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.DBRefSource;
28 import jalview.datamodel.Mapping;
29 import jalview.datamodel.Sequence;
30 import jalview.datamodel.SequenceFeature;
31 import jalview.datamodel.SequenceI;
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.Iterator;
39 import java.util.List;
42 * Functions for cross-referencing sequence databases.
50 * the dataset of the alignment for which we are searching for
51 * cross-references; in some cases we may resolve xrefs by
52 * searching in the dataset
54 private AlignmentI dataset;
57 * the sequences for which we are seeking cross-references
59 private SequenceI[] fromSeqs;
62 * matcher built from dataset
64 SequenceIdMatcher matcher;
67 * sequences found by cross-ref searches to fromSeqs
69 List<SequenceI> rseqs;
75 * the sequences for which we are seeking cross-references
77 * the containing alignment dataset (may be searched to resolve
80 public CrossRef(SequenceI[] seqs, AlignmentI ds)
83 dataset = ds.getDataset() == null ? ds : ds.getDataset();
87 * Returns a list of distinct database sources for which sequences have either
89 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
90 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
91 * reference from another sequence in the dataset which has a cross-reference
92 * to a direct DBRefEntry on the given sequence</li>
96 * - when true, cross-references *from* dna returned. When false,
97 * cross-references *from* protein are returned
100 public List<String> findXrefSourcesForSequences(boolean dna)
102 List<String> sources = new ArrayList<>();
103 for (SequenceI seq : fromSeqs)
107 findXrefSourcesForSequence(seq, dna, sources);
111 // hack to prevent EMBL xrefs resulting in redundant datasets
112 sources.remove(DBRefSource.EMBL);
116 // hack to prevent Ensembl xref option shown from cdna panel
117 sources.remove(DBRefSource.ENSEMBL);
123 * Returns a list of distinct database sources for which a sequence has either
125 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
126 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
127 * reference from another sequence in the dataset which has a cross-reference
128 * to a direct DBRefEntry on the given sequence</li>
132 * the sequence whose dbrefs we are searching against
134 * when true, context is DNA - so sources identifying protein
135 * products will be returned.
137 * a list of sources to add matches to
139 void findXrefSourcesForSequence(SequenceI seq, boolean fromDna,
140 List<String> sources)
143 * first find seq's xrefs (dna-to-peptide or peptide-to-dna)
145 DBRefEntry[] rfs = DBRefUtils.selectDbRefs(!fromDna, seq.getDBRefs());
146 addXrefsToSources(rfs, sources);
150 * find sequence's direct (dna-to-dna, peptide-to-peptide) xrefs
152 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(fromDna, seq.getDBRefs());
153 List<SequenceI> foundSeqs = new ArrayList<>();
156 * find sequences in the alignment which xref one of these DBRefs
157 * i.e. is xref-ed to a common sequence identifier
159 searchDatasetXrefs(fromDna, seq, lrfs, foundSeqs, null);
162 * add those sequences' (dna-to-peptide or peptide-to-dna) dbref sources
164 for (SequenceI rs : foundSeqs)
166 DBRefEntry[] xrs = DBRefUtils.selectDbRefs(!fromDna,
168 addXrefsToSources(xrs, sources);
174 * Helper method that adds the source identifiers of some cross-references to
175 * a (non-redundant) list of database sources
180 void addXrefsToSources(DBRefEntry[] xrefs, List<String> sources)
184 for (DBRefEntry ref : xrefs)
187 * avoid duplication e.g. ENSEMBL and Ensembl
189 String source = DBRefUtils.getCanonicalName(ref.getSource());
190 if (!sources.contains(source))
199 * Attempts to find cross-references from the sequences provided in the
200 * constructor to the given source database. Cross-references may be found
202 * <li>in dbrefs on the sequence which hold a mapping to a sequence
