < seqs.length; i++)
\r
77 if (seqs[i].getDatasetSequence() != null)
\r
78 ds[i] = seqs[i].getDatasetSequence();
\r
83 sfetcher = new SequenceFetcher();
\r
84 // select appropriate databases based on alignFrame context.
\r
85 if (af.getViewport().getAlignment().isNucleotide())
\r
87 dbSources = DBRefSource.DNACODINGDBS;
\r
91 dbSources = DBRefSource.PROTEINDBS;
\r
96 * start the fetcher thread
\r
98 * @param waitTillFinished
\r
99 * true to block until the fetcher has finished
\r
101 public void fetchDBRefs(boolean waitTillFinished)
\r
103 Thread thread = new Thread(this);
\r
107 if (waitTillFinished)
\r
114 } catch (Exception ex)
\r
122 * The sequence will be added to a vector of sequences belonging to key which
\r
123 * could be either seq name or dbref id
\r
130 void addSeqId(SequenceI seq, String key)
\r
132 key = key.toUpperCase();
\r
135 if (seqRefs.containsKey(key))
\r
137 seqs = (Vector) seqRefs.get(key);
\r
139 if (seqs != null && !seqs.contains(seq))
\r
141 seqs.addElement(seq);
\r
143 else if (seqs == null)
\r
145 seqs = new Vector();
\r
146 seqs.addElement(seq);
\r
152 seqs = new Vector();
\r
153 seqs.addElement(seq);
\r
156 seqRefs.put(key, seqs);
\r
164 if (dbSources == null)
\r
166 throw new Error("Implementation error. Must initialise dbSources");
\r
168 long startTime = System.currentTimeMillis();
\r
169 af.setProgressBar("Fetching db refs", startTime);
\r
172 Vector sdataset = new Vector();
\r
173 for (int s = 0; s < dataset.length; s++)
\r
175 sdataset.addElement(dataset[s]);
\r
177 while (sdataset.size() > 0 && db < dbSources.length)
\r
179 int maxqlen = 1; // default number of queries made to at one time
\r
180 System.err.println("Verifying against " + dbSources[db]);
\r
181 jalview.ws.seqfetcher.DbSourceProxy dbsource = sfetcher
\r
182 .getSourceProxy(dbSources[db]);
\r
183 if (dbsource == null)
\r
185 System.err.println("No proxy for " + dbSources[db]);
\r
189 if (dbsource.getDbSourceProperties()
\r
190 .containsKey(DBRefSource.MULTIACC))
\r
192 maxqlen = ((Integer) dbsource.getDbSourceProperties().get(
\r
193 DBRefSource.MULTIACC)).intValue();
\r
195 // iterate through db for each remaining un-verified sequence
\r
196 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
\r
197 sdataset.copyInto(currSeqs);// seqs that are to be validated against
\r
199 Vector queries = new Vector(); // generated queries curSeq
\r
200 seqRefs = new Hashtable();
\r
204 while (queries.size() > 0 || seqIndex < currSeqs.length)
\r
206 if (queries.size() > 0)
\r
208 // Still queries to make for current seqIndex
\r
209 StringBuffer queryString = new StringBuffer("");
\r
210 int nqSize = (maxqlen > queries.size()) ? queries.size()
\r
212 for (int nq = 0, numq = 0; nq < nqSize; nq++)
\r
214 String query = (String) queries.elementAt(nq);
\r
215 if (dbsource.isValidReference(query))
\r
217 queryString.append((nq == 0) ? "" : dbsource
\r
218 .getAccessionSeparator());
\r
219 queryString.append(query);
\r
223 for (int nq = 0; nq < nqSize; nq++)
\r
225 queries.removeElementAt(0);
\r
227 // make the queries and process the response
\r
228 AlignmentI retrieved = null;
\r
231 retrieved = dbsource.getSequenceRecords(queryString.toString());
\r
232 } catch (Exception ex)
\r
234 ex.printStackTrace();
\r
236 if (retrieved != null)
\r
238 transferReferences(sdataset, dbSources[db], retrieved);
\r
243 // make some more strings for use as queries
\r
244 for (int i = 0; (seqIndex < dataset.length) && (i < 50); seqIndex++, i++)
\r
246 SequenceI sequence = dataset[seqIndex];
\r
247 DBRefEntry[] uprefs = jalview.util.DBRefUtils.selectRefs(
\r
248 sequence.getDBRef(), new String[]
\r
249 { dbSources[db] }); // jalview.datamodel.DBRefSource.UNIPROT
\r
251 // check for existing dbrefs to use
\r
252 if (uprefs != null)
\r
254 for (int j = 0; j < uprefs.length; j++)
\r
256 addSeqId(sequence, uprefs[j].getAccessionId());
\r
258 .addElement(uprefs[j].getAccessionId()
\r
264 // generate queries from sequence ID string
\r
265 StringTokenizer st = new StringTokenizer(sequence.getName(),
\r
267 while (st.hasMoreTokens())
\r
269 String token = st.nextToken();
\r
270 addSeqId(sequence, token);
\r
271 queries.addElement(token.toUpperCase());
\r
277 // advance to next database
\r
279 } // all databases have been queries.
