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.
21 package jalview.analysis;
23 import jalview.datamodel.AlignedCodonFrame;
24 import jalview.datamodel.Alignment;
25 import jalview.datamodel.AlignmentI;
26 import jalview.datamodel.DBRefEntry;
27 import jalview.datamodel.Mapping;
28 import jalview.datamodel.Sequence;
29 import jalview.datamodel.SequenceFeature;
30 import jalview.datamodel.SequenceI;
31 import jalview.util.Comparison;
32 import jalview.util.DBRefUtils;
33 import jalview.util.MapList;
34 import jalview.ws.SequenceFetcherFactory;
35 import jalview.ws.seqfetcher.ASequenceFetcher;
37 import java.util.ArrayList;
38 import java.util.Arrays;
39 import java.util.List;
42 * Functions for cross-referencing sequence databases. user must first specify
43 * if cross-referencing from protein or dna (set dna==true)
51 * A sub-class that ignores Parent attribute when comparing sequence
52 * features. This avoids 'duplicate' CDS features that only
53 * differ in their parent Transcript ids.
55 class MySequenceFeature extends SequenceFeature
57 private SequenceFeature feat;
59 MySequenceFeature(SequenceFeature sf)
65 public boolean equals(Object o)
67 return feat.equals(o, true);
73 * Returns a list of distinct database sources for which sequences have either
75 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
76 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
77 * reference from another sequence in the dataset which has a cross-reference
78 * to a direct DBRefEntry on the given sequence</li>
82 * true if seqs are nucleotide
84 * sequences whose xrefs we are seeking
86 * an alignment to search for indirect references
89 public static List<String> findXrefSourcesForSequences(boolean dna,
90 SequenceI[] seqs, AlignmentI dataset)
92 List<String> sources = new ArrayList<String>();
93 for (SequenceI seq : seqs)
97 findXrefSourcesForSequence(seq, dna, dataset, sources);
104 * Returns a list of distinct database sources for which a sequence has either
106 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
107 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
108 * reference from another sequence in the dataset which has a cross-reference
109 * to a direct DBRefEntry on the given sequence</li>
113 * the sequence whose dbrefs we are searching against
115 * true if the sequence is nucleotide
117 * an alignment to search for indirect references
119 * a list of sources to add matches to
121 static void findXrefSourcesForSequence(SequenceI seq, boolean dna,
122 AlignmentI dataset, List<String> sources)
125 * first find seq's xrefs (dna-to-peptide or peptide-to-dna)
127 DBRefEntry[] rfs = DBRefUtils.selectDbRefs(!dna, seq.getDBRefs());
128 addXrefsToSources(rfs, sources);
132 * find sequence's direct (dna-to-dna, peptide-to-peptide) xrefs
134 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(dna, seq.getDBRefs());
135 List<SequenceI> rseqs = new ArrayList<SequenceI>();
138 * find sequences in the alignment which xref one of these DBRefs
139 * i.e. is xref-ed to a common sequence identifier
141 CrossRef.searchDatasetXrefs(seq, !dna, lrfs, dataset, rseqs, null);
144 * add those sequences' (dna-to-peptide or peptide-to-dna) dbref sources
146 for (SequenceI rs : rseqs)
148 DBRefEntry[] xrs = DBRefUtils.selectDbRefs(!dna, rs.getDBRefs());
149 addXrefsToSources(xrs, sources);
155 * Helper method that adds the source identifiers of some cross-references to
156 * a (non-redundant) list of database sources
161 static void addXrefsToSources(DBRefEntry[] xrefs, List<String> sources)
165 for (DBRefEntry ref : xrefs)
167 String source = ref.getSource();
168 if (!sources.contains(source))
179 * sequences whose xrefs are being retrieved
181 * true if sequences are nucleotide
184 * alignment to search for cross-referenced sequences (and possibly
186 * @return products (as dataset sequences)
188 public static Alignment findXrefSequences(SequenceI[] seqs,
189 final boolean dna, final String source, AlignmentI al)
191 AlignmentI dataset = al.getDataset() == null ? al : al.getDataset();
192 List<SequenceI> rseqs = new ArrayList<SequenceI>();
193 AlignedCodonFrame cf = new AlignedCodonFrame();
194 for (SequenceI seq : seqs)
197 while (dss.getDatasetSequence() != null)
199 dss = dss.getDatasetSequence();
201 boolean found = false;
202 DBRefEntry[] xrfs = DBRefUtils.selectDbRefs(!dna, dss.getDBRefs());
203 if ((xrfs == null || xrfs.length == 0) && dataset != null)
206 * found no suitable dbrefs on sequence - look for sequences in the
207 * alignment which share a dbref with this one
209 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(dna, seq.getDBRefs());
212 * find sequences (except this one!), of complementary type,
213 * which have a dbref to an accession id for this sequence,
214 * and add them to the results
216 found = CrossRef.searchDatasetXrefs(dss, !dna, lrfs, dataset,
219 for (int r = 0; xrfs != null && r < xrfs.length; r++)
221 DBRefEntry xref = xrfs[r];
222 if (source != null && !source.equals(xref.getSource()))
228 if (xref.getMap().getTo() != null)
231 SequenceI rsq = new Sequence(xref.getMap().getTo());
233 if (xref.getMap().getMap().getFromRatio() != xref
234 .getMap().getMap().getToRatio())
236 // get sense of map correct for adding to product alignment.
