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.List;
41 * Functions for cross-referencing sequence databases. user must first specify
42 * if cross-referencing from protein or dna (set dna==true)
50 * A sub-class that ignores Parent attribute when comparing sequence
51 * features. This avoids 'duplicate' CDS features that only
52 * differ in their parent Transcript ids.
54 class MySequenceFeature extends SequenceFeature
56 private SequenceFeature feat;
58 MySequenceFeature(SequenceFeature sf)
64 public boolean equals(Object o)
66 return feat.equals(o, true);
72 * Returns a list of distinct database sources for which sequences have either
74 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
75 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
76 * reference from another sequence in the dataset which has a cross-reference
77 * to a direct DBRefEntry on the given sequence</li>
81 * true if seqs are nucleotide
83 * sequences whose xrefs we are seeking
85 * an alignment to search for indirect references
88 public static List<String> findXrefSourcesForSequences(boolean dna,
89 SequenceI[] seqs, AlignmentI dataset)
91 List<String> sources = new ArrayList<String>();
92 for (SequenceI seq : seqs)
96 findXrefSourcesForSequence(seq, dna, dataset, sources);
103 * Returns a list of distinct database sources for which a sequence has either
105 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
106 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
107 * reference from another sequence in the dataset which has a cross-reference
108 * to a direct DBRefEntry on the given sequence</li>
112 * the sequence whose dbrefs we are searching against
114 * true if the sequence is nucleotide
116 * an alignment to search for indirect references
118 * a list of sources to add matches to
120 static void findXrefSourcesForSequence(SequenceI seq, boolean dna,
121 AlignmentI dataset, List<String> sources)
124 * first find seq's xrefs (dna-to-peptide or peptide-to-dna)
126 DBRefEntry[] rfs = DBRefUtils.selectDbRefs(!dna, seq.getDBRefs());
127 addXrefsToSources(rfs, sources);
131 * find sequence's direct (dna-to-dna, peptide-to-peptide) xrefs
133 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(dna, seq.getDBRefs());
134 List<SequenceI> rseqs = new ArrayList<SequenceI>();
137 * find sequences in the alignment which xref one of these DBRefs
138 * i.e. is xref-ed to a common sequence identifier
140 CrossRef.searchDatasetXrefs(seq, !dna, lrfs, dataset, rseqs, null);
143 * add those sequences' (dna-to-peptide or peptide-to-dna) dbref sources
145 for (SequenceI rs : rseqs)
147 DBRefEntry[] xrs = DBRefUtils.selectDbRefs(!dna, rs.getDBRefs());
148 addXrefsToSources(xrs, sources);
154 * Helper method that adds the source identifiers of some cross-references to
155 * a (non-redundant) list of database sources
160 static void addXrefsToSources(DBRefEntry[] xrefs, List<String> sources)
164 for (DBRefEntry ref : xrefs)
166 String source = ref.getSource();
167 if (!sources.contains(source))
178 * sequences whose xrefs are being retrieved
180 * true if sequences are nucleotide
183 * alignment to search for cross-referenced sequences (and possibly
185 * @return products (as dataset sequences)
187 public static Alignment findXrefSequences(SequenceI[] seqs,
188 final boolean dna, final String source, AlignmentI al)
190 AlignmentI dataset = al.getDataset() == null ? al : al.getDataset();
191 List<SequenceI> rseqs = new ArrayList<SequenceI>();
192 AlignedCodonFrame cf = new AlignedCodonFrame();
193 for (SequenceI seq : seqs)
196 while (dss.getDatasetSequence() != null)
198 dss = dss.getDatasetSequence();
200 boolean found = false;
201 DBRefEntry[] xrfs = DBRefUtils.selectDbRefs(!dna, dss.getDBRefs());
202 if ((xrfs == null || xrfs.length == 0) && dataset != null)
205 * found no suitable dbrefs on sequence - look for sequences in the
206 * alignment which share a dbref with this one
208 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(dna, seq.getDBRefs());
211 * find sequences (except this one!), of complementary type,
212 * which have a dbref to an accession id for this sequence,
213 * and add them to the results
215 found = CrossRef.searchDatasetXrefs(dss, !dna, lrfs, dataset,
218 for (int r = 0; xrfs != null && r < xrfs.length; r++)
220 DBRefEntry xref = xrfs[r];
221 if (source != null && !source.equals(xref.getSource()))
227 if (xref.getMap().getTo() != null)
230 SequenceI rsq = new Sequence(xref.getMap().getTo());
232 if (xref.getMap().getMap().getFromRatio() != xref
233 .getMap().getMap().getToRatio())
235 // get sense of map correct for adding to product alignment.
