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 against external databases
\r
35 * and then propagating references and features onto the sequence(s)
\r
38 * @version $Revision$
\r
40 public class DBRefFetcher implements Runnable
\r
42 SequenceI[] dataset;
\r
44 IProgressIndicator af;
\r
46 CutAndPasteTransfer output = new CutAndPasteTransfer();
\r
48 StringBuffer sbuffer = new StringBuffer();
\r
50 boolean running = false;
\r
52 // /This will be a collection of Vectors of sequenceI refs.
\r
53 // The key will be the seq name or accession id of the seq
\r
58 SequenceFetcher sfetcher;
\r
60 public DBRefFetcher()
\r
65 * Creates a new SequenceFeatureFetcher object.
\r
68 * fetch references for these sequences
\r
70 * the parent alignframe for progress bar monitoring.
\r
72 public DBRefFetcher(SequenceI[] seqs, AlignFrame af)
\r
75 SequenceI[] ds = new SequenceI[seqs.length];
\r
76 for (int i = 0; i < seqs.length; i++)
\r
78 if (seqs[i].getDatasetSequence() != null)
\r
79 ds[i] = seqs[i].getDatasetSequence();
\r
84 // TODO Jalview 2.5 lots of this code should be in the gui package!
\r
85 sfetcher = jalview.gui.SequenceFetcher.getSequenceFetcherSingleton(af);
\r
86 // select appropriate databases based on alignFrame context.
\r
87 if (af.getViewport().getAlignment().isNucleotide())
\r
89 dbSources = DBRefSource.DNACODINGDBS;
\r
93 dbSources = DBRefSource.PROTEINDBS;
\r
98 * start the fetcher thread
\r
100 * @param waitTillFinished
\r
101 * true to block until the fetcher has finished
\r
103 public void fetchDBRefs(boolean waitTillFinished)
\r
105 Thread thread = new Thread(this);
\r
109 if (waitTillFinished)
\r
116 } catch (Exception ex)
\r
124 * The sequence will be added to a vector of sequences belonging to key which
\r
125 * could be either seq name or dbref id
\r
132 void addSeqId(SequenceI seq, String key)
\r
134 key = key.toUpperCase();
\r
137 if (seqRefs.containsKey(key))
\r
139 seqs = (Vector) seqRefs.get(key);
\r
141 if (seqs != null && !seqs.contains(seq))
\r
143 seqs.addElement(seq);
\r
145 else if (seqs == null)
\r
147 seqs = new Vector();
\r
148 seqs.addElement(seq);
\r
154 seqs = new Vector();
\r
155 seqs.addElement(seq);
\r
158 seqRefs.put(key, seqs);
\r
166 if (dbSources == null)
\r
168 throw new Error("Implementation error. Must initialise dbSources");
\r
170 long startTime = System.currentTimeMillis();
\r
171 af.setProgressBar("Fetching db refs", startTime);
\r
174 Vector sdataset = new Vector();
\r
175 for (int s = 0; s < dataset.length; s++)
\r
177 sdataset.addElement(dataset[s]);
\r
179 while (sdataset.size() > 0 && db < dbSources.length)
\r
181 int maxqlen = 1; // default number of queries made to at one time
\r
182 System.err.println("Verifying against " + dbSources[db]);
\r
183 jalview.ws.seqfetcher.DbSourceProxy dbsource = sfetcher
\r
184 .getSourceProxy(dbSources[db]);
\r
185 if (dbsource == null)
\r
187 System.err.println("No proxy for " + dbSources[db]);
\r
191 if (dbsource.getDbSourceProperties()
\r
192 .containsKey(DBRefSource.MULTIACC))
\r
194 maxqlen = ((Integer) dbsource.getDbSourceProperties().get(
\r
195 DBRefSource.MULTIACC)).intValue();
\r
201 // iterate through db for each remaining un-verified sequence
\r
202 SequenceI[] currSeqs = new SequenceI[sdataset.size()];
\r
203 sdataset.copyInto(currSeqs);// seqs that are to be validated against
\r
205 Vector queries = new Vector(); // generated queries curSeq
\r
206 seqRefs = new Hashtable();
\r
210 while (queries.size() > 0 || seqIndex < currSeqs.length)
\r
212 if (queries.size() > 0)
\r
214 // Still queries to make for current seqIndex
\r
215 StringBuffer queryString = new StringBuffer("");
\r
216 int nqSize = (maxqlen > queries.size()) ? queries.size()
\r
218 for (int nq = 0, numq = 0; nq < nqSize; nq++)
\r
220 String query = (String) queries.elementAt(nq);
\r
221 if (dbsource.isValidReference(query))
\r
223 queryString.append((nq == 0) ? "" : dbsource
\r
224 .getAccessionSeparator());
\r
225 queryString.append(query);
\r
229 for (int nq = 0; nq < nqSize; nq++)
\r
231 queries.removeElementAt(0);
\r