204 * <li>provided with a fetched sequence (e.g. ENA translation), or</li>
205 * <li>populated previously after getting cross-references</li>
207 * <li>as other sequences in the alignment which share a dbref identifier with
209 * <li>by fetching from the remote database</li>
211 * The cross-referenced sequences, and mappings to them, are added to the
215 * @return cross-referenced sequences (as dataset sequences)
217 public Alignment findXrefSequences(String source, boolean fromDna)
220 rseqs = new ArrayList<>();
221 AlignedCodonFrame cf = new AlignedCodonFrame();
222 matcher = new SequenceIdMatcher(dataset.getSequences());
224 for (SequenceI seq : fromSeqs)
227 while (dss.getDatasetSequence() != null)
229 dss = dss.getDatasetSequence();
231 boolean found = false;
232 DBRefEntry[] xrfs = DBRefUtils.selectDbRefs(!fromDna,
234 // ENST & ENSP comes in to both Protein and nucleotide, so we need to
237 if ((xrfs == null || xrfs.length == 0) && dataset != null)
240 * found no suitable dbrefs on sequence - look for sequences in the
241 * alignment which share a dbref with this one
243 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(fromDna,
247 * find sequences (except this one!), of complementary type,
248 * which have a dbref to an accession id for this sequence,
249 * and add them to the results
251 found = searchDatasetXrefs(fromDna, dss, lrfs, rseqs, cf);
253 if (xrfs == null && !found)
256 * no dbref to source on this sequence or matched
257 * complementary sequence in the dataset
261 List<DBRefEntry> sourceRefs = DBRefUtils.searchRefsForSource(xrfs,
263 Iterator<DBRefEntry> refIterator = sourceRefs.iterator();
264 // At this point, if we are retrieving Ensembl, we still don't filter out
265 // ENST when looking for protein crossrefs.
266 while (refIterator.hasNext())
268 DBRefEntry xref = refIterator.next();
270 // we're only interested in coding cross-references, not
272 if (xref.hasMap() && xref.getMap().getMap().isTripletMap())
274 SequenceI mappedTo = xref.getMap().getTo();
275 if (mappedTo != null)
278 * dbref contains the sequence it maps to; add it to the
279 * results unless we have done so already (could happen if
280 * fetching xrefs for sequences which have xrefs in common)
281 * for example: UNIPROT {P0CE19, P0CE20} -> EMBL {J03321, X06707}
285 * problem: matcher.findIdMatch() is lenient - returns a sequence
286 * with a dbref to the search arg e.g. ENST for ENSP - wrong
287 * but findInDataset() matches ENSP when looking for Uniprot...
289 SequenceI matchInDataset = findInDataset(xref);
290 if (matchInDataset != null && xref.getMap().getTo() != null
291 && matchInDataset != xref.getMap().getTo())
294 "Implementation problem (reopen JAL-2154): CrossRef.findInDataset seems to have recovered a different sequence than the one explicitly mapped for xref."
295 + "Found:" + matchInDataset + "\nExpected:"
296 + xref.getMap().getTo() + "\nFor xref:"
299 /*matcher.findIdMatch(mappedTo);*/
300 if (matchInDataset != null)
302 if (!rseqs.contains(matchInDataset))
304 rseqs.add(matchInDataset);
306 // even if rseqs contained matchInDataset - check mappings between
307 // these seqs are added
308 // need to try harder to only add unique mappings
309 if (xref.getMap().getMap().isTripletMap()
310 && dataset.getMapping(seq, matchInDataset) == null
311 && cf.getMappingBetween(seq, matchInDataset) == null)
313 // materialise a mapping for highlighting between these
317 cf.addMap(dss, matchInDataset, xref.getMap().getMap(),
318 xref.getMap().getMappedFromId());
322 cf.addMap(matchInDataset, dss,
323 xref.getMap().getMap().getInverse(),
324 xref.getMap().getMappedFromId());
328 refIterator.remove();
331 // TODO: need to determine if this should be a deriveSequence
332 SequenceI rsq = new Sequence(mappedTo);
334 if (xref.getMap().getMap().isTripletMap())
336 // get sense of map correct for adding to product alignment.
