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);
110 sources.remove(DBRefSource.EMBL); // hack to prevent EMBL xrefs resulting in
111 // redundant datasets
114 sources.remove(DBRefSource.ENSEMBL); // hack to prevent Ensembl and
115 // EnsemblGenomes xref option shown
117 sources.remove(DBRefSource.ENSEMBLGENOMES);
119 // redundant datasets
124 * Returns a list of distinct database sources for which a sequence has either
126 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
127 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
128 * reference from another sequence in the dataset which has a cross-reference
129 * to a direct DBRefEntry on the given sequence</li>
133 * the sequence whose dbrefs we are searching against
135 * when true, context is DNA - so sources identifying protein
136 * products will be returned.
138 * a list of sources to add matches to
140 void findXrefSourcesForSequence(SequenceI seq, boolean fromDna,
141 List<String> sources)
144 * first find seq's xrefs (dna-to-peptide or peptide-to-dna)
146 List<DBRefEntry> rfs = DBRefUtils.selectDbRefs(!fromDna, seq.getDBRefs());
147 addXrefsToSources(rfs, sources);
151 * find sequence's direct (dna-to-dna, peptide-to-peptide) xrefs
153 List<DBRefEntry> lrfs = DBRefUtils.selectDbRefs(fromDna, seq.getDBRefs());
154 List<SequenceI> foundSeqs = new ArrayList<>();
157 * find sequences in the alignment which xref one of these DBRefs
158 * i.e. is xref-ed to a common sequence identifier
160 searchDatasetXrefs(fromDna, seq, lrfs, foundSeqs, null);
163 * add those sequences' (dna-to-peptide or peptide-to-dna) dbref sources
165 for (SequenceI rs : foundSeqs)
167 List<DBRefEntry> xrs = DBRefUtils.selectDbRefs(!fromDna,
169 addXrefsToSources(xrs, sources);
175 * Helper method that adds the source identifiers of some cross-references to
176 * a (non-redundant) list of database sources
181 void addXrefsToSources(List<DBRefEntry> xrefs, List<String> sources)
185 for (DBRefEntry ref : xrefs)
188 * avoid duplication e.g. ENSEMBL and Ensembl
190 String source = DBRefUtils.getCanonicalName(ref.getSource());
191 if (!sources.contains(source))
200 * Attempts to find cross-references from the sequences provided in the
201 * constructor to the given source database. Cross-references may be found
203 * <li>in dbrefs on the sequence which hold a mapping to a sequence
205 * <li>provided with a fetched sequence (e.g. ENA translation), or</li>
206 * <li>populated previously after getting cross-references</li>
208 * <li>as other sequences in the alignment which share a dbref identifier with
210 * <li>by fetching from the remote database</li>
212 * The cross-referenced sequences, and mappings to them, are added to the
216 * @return cross-referenced sequences (as dataset sequences)
218 public Alignment findXrefSequences(String source, boolean fromDna)
221 rseqs = new ArrayList<>();
222 AlignedCodonFrame cf = new AlignedCodonFrame();
223 matcher = new SequenceIdMatcher(dataset.getSequences());
225 for (SequenceI seq : fromSeqs)
228 while (dss.getDatasetSequence() != null)
230 dss = dss.getDatasetSequence();
232 boolean found = false;
233 List<DBRefEntry> xrfs = DBRefUtils.selectDbRefs(!fromDna,
235 // ENST & ENSP comes in to both Protein and nucleotide, so we need to
238 if ((xrfs == null || xrfs.size() == 0) && dataset != null)
241 * found no suitable dbrefs on sequence - look for sequences in the
242 * alignment which share a dbref with this one
244 List<DBRefEntry> lrfs = DBRefUtils.selectDbRefs(fromDna,
248 * find sequences (except this one!), of complementary type,
249 * which have a dbref to an accession id for this sequence,
250 * and add them to the results
252 found = searchDatasetXrefs(fromDna, dss, lrfs, rseqs, cf);
254 if (xrfs == null && !found)
257 * no dbref to source on this sequence or matched
258 * complementary sequence in the dataset
262 List<DBRefEntry> sourceRefs = DBRefUtils.searchRefsForSource(xrfs,
264 Iterator<DBRefEntry> refIterator = sourceRefs.iterator();
265 // At this point, if we are retrieving Ensembl, we still don't filter out
266 // ENST when looking for protein crossrefs.