238 // map is from dna seq to a protein product
239 cf.addMap(dss, rsq, xref.getMap().getMap());
243 // map should be from protein seq to its coding dna
244 cf.addMap(rsq, dss, xref.getMap().getMap().getInverse());
251 // do a bit more work - search for sequences with references matching
252 // xrefs on this sequence.
255 found = searchDataset(dss, xref, dataset, rseqs, cf, false,/*true?*/
259 xrfs[r] = null; // we've recovered seqs for this one.
266 if (xrfs != null && xrfs.length > 0)
268 ASequenceFetcher sftch = SequenceFetcherFactory
269 .getSequenceFetcher();
270 SequenceI[] retrieved = null;
272 for (int r = 0; r < xrfs.length; r++)
274 // filter out any irrelevant or irretrievable references
276 || ((source != null && !source.equals(xrfs[r]
277 .getSource())) || !sftch.isFetchable(xrfs[r]
287 // .println("Attempting to retrieve cross referenced sequences.");
288 DBRefEntry[] t = new DBRefEntry[l];
290 for (int r = 0; r < xrfs.length; r++)
300 retrieved = sftch.getSequences(xrfs, !dna);
301 // problem here is we don't know which of xrfs resulted in which
303 } catch (Exception e)
306 .println("Problem whilst retrieving cross references for Sequence : "
311 if (retrieved != null)
313 updateDbrefMappings(dna, seq, xrfs, retrieved, cf);
315 SequenceIdMatcher matcher = new SequenceIdMatcher(
316 dataset.getSequences());
317 List<SequenceFeature> copiedFeatures = new ArrayList<SequenceFeature>();
318 CrossRef me = new CrossRef();
319 for (int rs = 0; rs < retrieved.length; rs++)
321 // TODO: examine each sequence for 'redundancy'
322 DBRefEntry[] dbr = retrieved[rs].getDBRefs();
323 if (dbr != null && dbr.length > 0)
325 for (int di = 0; di < dbr.length; di++)
327 // find any entry where we should put in the sequence being
328 // cross-referenced into the map
329 Mapping map = dbr[di].getMap();
332 if (map.getTo() != null && map.getMap() != null)
334 SequenceI matched = matcher
335 .findIdMatch(map.getTo());
339 * already got an xref to this sequence; update this
340 * map to point to the same sequence, and add
341 * any new dbrefs to it
343 for (DBRefEntry ref : map.getTo().getDBRefs())
345 matched.addDBRef(ref); // add or update mapping
351 matcher.add(map.getTo());
355 // compare ms with dss and replace with dss in mapping
356 // if map is congruent
357 SequenceI ms = map.getTo();
358 int sf = map.getMap().getToLowest();
359 int st = map.getMap().getToHighest();
360 SequenceI mappedrg = ms.getSubSequence(sf, st);
361 // SequenceI loc = dss.getSubSequence(sf, st);
362 if (mappedrg.getLength() > 0
363 && ms.getSequenceAsString().equals(
364 dss.getSequenceAsString()))
365 // && mappedrg.getSequenceAsString().equals(
366 // loc.getSequenceAsString()))
368 String msg = "Mapping updated from "
370 + " to retrieved crossreference "
372 System.out.println(msg);
373 // method to update all refs of existing To on
374 // retrieved sequence with dss and merge any props
378 * copy sequence features as well, avoiding
379 * duplication (e.g. same variation from 2
382 SequenceFeature[] sfs = ms
383 .getSequenceFeatures();
386 for (SequenceFeature feat : sfs)
389 * we override SequenceFeature.equals here (but
390 * not elsewhere) to ignore Parent attribute
391 * TODO not quite working yet!
394 .contains(me.new MySequenceFeature(
397 dss.addSequenceFeature(feat);
398 copiedFeatures.add(feat);
402 cf.addMap(retrieved[rs].getDatasetSequence(),
405 // TODO remove this 'else' and the cf.addMap above?