239 // map is from dna seq to a protein product
240 cf.addMap(dss, rsq, xref.getMap().getMap());
244 // map should be from protein seq to its coding dna
245 cf.addMap(rsq, dss, xref.getMap().getMap().getInverse());
252 // do a bit more work - search for sequences with references matching
253 // xrefs on this sequence.
256 found = searchDataset(dss, xref, dataset, rseqs, cf, false,/*true?*/
260 xrfs[r] = null; // we've recovered seqs for this one.
267 if (xrfs != null && xrfs.length > 0)
269 ASequenceFetcher sftch = SequenceFetcherFactory
270 .getSequenceFetcher();
271 SequenceI[] retrieved = null;
273 for (int r = 0; r < xrfs.length; r++)
275 // filter out any irrelevant or irretrievable references
277 || ((source != null && !source.equals(xrfs[r]
278 .getSource())) || !sftch.isFetchable(xrfs[r]
288 // .println("Attempting to retrieve cross referenced sequences.");
289 DBRefEntry[] t = new DBRefEntry[l];
291 for (int r = 0; r < xrfs.length; r++)
301 retrieved = sftch.getSequences(Arrays.asList(xrfs), !dna);
302 // problem here is we don't know which of xrfs resulted in which
304 } catch (Exception e)
307 .println("Problem whilst retrieving cross references for Sequence : "
312 if (retrieved != null)
314 updateDbrefMappings(dna, seq, xrfs, retrieved, cf);
316 SequenceIdMatcher matcher = new SequenceIdMatcher(
317 dataset.getSequences());
318 List<SequenceFeature> copiedFeatures = new ArrayList<SequenceFeature>();
319 CrossRef me = new CrossRef();
320 for (int rs = 0; rs < retrieved.length; rs++)
322 // TODO: examine each sequence for 'redundancy'
323 DBRefEntry[] dbr = retrieved[rs].getDBRefs();
324 if (dbr != null && dbr.length > 0)
326 for (int di = 0; di < dbr.length; di++)
328 // find any entry where we should put in the sequence being
329 // cross-referenced into the map
330 Mapping map = dbr[di].getMap();
333 if (map.getTo() != null && map.getMap() != null)
335 SequenceI matched = matcher
336 .findIdMatch(map.getTo());
340 * already got an xref to this sequence; update this
341 * map to point to the same sequence, and add
342 * any new dbrefs to it
344 for (DBRefEntry ref : map.getTo().getDBRefs())
346 matched.addDBRef(ref); // add or update mapping
352 matcher.add(map.getTo());
356 // compare ms with dss and replace with dss in mapping
357 // if map is congruent
358 SequenceI ms = map.getTo();
359 int sf = map.getMap().getToLowest();
360 int st = map.getMap().getToHighest();
361 SequenceI mappedrg = ms.getSubSequence(sf, st);
362 // SequenceI loc = dss.getSubSequence(sf, st);
363 if (mappedrg.getLength() > 0
364 && ms.getSequenceAsString().equals(
365 dss.getSequenceAsString()))
366 // && mappedrg.getSequenceAsString().equals(
367 // loc.getSequenceAsString()))
369 String msg = "Mapping updated from "
371 + " to retrieved crossreference "
373 System.out.println(msg);
374 // method to update all refs of existing To on
375 // retrieved sequence with dss and merge any props
379 * copy sequence features as well, avoiding
380 * duplication (e.g. same variation from 2
383 SequenceFeature[] sfs = ms
384 .getSequenceFeatures();
387 for (SequenceFeature feat : sfs)
390 * we override SequenceFeature.equals here (but
391 * not elsewhere) to ignore Parent attribute
392 * TODO not quite working yet!
395 .contains(me.new MySequenceFeature(
398 dss.addSequenceFeature(feat);
399 copiedFeatures.add(feat);
403 cf.addMap(retrieved[rs].getDatasetSequence(),
406 // TODO remove this 'else' and the cf.addMap above?