267 while (refIterator.hasNext())
269 DBRefEntry xref = refIterator.next();
271 // we're only interested in coding cross-references, not
273 if (xref.hasMap() && xref.getMap().getMap().isTripletMap())
275 SequenceI mappedTo = xref.getMap().getTo();
276 if (mappedTo != null)
279 * dbref contains the sequence it maps to; add it to the
280 * results unless we have done so already (could happen if
281 * fetching xrefs for sequences which have xrefs in common)
282 * for example: UNIPROT {P0CE19, P0CE20} -> EMBL {J03321, X06707}
286 * problem: matcher.findIdMatch() is lenient - returns a sequence
287 * with a dbref to the search arg e.g. ENST for ENSP - wrong
288 * but findInDataset() matches ENSP when looking for Uniprot...
290 SequenceI matchInDataset = findInDataset(xref);
291 if (matchInDataset != null && xref.getMap().getTo() != null
292 && matchInDataset != xref.getMap().getTo())
295 "Implementation problem (reopen JAL-2154): CrossRef.findInDataset seems to have recovered a different sequence than the one explicitly mapped for xref."
296 + "Found:" + matchInDataset + "\nExpected:"
297 + xref.getMap().getTo() + "\nFor xref:"
300 /*matcher.findIdMatch(mappedTo);*/
301 if (matchInDataset != null)
303 if (!rseqs.contains(matchInDataset))
305 rseqs.add(matchInDataset);
307 // even if rseqs contained matchInDataset - check mappings between
308 // these seqs are added
309 // need to try harder to only add unique mappings
310 if (xref.getMap().getMap().isTripletMap()
311 && dataset.getMapping(seq, matchInDataset) == null
312 && cf.getMappingBetween(seq, matchInDataset) == null)
314 // materialise a mapping for highlighting between these
318 cf.addMap(dss, matchInDataset, xref.getMap().getMap(),
319 xref.getMap().getMappedFromId());
323 cf.addMap(matchInDataset, dss,
324 xref.getMap().getMap().getInverse(),
325 xref.getMap().getMappedFromId());
329 refIterator.remove();
332 // TODO: need to determine if this should be a deriveSequence
333 SequenceI rsq = new Sequence(mappedTo);
335 if (xref.getMap().getMap().isTripletMap())
337 // get sense of map correct for adding to product alignment.
340 // map is from dna seq to a protein product
341 cf.addMap(dss, rsq, xref.getMap().getMap(),
342 xref.getMap().getMappedFromId());
346 // map should be from protein seq to its coding dna
347 cf.addMap(rsq, dss, xref.getMap().getMap().getInverse(),
348 xref.getMap().getMappedFromId());
356 SequenceI matchedSeq = matcher.findIdMatch(
357 xref.getSource() + "|" + xref.getAccessionId());
358 // if there was a match, check it's at least the right type of
360 if (matchedSeq != null && matchedSeq.isProtein() == fromDna)
362 if (constructMapping(seq, matchedSeq, xref, cf, fromDna))
371 // do a bit more work - search for sequences with references matching
372 // xrefs on this sequence.