339 // map is from dna seq to a protein product
340 cf.addMap(dss, rsq, xref.getMap().getMap(),
341 xref.getMap().getMappedFromId());
345 // map should be from protein seq to its coding dna
346 cf.addMap(rsq, dss, xref.getMap().getMap().getInverse(),
347 xref.getMap().getMappedFromId());
355 SequenceI matchedSeq = matcher.findIdMatch(
356 xref.getSource() + "|" + xref.getAccessionId());
357 // if there was a match, check it's at least the right type of
359 if (matchedSeq != null && matchedSeq.isProtein() == fromDna)
361 if (constructMapping(seq, matchedSeq, xref, cf, fromDna))
370 // do a bit more work - search for sequences with references matching
371 // xrefs on this sequence.
372 found = searchDataset(fromDna, dss, xref, rseqs, cf, false);
376 refIterator.remove();
381 * fetch from source database any dbrefs we haven't resolved up to here
383 if (!sourceRefs.isEmpty())
385 retrieveCrossRef(sourceRefs, seq, xrfs, fromDna, cf);
389 Alignment ral = null;
390 if (rseqs.size() > 0)
392 ral = new Alignment(rseqs.toArray(new SequenceI[rseqs.size()]));
395 dataset.addCodonFrame(cf);
401 private void retrieveCrossRef(List<DBRefEntry> sourceRefs, SequenceI seq,
402 DBRefEntry[] xrfs, boolean fromDna, AlignedCodonFrame cf)
404 ASequenceFetcher sftch = SequenceFetcherFactory.getSequenceFetcher();
405 SequenceI[] retrieved = null;
406 SequenceI dss = seq.getDatasetSequence() == null ? seq
407 : seq.getDatasetSequence();
408 // first filter in case we are retrieving crossrefs that have already been
409 // retrieved. this happens for cases where a database record doesn't yield
410 // protein products for CDS
411 removeAlreadyRetrievedSeqs(sourceRefs, fromDna);
412 if (sourceRefs.size() == 0)
414 // no more work to do! We already had all requested sequence records in
420 retrieved = sftch.getSequences(sourceRefs, !fromDna);
421 } catch (Exception e)
424 "Problem whilst retrieving cross references for Sequence : "
429 if (retrieved != null)
431 boolean addedXref = false;
432 List<SequenceI> newDsSeqs = new ArrayList<>(),
433 doNotAdd = new ArrayList<>();
435 for (SequenceI retrievedSequence : retrieved)
437 // dataset gets contaminated ccwith non-ds sequences. why ??!
438 // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL->
439 SequenceI retrievedDss = retrievedSequence
440 .getDatasetSequence() == null ? retrievedSequence
441 : retrievedSequence.getDatasetSequence();
442 addedXref |= importCrossRefSeq(cf, newDsSeqs, doNotAdd, dss,
447 // try again, after looking for matching IDs
448 // shouldn't need to do this unless the dbref mechanism has broken.
449 updateDbrefMappings(seq, xrfs, retrieved, cf, fromDna);
450 for (SequenceI retrievedSequence : retrieved)
452 // dataset gets contaminated ccwith non-ds sequences. why ??!
453 // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL->
454 SequenceI retrievedDss = retrievedSequence
455 .getDatasetSequence() == null ? retrievedSequence
456 : retrievedSequence.getDatasetSequence();
457 addedXref |= importCrossRefSeq(cf, newDsSeqs, doNotAdd, dss,
461 for (SequenceI newToSeq : newDsSeqs)
463 if (!doNotAdd.contains(newToSeq)
464 && dataset.findIndex(newToSeq) == -1)
466 dataset.addSequence(newToSeq);
467 matcher.add(newToSeq);
474 * Search dataset for sequences with a primary reference contained in
478 * - list of references to filter.