233 // make the queries and process the response
\r
234 AlignmentI retrieved = null;
\r
237 retrieved = dbsource.getSequenceRecords(queryString.toString());
\r
238 } catch (Exception ex)
\r
240 ex.printStackTrace();
\r
241 } catch (OutOfMemoryError err)
\r
243 new OOMWarning("retrieving database references ("
\r
244 + queryString.toString() + ")", err);
\r
246 if (retrieved != null)
\r
248 transferReferences(sdataset, dbSources[db], retrieved);
\r
253 // make some more strings for use as queries
\r
254 for (int i = 0; (seqIndex < dataset.length) && (i < 50); seqIndex++, i++)
\r
256 SequenceI sequence = dataset[seqIndex];
\r
257 DBRefEntry[] uprefs = jalview.util.DBRefUtils.selectRefs(
\r
258 sequence.getDBRef(), new String[]
\r
259 { dbSources[db] }); // jalview.datamodel.DBRefSource.UNIPROT
\r
261 // check for existing dbrefs to use
\r
262 if (uprefs != null)
\r
264 for (int j = 0; j < uprefs.length; j++)
\r
266 addSeqId(sequence, uprefs[j].getAccessionId());
\r
268 .addElement(uprefs[j].getAccessionId()
\r
274 // generate queries from sequence ID string
\r
275 StringTokenizer st = new StringTokenizer(sequence.getName(),
\r
277 while (st.hasMoreTokens())
\r
279 String token = st.nextToken();
\r
280 addSeqId(sequence, token);
\r
281 queries.addElement(token.toUpperCase());
\r
287 // advance to next database
\r
289 } // all databases have been queries.
\r
290 if (sbuffer.length() > 0)
\r
293 .setText("Your sequences have been verified against known sequence databases. Some of the ids have been\n"
\r
294 + "altered, most likely the start/end residue will have been updated.\n"
\r
295 + "Save your alignment to maintain the updated id.\n\n"
\r
296 + sbuffer.toString());
\r
297 Desktop.addInternalFrame(output, "Sequence names updated ", 600, 300);
\r
298 // The above is the dataset, we must now find out the index
\r
299 // of the viewed sequence
\r
303 af.setProgressBar("DBRef search completed", startTime);
\r
304 // promptBeforeBlast();
\r
311 * Verify local sequences in seqRefs against the retrieved sequence database
\r
315 void transferReferences(Vector sdataset, String dbSource,
\r
316 AlignmentI retrievedAl) // File
\r
320 if (retrievedAl == null || retrievedAl.getHeight() == 0)
\r
324 SequenceI[] retrieved = retrievedAl.getSequencesArray();
\r
325 SequenceI sequence = null;
\r
327 // Vector entries = new Uniprot().getUniprotEntries(file);
\r
329 int i, iSize = retrieved.length; // entries == null ? 0 : entries.size();
\r
330 // UniprotEntry entry;
\r
331 for (i = 0; i < iSize; i++)
\r
333 SequenceI entry = retrieved[i]; // (UniprotEntry) entries.elementAt(i);
\r
335 // Work out which sequences this sequence matches,
\r
336 // taking into account all accessionIds and names in the file
\r
337 Vector sequenceMatches = new Vector();
\r
338 // look for corresponding accession ids
\r
339 DBRefEntry[] entryRefs = jalview.util.DBRefUtils.selectRefs(entry
\r
340 .getDBRef(), new String[]
\r
342 for (int j = 0; j < entryRefs.length; j++)
\r
344 String accessionId = entryRefs[j].getAccessionId(); // .getAccession().elementAt(j).toString();
\r
345 // match up on accessionId
\r
346 if (seqRefs.containsKey(accessionId.toUpperCase()))
\r
348 Vector seqs = (Vector) seqRefs.get(accessionId);
\r
349 for (int jj = 0; jj < seqs.size(); jj++)
\r
351 sequence = (SequenceI) seqs.elementAt(jj);
\r
352 if (!sequenceMatches.contains(sequence))
\r
354 sequenceMatches.addElement(sequence);
\r
359 if (sequenceMatches.size() == 0)
\r
361 // failed to match directly on accessionId==query so just compare all
\r
362 // sequences to entry
\r
363 Enumeration e = seqRefs.keys();
\r
364 while (e.hasMoreElements())
\r
366 Vector sqs = (Vector) seqRefs.get(e.nextElement());
\r
367 if (sqs != null && sqs.size() > 0)
\r
369 Enumeration sqe = sqs.elements();
\r
370 while (sqe.hasMoreElements())
\r
372 sequenceMatches.addElement(sqe.nextElement());
\r
377 // look for corresponding names
\r
378 // this is uniprot specific ?
\r
379 // could be useful to extend this so we try to find any 'significant'
\r
380 // information in common between two sequence objects.