373 found = searchDataset(fromDna, dss, xref, rseqs, cf, false, DBRefUtils.SEARCH_MODE_FULL);
377 refIterator.remove();
382 * fetch from source database any dbrefs we haven't resolved up to here
384 if (!sourceRefs.isEmpty())
386 retrieveCrossRef(sourceRefs, seq, xrfs, fromDna, cf);
390 Alignment ral = null;
391 if (rseqs.size() > 0)
393 ral = new Alignment(rseqs.toArray(new SequenceI[rseqs.size()]));
396 dataset.addCodonFrame(cf);
402 private void retrieveCrossRef(List<DBRefEntry> sourceRefs, SequenceI seq,
403 List<DBRefEntry> xrfs, boolean fromDna, AlignedCodonFrame cf)
405 ASequenceFetcher sftch = SequenceFetcherFactory.getSequenceFetcher();
406 SequenceI[] retrieved = null;
407 SequenceI dss = seq.getDatasetSequence() == null ? seq
408 : seq.getDatasetSequence();
409 // first filter in case we are retrieving crossrefs that have already been
410 // retrieved. this happens for cases where a database record doesn't yield
411 // protein products for CDS
412 removeAlreadyRetrievedSeqs(sourceRefs, fromDna);
413 if (sourceRefs.size() == 0)
415 // no more work to do! We already had all requested sequence records in
421 retrieved = sftch.getSequences(sourceRefs, !fromDna);
422 } catch (Exception e)
425 "Problem whilst retrieving cross references for Sequence : "
430 if (retrieved != null)
432 boolean addedXref = false;
433 List<SequenceI> newDsSeqs = new ArrayList<>(),
434 doNotAdd = new ArrayList<>();
436 for (SequenceI retrievedSequence : retrieved)
438 // dataset gets contaminated ccwith non-ds sequences. why ??!
439 // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL->
440 SequenceI retrievedDss = retrievedSequence
441 .getDatasetSequence() == null ? retrievedSequence
442 : retrievedSequence.getDatasetSequence();
443 addedXref |= importCrossRefSeq(cf, newDsSeqs, doNotAdd, dss,
448 // try again, after looking for matching IDs
449 // shouldn't need to do this unless the dbref mechanism has broken.
450 updateDbrefMappings(seq, xrfs, retrieved, cf, fromDna);
451 for (SequenceI retrievedSequence : retrieved)
453 // dataset gets contaminated ccwith non-ds sequences. why ??!
454 // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL->
455 SequenceI retrievedDss = retrievedSequence
456 .getDatasetSequence() == null ? retrievedSequence
457 : retrievedSequence.getDatasetSequence();
458 addedXref |= importCrossRefSeq(cf, newDsSeqs, doNotAdd, dss,
462 for (SequenceI newToSeq : newDsSeqs)
464 if (!doNotAdd.contains(newToSeq)
465 && dataset.findIndex(newToSeq) == -1)
467 dataset.addSequence(newToSeq);
468 matcher.add(newToSeq);
475 * Search dataset for sequences with a primary reference contained in
479 * - list of references to filter.
481 * - type of sequence to search for matching primary reference.