480 * - type of sequence to search for matching primary reference.
482 private void removeAlreadyRetrievedSeqs(List<DBRefEntry> sourceRefs,
485 DBRefEntry[] dbrSourceSet = sourceRefs.toArray(new DBRefEntry[0]);
486 for (SequenceI sq : dataset.getSequences())
488 boolean dupeFound = false;
489 // !fromDna means we are looking only for nucleotide sequences, not
491 if (sq.isProtein() == fromDna)
493 for (DBRefEntry dbr : sq.getPrimaryDBRefs())
495 for (DBRefEntry found : DBRefUtils.searchRefs(dbrSourceSet, dbr))
497 sourceRefs.remove(found);
504 // rebuild the search array from the filtered sourceRefs list
505 dbrSourceSet = sourceRefs.toArray(new DBRefEntry[0]);
511 * process sequence retrieved via a dbref on source sequence to resolve and
515 * @param sourceSequence
516 * @param retrievedSequence
517 * @return true if retrieveSequence was imported
519 private boolean importCrossRefSeq(AlignedCodonFrame cf,
520 List<SequenceI> newDsSeqs, List<SequenceI> doNotAdd,
521 SequenceI sourceSequence, SequenceI retrievedSequence)
524 * set when retrievedSequence has been verified as a crossreference for
527 boolean imported = false;
528 DBRefEntry[] dbr = retrievedSequence.getDBRefs();
531 for (DBRefEntry dbref : dbr)
533 SequenceI matched = findInDataset(dbref);
534 if (matched == sourceSequence)
536 // verified retrieved and source sequence cross-reference each other
539 // find any entry where we should put in the sequence being
540 // cross-referenced into the map
541 Mapping map = dbref.getMap();
544 if (map.getTo() != null && map.getMap() != null)
546 if (map.getTo() == sourceSequence)
548 // already called to import once, and most likely this sequence
549 // already imported !
555 * sequence is new to dataset, so save a reference so it can be added.
557 newDsSeqs.add(map.getTo());
562 * there was a matching sequence in dataset, so now, check to see if we can update the map.getTo() sequence to the existing one.
567 // compare ms with dss and replace with dss in mapping
568 // if map is congruent
569 SequenceI ms = map.getTo();
570 // TODO findInDataset requires exact sequence match but
571 // 'congruent' test is only for the mapped part
572 // maybe not a problem in practice since only ENA provide a
573 // mapping and it is to the full protein translation of CDS
574 // matcher.findIdMatch(map.getTo());
575 // TODO addendum: if matched is shorter than getTo, this will fail
576 // - when it should really succeed.
577 int sf = map.getMap().getToLowest();
578 int st = map.getMap().getToHighest();
579 SequenceI mappedrg = ms.getSubSequence(sf, st);
580 if (mappedrg.getLength() > 0 && ms.getSequenceAsString()
581 .equals(matched.getSequenceAsString()))
584 * sequences were a match,
586 String msg = "Mapping updated from " + ms.getName()
587 + " to retrieved crossreference "
589 System.out.println(msg);
591 DBRefEntry[] toRefs = map.getTo().getDBRefs();
595 * transfer database refs
597 for (DBRefEntry ref : toRefs)
599 if (dbref.getSrcAccString()
600 .equals(ref.getSrcAccString()))
602 continue; // avoid overwriting the ref on source sequence
604 matched.addDBRef(ref); // add or update mapping
607 doNotAdd.add(map.getTo());
611 * give the reverse reference the inverse mapping
612 * (if it doesn't have one already)
614 setReverseMapping(matched, dbref, cf);
617 * copy sequence features as well, avoiding
618 * duplication (e.g. same variation from two
621 List<SequenceFeature> sfs = ms.getFeatures()
623 for (SequenceFeature feat : sfs)
626 * make a flyweight feature object which ignores Parent
627 * attribute in equality test; this avoids creating many
628 * otherwise duplicate exon features on genomic sequence
630 SequenceFeature newFeature = new SequenceFeature(feat)
633 public boolean equals(Object o)
635 return super.equals(o, true);
638 matched.addSequenceFeature(newFeature);
641 cf.addMap(retrievedSequence, map.getTo(), map.getMap());
642 } catch (Exception e)
645 "Exception when consolidating Mapped sequence set...");
646 e.printStackTrace(System.err);
654 retrievedSequence.updatePDBIds();
655 rseqs.add(retrievedSequence);
656 if (dataset.findIndex(retrievedSequence) == -1)
658 dataset.addSequence(retrievedSequence);
659 matcher.add(retrievedSequence);
666 * Sets the inverse sequence mapping in the corresponding dbref of the mapped
667 * to sequence (if any). This is used after fetching a cross-referenced
668 * sequence, if the fetched sequence has a mapping to the original sequence,
669 * to set the mapping in the original sequence's dbref.