409 cf.addMap(retrieved[rs].getDatasetSequence(),
410 map.getTo(), map.getMap());
412 } catch (Exception e)
415 .println("Exception when consolidating Mapped sequence set...");
416 e.printStackTrace(System.err);
422 retrieved[rs].updatePDBIds();
423 rseqs.add(retrieved[rs]);
431 Alignment ral = null;
432 if (rseqs.size() > 0)
434 ral = new Alignment(rseqs.toArray(new SequenceI[rseqs.size()]));
435 if (cf != null && !cf.isEmpty())
437 ral.addCodonFrame(cf);
444 * Updates any empty mappings in the cross-references with one to a compatible
445 * retrieved sequence if found, and adds any new mappings to the
454 static void updateDbrefMappings(boolean dna, SequenceI mapFrom,
455 DBRefEntry[] xrefs, SequenceI[] retrieved, AlignedCodonFrame acf)
457 SequenceIdMatcher matcher = new SequenceIdMatcher(retrieved);
458 for (DBRefEntry xref : xrefs)
462 String targetSeqName = xref.getSource() + "|"
463 + xref.getAccessionId();
464 SequenceI[] matches = matcher.findAllIdMatches(targetSeqName);
469 for (SequenceI seq : matches)
471 MapList mapping = null;
474 mapping = AlignmentUtils.mapCdnaToProtein(seq, mapFrom);
478 mapping = AlignmentUtils.mapCdnaToProtein(mapFrom, seq);
481 mapping = mapping.getInverse();
486 xref.setMap(new Mapping(seq, mapping));
489 AlignmentUtils.computeProteinFeatures(mapFrom, seq, mapping);
493 acf.addMap(mapFrom, seq, mapping);
497 acf.addMap(seq, mapFrom, mapping.getInverse());
507 * find references to lrfs in the cross-reference set of each sequence in
508 * dataset (that is not equal to sequenceI) Identifies matching DBRefEntry
509 * based on source and accession string only - Map and Version are nulled.
515 * @return true if matches were found.
517 private static boolean searchDatasetXrefs(SequenceI sequenceI,
518 boolean dna, DBRefEntry[] lrfs, AlignmentI dataset,
519 List<SequenceI> rseqs, AlignedCodonFrame cf)
521 boolean found = false;
526 for (int i = 0; i < lrfs.length; i++)
528 DBRefEntry xref = new DBRefEntry(lrfs[i]);
530 xref.setVersion(null);
532 found |= searchDataset(sequenceI, xref, dataset, rseqs, cf, false,
539 * Searches dataset for DBRefEntrys matching the given one (xrf) and adds the
540 * associated sequence to rseqs
543 * a sequence to ignore (start point of search)
545 * a cross-reference to try to match
547 * sequences to search in
549 * result list to add to
551 * a set of sequence mappings to add to
553 * - search all references or only subset
555 * search dna or protein xrefs (if direct=false)
556 * @return true if relationship found and sequence added.
558 public static boolean searchDataset(SequenceI sequenceI, DBRefEntry xrf,
559 AlignmentI dataset, List<SequenceI> rseqs, AlignedCodonFrame cf,
560 boolean direct, boolean dna)
562 boolean found = false;
567 if (dataset.getSequences() == null)
569 System.err.println("Empty dataset sequence set - NO VECTOR");
573 synchronized (ds = dataset.getSequences())
575 for (SequenceI nxt : ds)
579 if (nxt.getDatasetSequence() != null)
582 .println("Implementation warning: getProducts passed a dataset alignment without dataset sequences in it!");
584 if (nxt == sequenceI || nxt == sequenceI.getDatasetSequence())
588 // check if this is the correct sequence type
590 // TODO 'direct' is always set to false - remove?
591 // or should it be 'true' from findXrefSequences?
592 // also its Javadoc conflicts with its use:
593 // test below implies 'direct' means find complementary sequences,
594 // !direct means select same molecule type
595 boolean isDna = Comparison
596 .isNucleotide(new SequenceI[] { nxt });
597 if ((direct && isDna == dna) || (!direct && isDna != dna))
599 // skip this sequence because it is wrong molecule type
604 // look for direct or indirect references in common
605 DBRefEntry[] poss = nxt.getDBRefs();
606 List<DBRefEntry> cands = null;
608 * TODO does this make any sense?
609 * if 'direct', search the dbrefs for xrf
610 * else, filter the dbrefs by type and then search for xrf
611 * - the result is the same isn't it?
615 cands = DBRefUtils.searchRefs(poss, xrf);
619 poss = DBRefUtils.selectDbRefs(!dna, poss);
620 cands = DBRefUtils.searchRefs(poss, xrf);
622 if (!cands.isEmpty())
624 if (!rseqs.contains(nxt))
630 // don't search if we aren't given a codon map object
631 for (DBRefEntry candidate : cands)
633 Mapping mapping = candidate.getMap();
636 MapList map = mapping.getMap();
637 if (mapping.getTo() != null
638 && map.getFromRatio() != map.getToRatio())
640 // get sense of map correct for adding to product
644 // map is from dna seq to a protein product
645 cf.addMap(sequenceI, nxt, map);
649 // map should be from protein seq to its coding dna
650 cf.addMap(nxt, sequenceI, map.getInverse());
656 // TODO: add mapping between sequences if necessary