\r
280 if (sbuffer.length() > 0)
\r
283 .setText("Your sequences have been verified against known sequence databases. Some of the ids have been\n"
\r
284 + "altered, most likely the start/end residue will have been updated.\n"
\r
285 + "Save your alignment to maintain the updated id.\n\n"
\r
286 + sbuffer.toString());
\r
287 Desktop.addInternalFrame(output, "Sequence names updated ", 600, 300);
\r
288 // The above is the dataset, we must now find out the index
\r
289 // of the viewed sequence
\r
293 af.setProgressBar("DBRef search completed", startTime);
\r
294 // promptBeforeBlast();
\r
301 * Verify local sequences in seqRefs against the retrieved sequence database
\r
305 void transferReferences(Vector sdataset, String dbSource,
\r
306 AlignmentI retrievedAl) // File
\r
310 if (retrievedAl == null || retrievedAl.getHeight() == 0)
\r
314 SequenceI[] retrieved = retrievedAl.getSequencesArray();
\r
315 SequenceI sequence = null;
\r
317 // Vector entries = new Uniprot().getUniprotEntries(file);
\r
319 int i, iSize = retrieved.length; // entries == null ? 0 : entries.size();
\r
320 // UniprotEntry entry;
\r
321 for (i = 0; i < iSize; i++)
\r
323 SequenceI entry = retrieved[i]; // (UniprotEntry) entries.elementAt(i);
\r
325 // Work out which sequences this sequence matches,
\r
326 // taking into account all accessionIds and names in the file
\r
327 Vector sequenceMatches = new Vector();
\r
328 // look for corresponding accession ids
\r
329 DBRefEntry[] entryRefs = jalview.util.DBRefUtils.selectRefs(entry
\r
330 .getDBRef(), new String[]
\r
332 for (int j = 0; j < entryRefs.length; j++)
\r
334 String accessionId = entryRefs[j].getAccessionId(); // .getAccession().elementAt(j).toString();
\r
335 // match up on accessionId
\r
336 if (seqRefs.containsKey(accessionId.toUpperCase()))
\r
338 Vector seqs = (Vector) seqRefs.get(accessionId);
\r
339 for (int jj = 0; jj < seqs.size(); jj++)
\r
341 sequence = (SequenceI) seqs.elementAt(jj);
\r
342 if (!sequenceMatches.contains(sequence))
\r
344 sequenceMatches.addElement(sequence);
\r
349 if (sequenceMatches.size()==0)
\r
351 // failed to match directly on accessionId==query so just compare all sequences to entry
\r
352 Enumeration e = seqRefs.keys();
\r
353 while (e.hasMoreElements())
\r
355 Vector sqs = (Vector) seqRefs.get(e.nextElement());
\r
356 if (sqs!=null && sqs.size()>0)
\r
358 Enumeration sqe = sqs.elements();
\r
359 while (sqe.hasMoreElements())
\r
361 sequenceMatches.addElement(sqe.nextElement());
\r
366 // look for corresponding names
\r
367 // this is uniprot specific ?
\r
368 // could be useful to extend this so we try to find any 'significant'
\r
369 // information in common between two sequence objects.