672 * the sequence mapped from
676 void setReverseMapping(SequenceI mapFrom, DBRefEntry dbref,
677 AlignedCodonFrame mappings)
679 SequenceI mapTo = dbref.getMap().getTo();
684 DBRefEntry[] dbrefs = mapTo.getDBRefs();
689 for (DBRefEntry toRef : dbrefs)
691 if (toRef.hasMap() && mapFrom == toRef.getMap().getTo())
694 * found the reverse dbref; update its mapping if null
696 if (toRef.getMap().getMap() == null)
698 MapList inverse = dbref.getMap().getMap().getInverse();
699 toRef.getMap().setMap(inverse);
700 mappings.addMap(mapTo, mapFrom, inverse);
707 * Returns null or the first sequence in the dataset which is identical to
708 * xref.mapTo, and has a) a primary dbref matching xref, or if none found, the
709 * first one with an ID source|xrefacc
712 * with map and mapped-to sequence
715 SequenceI findInDataset(DBRefEntry xref)
717 if (xref == null || !xref.hasMap() || xref.getMap().getTo() == null)
721 SequenceI mapsTo = xref.getMap().getTo();
722 String name = xref.getAccessionId();
723 String name2 = xref.getSource() + "|" + name;
724 SequenceI dss = mapsTo.getDatasetSequence() == null ? mapsTo
725 : mapsTo.getDatasetSequence();
726 // first check ds if ds is directly referenced
727 if (dataset.findIndex(dss) > -1)
731 DBRefEntry template = new DBRefEntry(xref.getSource(), null,
732 xref.getAccessionId());
734 * remember the first ID match - in case we don't find a match to template
736 SequenceI firstIdMatch = null;
737 for (SequenceI seq : dataset.getSequences())
739 // first check primary refs.
740 List<DBRefEntry> match = DBRefUtils.searchRefs(
741 seq.getPrimaryDBRefs().toArray(new DBRefEntry[0]), template);
742 if (match != null && match.size() == 1 && sameSequence(seq, dss))
747 * clumsy alternative to using SequenceIdMatcher which currently
748 * returns sequences with a dbref to the matched accession id
749 * which we don't want
751 if (firstIdMatch == null && (name.equals(seq.getName())
752 || seq.getName().startsWith(name2)))
754 if (sameSequence(seq, dss))
764 * Answers true if seq1 and seq2 contain exactly the same characters (ignoring
765 * case), else false. This method compares the lengths, then each character in
766 * turn, in order to 'fail fast'. For case-sensitive comparison, it would be
767 * possible to use Arrays.equals(seq1.getSequence(), seq2.getSequence()).