\r
382 * DBRefEntry[] entryRefs =
\r
383 * jalview.util.DBRefUtils.selectRefs(entry.getDBRef(), new String[] {
\r
384 * dbSource }); for (int j = 0; j < entry.getName().size(); j++) { String
\r
385 * name = entry.getName().elementAt(j).toString(); if
\r
386 * (seqRefs.containsKey(name)) { Vector seqs = (Vector) seqRefs.get(name);
\r
387 * for (int jj = 0; jj < seqs.size(); jj++) { sequence = (SequenceI)
\r
388 * seqs.elementAt(jj); if (!sequenceMatches.contains(sequence)) {
\r
389 * sequenceMatches.addElement(sequence); } } } }
\r
391 // sequenceMatches now contains the set of all sequences associated with
\r
392 // the returned db record
\r
393 String entrySeq = entry.getSequenceAsString().toUpperCase();
\r
394 for (int m = 0; m < sequenceMatches.size(); m++)
\r
396 sequence = (SequenceI) sequenceMatches.elementAt(m);
\r
397 // only update start and end positions and shift features if there are
\r
398 // no existing references
\r
399 // TODO: test for legacy where uniprot or EMBL refs exist but no
\r
400 // mappings are made (but content matches retrieved set)
\r
401 boolean updateRefFrame = sequence.getDBRef() == null
\r
402 || sequence.getDBRef().length == 0;
\r
403 // verify sequence against the entry sequence
\r
405 String nonGapped = AlignSeq.extractGaps("-. ",
\r
406 sequence.getSequenceAsString()).toUpperCase();
\r
408 int absStart = entrySeq.indexOf(nonGapped);
\r
409 int mapStart = entry.getStart();
\r
410 jalview.datamodel.Mapping mp;
\r
412 if (absStart == -1)
\r
414 // Is local sequence contained in dataset sequence?
\r
415 absStart = nonGapped.indexOf(entrySeq);
\r
416 if (absStart == -1)
\r
417 { // verification failed.
\r
418 sbuffer.append(sequence.getName()
\r
419 + " SEQUENCE NOT %100 MATCH \n");
\r
423 sbuffer.append(sequence.getName() + " HAS " + absStart
\r
424 + " PREFIXED RESIDUES COMPARED TO " + dbSource + "\n");
\r
426 // + " - ANY SEQUENCE FEATURES"
\r
427 // + " HAVE BEEN ADJUSTED ACCORDINGLY \n");
\r
429 // create valid mapping between matching region of local sequence and
\r
430 // the mapped sequence
\r
431 mp = new Mapping(null, new int[]
\r
432 { sequence.getStart() + absStart,
\r
433 sequence.getStart() + absStart + entrySeq.length() - 1 },
\r
435 { entry.getStart(),
\r
436 entry.getStart() + entrySeq.length() - 1 }, 1, 1);
\r
437 updateRefFrame = false; // mapping is based on current start/end so
\r
438 // don't modify start and end
\r
442 // update start and end of local sequence to place it in entry's
\r
443 // reference frame.
\r
444 // apply identity map map from whole of local sequence to matching
\r
445 // region of database
\r
447 mp = null; // Mapping.getIdentityMap();
\r
448 // new Mapping(null,
\r
449 // new int[] { absStart+sequence.getStart(),
\r
450 // absStart+sequence.getStart()+entrySeq.length()-1},
\r
451 // new int[] { entry.getStart(), entry.getEnd() }, 1, 1);
\r
452 // relocate local features for updated start
\r
453 if (updateRefFrame && sequence.getSequenceFeatures() != null)
\r
455 SequenceFeature[] sf = sequence.getSequenceFeatures();
\r
456 int start = sequence.getStart();
\r
457 int end = sequence.getEnd();
\r
458 int startShift = 1 - absStart - start; // how much the features are
\r
459 // to be shifted by
\r
460 for (int sfi = 0; sfi < sf.length; sfi++)
\r
462 if (sf[sfi].getBegin() >= start && sf[sfi].getEnd() <= end)
\r
464 // shift feature along by absstart
\r
465 sf[sfi].setBegin(sf[sfi].getBegin() + startShift);
\r
466 sf[sfi].setEnd(sf[sfi].getEnd() + startShift);
\r
472 System.out.println("Adding dbrefs to " + sequence.getName()
\r
473 + " from " + dbSource + " sequence : " + entry.getName());
\r
474 sequence.transferAnnotation(entry, mp);
\r
475 // unknownSequences.remove(sequence);
\r
476 int absEnd = absStart + nonGapped.length();
\r
478 if (updateRefFrame)
\r
480 // finally, update local sequence reference frame if we're allowed
\r
481 sequence.setStart(absStart);
\r
482 sequence.setEnd(absEnd);
\r
484 // and remove it from the rest
\r
485 // TODO: decide if we should remove annotated sequence from set
\r
486 sdataset.remove(sequence);
\r