408 cf.addMap(retrieved[rs].getDatasetSequence(),
409 map.getTo(), map.getMap());
411 } catch (Exception e)
414 .println("Exception when consolidating Mapped sequence set...");
415 e.printStackTrace(System.err);
421 retrieved[rs].updatePDBIds();
422 rseqs.add(retrieved[rs]);
430 Alignment ral = null;
431 if (rseqs.size() > 0)
433 ral = new Alignment(rseqs.toArray(new SequenceI[rseqs.size()]));
434 if (cf != null && !cf.isEmpty())
436 ral.addCodonFrame(cf);
443 * Updates any empty mappings in the cross-references with one to a compatible
444 * retrieved sequence if found, and adds any new mappings to the
453 static void updateDbrefMappings(boolean dna, SequenceI mapFrom,
454 DBRefEntry[] xrefs, SequenceI[] retrieved, AlignedCodonFrame acf)
456 SequenceIdMatcher matcher = new SequenceIdMatcher(retrieved);
457 for (DBRefEntry xref : xrefs)
461 String targetSeqName = xref.getSource() + "|"
462 + xref.getAccessionId();
463 SequenceI[] matches = matcher.findAllIdMatches(targetSeqName);
468 for (SequenceI seq : matches)
470 MapList mapping = null;
473 mapping = AlignmentUtils.mapCdnaToProtein(seq, mapFrom);
477 mapping = AlignmentUtils.mapCdnaToProtein(mapFrom, seq);
480 mapping = mapping.getInverse();
485 xref.setMap(new Mapping(seq, mapping));
488 AlignmentUtils.computeProteinFeatures(mapFrom, seq, mapping);
492 acf.addMap(mapFrom, seq, mapping);
496 acf.addMap(seq, mapFrom, mapping.getInverse());
506 * find references to lrfs in the cross-reference set of each sequence in
507 * dataset (that is not equal to sequenceI) Identifies matching DBRefEntry
508 * based on source and accession string only - Map and Version are nulled.
514 * @return true if matches were found.
516 private static boolean searchDatasetXrefs(SequenceI sequenceI,
517 boolean dna, DBRefEntry[] lrfs, AlignmentI dataset,
518 List<SequenceI> rseqs, AlignedCodonFrame cf)
520 boolean found = false;
525 for (int i = 0; i < lrfs.length; i++)
527 DBRefEntry xref = new DBRefEntry(lrfs[i]);
529 xref.setVersion(null);
531 found |= searchDataset(sequenceI, xref, dataset, rseqs, cf, false,
538 * Searches dataset for DBRefEntrys matching the given one (xrf) and adds the
539 * associated sequence to rseqs
542 * a sequence to ignore (start point of search)
544 * a cross-reference to try to match
546 * sequences to search in
548 * result list to add to
550 * a set of sequence mappings to add to
552 * - search all references or only subset
554 * search dna or protein xrefs (if direct=false)
555 * @return true if relationship found and sequence added.
557 public static boolean searchDataset(SequenceI sequenceI, DBRefEntry xrf,
558 AlignmentI dataset, List<SequenceI> rseqs, AlignedCodonFrame cf,
559 boolean direct, boolean dna)
561 boolean found = false;
566 if (dataset.getSequences() == null)
568 System.err.println("Empty dataset sequence set - NO VECTOR");
572 synchronized (ds = dataset.getSequences())
574 for (SequenceI nxt : ds)
578 if (nxt.getDatasetSequence() != null)
581 .println("Implementation warning: getProducts passed a dataset alignment without dataset sequences in it!");
583 if (nxt == sequenceI || nxt == sequenceI.getDatasetSequence())
587 // check if this is the correct sequence type
589 // TODO 'direct' is always set to false - remove?
590 // or should it be 'true' from findXrefSequences?
591 // also its Javadoc conflicts with its use:
592 // test below implies 'direct' means find complementary sequences,
593 // !direct means select same molecule type
594 boolean isDna = Comparison
595 .isNucleotide(new SequenceI[] { nxt });
596 if ((direct && isDna == dna) || (!direct && isDna != dna))
598 // skip this sequence because it is wrong molecule type
603 // look for direct or indirect references in common
604 DBRefEntry[] poss = nxt.getDBRefs();
605 DBRefEntry[] cands = null;
607 * TODO does this make any sense?
608 * if 'direct', search the dbrefs for xrf
609 * else, filter the dbrefs by type and then search for xrf
610 * - the result is the same isn't it?
614 cands = DBRefUtils.searchRefs(poss, xrf);
618 poss = DBRefUtils.selectDbRefs(!dna, poss);
619 cands = DBRefUtils.searchRefs(poss, xrf);
623 if (!rseqs.contains(nxt))
627 boolean foundmap = cf != null;
628 // don't search if we aren't given a codon map object
629 for (int r = 0; foundmap && r < cands.length; r++)
631 if (cands[r].hasMap())
633 Mapping mapping = cands[r].getMap();
634 MapList map = mapping.getMap();
635 if (mapping.getTo() != null
636 && map.getFromRatio() != map.getToRatio())
638 // get sense of map correct for adding to product
642 // map is from dna seq to a protein product
643 cf.addMap(sequenceI, nxt, map);
647 // map should be from protein seq to its coding dna
648 cf.addMap(nxt, sequenceI, map.getInverse());
653 // TODO: add mapping between sequences if necessary