773 // TODO move to Sequence / SequenceI
774 static boolean sameSequence(SequenceI seq1, SequenceI seq2)
780 if (seq1 == null || seq2 == null)
785 if (seq1.getLength() != seq2.getLength())
789 int length = seq1.getLength();
790 for (int i = 0; i < length; i++)
792 int diff = seq1.getCharAt(i) - seq2.getCharAt(i);
794 * same char or differ in case only ('a'-'A' == 32)
796 if (diff != 0 && diff != 32 && diff != -32)
805 * Updates any empty mappings in the cross-references with one to a compatible
806 * retrieved sequence if found, and adds any new mappings to the
814 void updateDbrefMappings(SequenceI mapFrom, DBRefEntry[] xrefs,
815 SequenceI[] retrieved, AlignedCodonFrame acf, boolean fromDna)
817 SequenceIdMatcher idMatcher = new SequenceIdMatcher(retrieved);
818 for (DBRefEntry xref : xrefs)
822 String targetSeqName = xref.getSource() + "|"
823 + xref.getAccessionId();
824 SequenceI[] matches = idMatcher.findAllIdMatches(targetSeqName);
829 for (SequenceI seq : matches)
831 constructMapping(mapFrom, seq, xref, acf, fromDna);
838 * Tries to make a mapping between sequences. If successful, adds the mapping
839 * to the dbref and the mappings collection and answers true, otherwise
840 * answers false. The following methods of making are mapping are tried in
843 * <li>if 'mapTo' holds a mapping to 'mapFrom', take the inverse; this is, for
844 * example, the case after fetching EMBL cross-references for a Uniprot
846 * <li>else check if the dna translates exactly to the protein (give or take
847 * start and stop codons></li>
848 * <li>else try to map based on CDS features on the dna sequence</li>
857 boolean constructMapping(SequenceI mapFrom, SequenceI mapTo,
858 DBRefEntry xref, AlignedCodonFrame mappings, boolean fromDna)
860 MapList mapping = null;
861 SequenceI dsmapFrom = mapFrom.getDatasetSequence() == null ? mapFrom
862 : mapFrom.getDatasetSequence();
863 SequenceI dsmapTo = mapTo.getDatasetSequence() == null ? mapTo
864 : mapTo.getDatasetSequence();
866 * look for a reverse mapping, if found make its inverse.
867 * Note - we do this on dataset sequences only.
869 if (dsmapTo.getDBRefs() != null)
871 for (DBRefEntry dbref : dsmapTo.getDBRefs())
873 String name = dbref.getSource() + "|" + dbref.getAccessionId();
874 if (dbref.hasMap() && dsmapFrom.getName().startsWith(name))
877 * looks like we've found a map from 'mapTo' to 'mapFrom'
878 * - invert it to make the mapping the other way
880 MapList reverse = dbref.getMap().getMap().getInverse();
881 xref.setMap(new Mapping(dsmapTo, reverse));
882 mappings.addMap(mapFrom, dsmapTo, reverse);
890 mapping = AlignmentUtils.mapCdnaToProtein(mapTo, mapFrom);
894 mapping = AlignmentUtils.mapCdnaToProtein(mapFrom, mapTo);
897 mapping = mapping.getInverse();
904 xref.setMap(new Mapping(mapTo, mapping));
907 * and add a reverse DbRef with the inverse mapping
909 if (mapFrom.getDatasetSequence() != null && false)
910 // && mapFrom.getDatasetSequence().getSourceDBRef() != null)
912 // possible need to search primary references... except, why doesn't xref
913 // == getSourceDBRef ??
914 // DBRefEntry dbref = new DBRefEntry(mapFrom.getDatasetSequence()
915 // .getSourceDBRef());
916 // dbref.setMap(new Mapping(mapFrom.getDatasetSequence(), mapping
918 // mapTo.addDBRef(dbref);
923 AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping);
924 mappings.addMap(mapFrom, mapTo, mapping);
928 mappings.addMap(mapTo, mapFrom, mapping.getInverse());
935 * find references to lrfs in the cross-reference set of each sequence in
936 * dataset (that is not equal to sequenceI) Identifies matching DBRefEntry
937 * based on source and accession string only - Map and Version are nulled.
940 * - true if context was searching from Dna sequences, false if
941 * context was searching from Protein sequences
945 * @return true if matches were found.