483 private void removeAlreadyRetrievedSeqs(List<DBRefEntry> sourceRefs,
486 List<DBRefEntry> dbrSourceSet = new ArrayList<DBRefEntry>(sourceRefs);
487 List<SequenceI> dsSeqs = dataset.getSequences();
488 for (int ids = 0, nds = dsSeqs.size(); ids < nds; ids++)
490 SequenceI sq = dsSeqs.get(ids);
491 boolean dupeFound = false;
492 // !fromDna means we are looking only for nucleotide sequences, not
494 if (sq.isProtein() == fromDna)
496 List<DBRefEntry> sqdbrefs = sq.getPrimaryDBRefs();
497 for (int idb = 0, ndb = sqdbrefs.size(); idb < ndb; idb++)
499 DBRefEntry dbr = sqdbrefs.get(idb);
500 List<DBRefEntry> searchrefs = DBRefUtils.searchRefs(dbrSourceSet, dbr, DBRefUtils.SEARCH_MODE_FULL);
501 for (int isr = 0, nsr = searchrefs.size(); isr < nsr; isr++)
503 sourceRefs.remove(searchrefs.get(isr));
510 // rebuild the search array from the filtered sourceRefs list
511 dbrSourceSet.clear();
512 dbrSourceSet.addAll(sourceRefs);
518 * process sequence retrieved via a dbref on source sequence to resolve and
522 * @param sourceSequence
523 * @param retrievedSequence
524 * @return true if retrieveSequence was imported
526 private boolean importCrossRefSeq(AlignedCodonFrame cf,
527 List<SequenceI> newDsSeqs, List<SequenceI> doNotAdd,
528 SequenceI sourceSequence, SequenceI retrievedSequence)
531 * set when retrievedSequence has been verified as a crossreference for
534 boolean imported = false;
535 List<DBRefEntry> dbr = retrievedSequence.getDBRefs();
538 for (int ib = 0, nb = dbr.size(); ib < nb; ib++)
541 DBRefEntry dbref = dbr.get(ib);
542 SequenceI matched = findInDataset(dbref);
543 if (matched == sourceSequence)
545 // verified retrieved and source sequence cross-reference each other
548 // find any entry where we should put in the sequence being
549 // cross-referenced into the map
550 Mapping map = dbref.getMap();
553 SequenceI ms = map.getTo();
554 if (ms != null && map.getMap() != null)
556 if (ms == sourceSequence)
558 // already called to import once, and most likely this sequence
559 // already imported !
565 * sequence is new to dataset, so save a reference so it can be added.
572 * there was a matching sequence in dataset, so now, check to see if we can update the map.getTo() sequence to the existing one.
577 // compare ms with dss and replace with dss in mapping
578 // if map is congruent
579 // TODO findInDataset requires exact sequence match but
580 // 'congruent' test is only for the mapped part
581 // maybe not a problem in practice since only ENA provide a
582 // mapping and it is to the full protein translation of CDS
583 // matcher.findIdMatch(map.getTo());
584 // TODO addendum: if matched is shorter than getTo, this will fail
585 // - when it should really succeed.
586 int sf = map.getMap().getToLowest();
587 int st = map.getMap().getToHighest();
588 SequenceI mappedrg = ms.getSubSequence(sf, st);
589 if (mappedrg.getLength() > 0 && ms.getSequenceAsString()
590 .equals(matched.getSequenceAsString()))
593 * sequences were a match,
595 String msg = "Mapping updated from " + ms.getName()
596 + " to retrieved crossreference "
598 System.out.println(msg);
600 List<DBRefEntry> toRefs = map.getTo().getDBRefs();
604 * transfer database refs
606 for (DBRefEntry ref : toRefs)
608 if (dbref.getSrcAccString()
609 .equals(ref.getSrcAccString()))
611 continue; // avoid overwriting the ref on source sequence
613 matched.addDBRef(ref); // add or update mapping
616 doNotAdd.add(map.getTo());
620 * give the reverse reference the inverse mapping
621 * (if it doesn't have one already)
623 setReverseMapping(matched, dbref, cf);
626 * copy sequence features as well, avoiding
627 * duplication (e.g. same variation from two
630 List<SequenceFeature> sfs = ms.getFeatures()
632 for (SequenceFeature feat : sfs)
635 * make a flyweight feature object which ignores Parent
636 * attribute in equality test; this avoids creating many
637 * otherwise duplicate exon features on genomic sequence
639 SequenceFeature newFeature = new SequenceFeature(feat)
642 public boolean equals(Object o)
644 return super.equals(o, true);
647 matched.addSequenceFeature(newFeature);
650 cf.addMap(retrievedSequence, map.getTo(), map.getMap());
651 } catch (Exception e)
654 "Exception when consolidating Mapped sequence set...");
655 e.printStackTrace(System.err);
663 retrievedSequence.updatePDBIds();
664 rseqs.add(retrievedSequence);
665 if (dataset.findIndex(retrievedSequence) == -1)
667 dataset.addSequence(retrievedSequence);
668 matcher.add(retrievedSequence);
675 * Sets the inverse sequence mapping in the corresponding dbref of the mapped
676 * to sequence (if any). This is used after fetching a cross-referenced
677 * sequence, if the fetched sequence has a mapping to the original sequence,
678 * to set the mapping in the original sequence's dbref.