\r
371 * DBRefEntry[] entryRefs =
\r
372 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
\r
373 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
\r
374 * name = entry.getName().elementAt(j).toString(); if
\r
375 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
\r
376 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
\r
377 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
\r
378 * sequenceMatches.addElement(sequence); } } } }
\r
380 // sequenceMatches now contains the set of all sequences associated with
\r
381 // the returned db record
\r
382 String entrySeq = entry.getSequenceAsString().toUpperCase();
\r
383 for (int m = 0; m < sequenceMatches.size(); m++)
\r
385 sequence = (SequenceI) sequenceMatches.elementAt(m);
\r
386 // only update start and end positions and shift features if there are no existing references
\r
387 // TODO: test for legacy where uniprot or EMBL refs exist but no mappings are made (but content matches retrieved set)
\r
388 boolean updateRefFrame = sequence.getDBRef()==null || sequence.getDBRef().length==0;
\r
389 // verify sequence against the entry sequence
\r
391 String nonGapped = AlignSeq.extractGaps("-. ",
\r
392 sequence.getSequenceAsString()).toUpperCase();
\r
394 int absStart = entrySeq.indexOf(nonGapped);
\r
395 int mapStart = entry.getStart();
\r
396 jalview.datamodel.Mapping mp;
\r
398 if (absStart == -1)
\r
400 // Is local sequence contained in dataset sequence?
\r
401 absStart = nonGapped.indexOf(entrySeq);
\r
402 if (absStart == -1)
\r
403 { // verification failed.
\r
404 sbuffer.append(sequence.getName()
\r
405 + " SEQUENCE NOT %100 MATCH \n");
\r
409 sbuffer.append(sequence.getName() + " HAS " + absStart
\r
410 + " PREFIXED RESIDUES COMPARED TO " + dbSource+"\n");
\r
412 // + " - ANY SEQUENCE FEATURES"
\r
413 // + " HAVE BEEN ADJUSTED ACCORDINGLY \n");
\r
415 // create valid mapping between matching region of local sequence and
\r
416 // the mapped sequence
\r
417 mp = new Mapping(null, new int[]
\r
418 { sequence.getStart()+absStart, sequence.getStart()+absStart+entrySeq.length()-1 }, new int[]
\r
419 { entry.getStart(),
\r
420 entry.getStart() + entrySeq.length() - 1 }, 1, 1);
\r
421 updateRefFrame=false; // mapping is based on current start/end so don't modify start and end
\r
425 // update start and end of local sequence to place it in entry's
\r
426 // reference frame.
\r
427 // apply identity map map from whole of local sequence to matching
\r
428 // region of database
\r
430 mp = null; // Mapping.getIdentityMap();
\r
431 // new Mapping(null,
\r
432 // new int[] { absStart+sequence.getStart(),
\r
433 // absStart+sequence.getStart()+entrySeq.length()-1},
\r
434 // new int[] { entry.getStart(), entry.getEnd() }, 1, 1);
\r
435 // relocate local features for updated start
\r
436 if (updateRefFrame && sequence.getSequenceFeatures() != null)
\r
438 SequenceFeature[] sf = sequence.getSequenceFeatures();
\r
439 int start = sequence.getStart();
\r
440 int end = sequence.getEnd();
\r
441 int startShift = 1-absStart-start; // how much the features are to be shifted by
\r
442 for (int sfi = 0; sfi < sf.length; sfi++)
\r
444 if (sf[sfi].getBegin() >= start && sf[sfi].getEnd() <= end)
\r
446 // shift feature along by absstart
\r
447 sf[sfi].setBegin(sf[sfi].getBegin() + startShift);
\r
448 sf[sfi].setEnd(sf[sfi].getEnd() + startShift);
\r
454 System.out.println("Adding dbrefs to " + sequence.getName()
\r
455 + " from " + dbSource + " sequence : " + entry.getName());
\r
456 sequence.transferAnnotation(entry, mp);
\r
457 // unknownSequences.remove(sequence);
\r
458 int absEnd = absStart + nonGapped.length();
\r
460 if (updateRefFrame)
\r
462 // finally, update local sequence reference frame if we're allowed
\r
463 sequence.setStart(absStart);
\r
464 sequence.setEnd(absEnd);
\r
466 // and remove it from the rest
\r
467 // TODO: decide if we should remove annotated sequence from set
\r
468 sdataset.remove(sequence);
\r