947 private boolean searchDatasetXrefs(boolean fromDna, SequenceI sequenceI,
948 DBRefEntry[] lrfs, List<SequenceI> foundSeqs,
949 AlignedCodonFrame cf)
951 boolean found = false;
956 for (int i = 0; i < lrfs.length; i++)
958 DBRefEntry xref = new DBRefEntry(lrfs[i]);
960 xref.setVersion(null);
962 found |= searchDataset(fromDna, sequenceI, xref, foundSeqs, cf,
969 * Searches dataset for DBRefEntrys matching the given one (xrf) and adds the
970 * associated sequence to rseqs
973 * true if context was searching for refs *from* dna sequence, false
974 * if context was searching for refs *from* protein sequence
976 * a sequence to ignore (start point of search)
978 * a cross-reference to try to match
980 * result list to add to
982 * a set of sequence mappings to add to
984 * - indicates the type of relationship between returned sequences,
985 * xrf, and sequenceI that is required.
987 * <li>direct implies xrf is a primary reference for sequenceI AND
988 * the sequences to be located (eg a uniprot ID for a protein
989 * sequence, and a uniprot ref on a transcript sequence).</li>
990 * <li>indirect means xrf is a cross reference with respect to
991 * sequenceI or all the returned sequences (eg a genomic reference
992 * associated with a locus and one or more transcripts)</li>
994 * @return true if relationship found and sequence added.
996 boolean searchDataset(boolean fromDna, SequenceI fromSeq, DBRefEntry xrf,
997 List<SequenceI> foundSeqs, AlignedCodonFrame mappings,
1000 boolean found = false;
1001 if (dataset == null)
1005 if (dataset.getSequences() == null)
1007 System.err.println("Empty dataset sequence set - NO VECTOR");
1011 synchronized (ds = dataset.getSequences())
1013 for (SequenceI nxt : ds)
1017 if (nxt.getDatasetSequence() != null)
1020 "Implementation warning: CrossRef initialised with a dataset alignment with non-dataset sequences in it! ("
1021 + nxt.getDisplayId(true) + " has ds reference "
1022 + nxt.getDatasetSequence().getDisplayId(true)
1025 if (nxt == fromSeq || nxt == fromSeq.getDatasetSequence())
1030 * only look at same molecule type if 'direct', or
1031 * complementary type if !direct
1034 boolean isDna = !nxt.isProtein();
1035 if (direct ? (isDna != fromDna) : (isDna == fromDna))
1037 // skip this sequence because it is wrong molecule type
1042 // look for direct or indirect references in common
1043 DBRefEntry[] poss = nxt.getDBRefs();
1044 List<DBRefEntry> cands = null;
1046 // todo: indirect specifies we select either direct references to nxt
1047 // that match xrf which is indirect to sequenceI, or indirect
1048 // references to nxt that match xrf which is direct to sequenceI
1049 cands = DBRefUtils.searchRefs(poss, xrf);
1052 // poss = DBRefUtils.selectDbRefs(nxt.isProtein()!fromDna, poss);
1053 // cands = DBRefUtils.searchRefs(poss, xrf);
1055 if (!cands.isEmpty())
1057 if (foundSeqs.contains(nxt))
1063 if (mappings != null && !direct)
1066 * if the matched sequence has mapped dbrefs to
1067 * protein product / cdna, add equivalent mappings to
1068 * our source sequence
1070 for (DBRefEntry candidate : cands)
1072 Mapping mapping = candidate.getMap();
1073 if (mapping != null)
1075 MapList map = mapping.getMap();
1076 if (mapping.getTo() != null
1077 && map.getFromRatio() != map.getToRatio())
1080 * add a mapping, as from dna to peptide sequence
1082 if (map.getFromRatio() == 3)
1084 mappings.addMap(nxt, fromSeq, map);
1088 mappings.addMap(nxt, fromSeq, map.getInverse());