681 * the sequence mapped from
685 void setReverseMapping(SequenceI mapFrom, DBRefEntry dbref,
686 AlignedCodonFrame mappings)
688 SequenceI mapTo = dbref.getMap().getTo();
693 List<DBRefEntry> dbrefs = mapTo.getDBRefs();
698 for (DBRefEntry toRef : dbrefs)
700 if (toRef.hasMap() && mapFrom == toRef.getMap().getTo())
703 * found the reverse dbref; update its mapping if null
705 if (toRef.getMap().getMap() == null)
707 MapList inverse = dbref.getMap().getMap().getInverse();
708 toRef.getMap().setMap(inverse);
709 mappings.addMap(mapTo, mapFrom, inverse);
716 * Returns null or the first sequence in the dataset which is identical to
717 * xref.mapTo, and has a) a primary dbref matching xref, or if none found, the
718 * first one with an ID source|xrefacc
721 * with map and mapped-to sequence
724 SequenceI findInDataset(DBRefEntry xref)
726 if (xref == null || !xref.hasMap() || xref.getMap().getTo() == null)
730 SequenceI mapsTo = xref.getMap().getTo();
731 String name = xref.getAccessionId();
732 String name2 = xref.getSource() + "|" + name;
733 SequenceI dss = mapsTo.getDatasetSequence() == null ? mapsTo
734 : mapsTo.getDatasetSequence();
735 // first check ds if ds is directly referenced
736 if (dataset.findIndex(dss) > -1)
740 DBRefEntry template = new DBRefEntry(xref.getSource(), null,
741 xref.getAccessionId());
743 * remember the first ID match - in case we don't find a match to template
745 SequenceI firstIdMatch = null;
746 for (SequenceI seq : dataset.getSequences())
748 // first check primary refs.
749 List<DBRefEntry> match = DBRefUtils.searchRefs(
750 seq.getPrimaryDBRefs(), template, DBRefUtils.SEARCH_MODE_FULL);
751 if (match != null && match.size() == 1 && sameSequence(seq, dss))
756 * clumsy alternative to using SequenceIdMatcher which currently
757 * returns sequences with a dbref to the matched accession id
758 * which we don't want
760 if (firstIdMatch == null && (name.equals(seq.getName())
761 || seq.getName().startsWith(name2)))
763 if (sameSequence(seq, dss))
773 * Answers true if seq1 and seq2 contain exactly the same characters (ignoring
774 * case), else false. This method compares the lengths, then each character in
775 * turn, in order to 'fail fast'. For case-sensitive comparison, it would be
776 * possible to use Arrays.equals(seq1.getSequence(), seq2.getSequence()).
782 // TODO move to Sequence / SequenceI
783 static boolean sameSequence(SequenceI seq1, SequenceI seq2)
789 if (seq1 == null || seq2 == null)
794 if (seq1.getLength() != seq2.getLength())
798 int length = seq1.getLength();
799 for (int i = 0; i < length; i++)
801 int diff = seq1.getCharAt(i) - seq2.getCharAt(i);
803 * same char or differ in case only ('a'-'A' == 32)
805 if (diff != 0 && diff != 32 && diff != -32)
814 * Updates any empty mappings in the cross-references with one to a compatible
815 * retrieved sequence if found, and adds any new mappings to the
823 void updateDbrefMappings(SequenceI mapFrom, List<DBRefEntry> xrefs,
824 SequenceI[] retrieved, AlignedCodonFrame acf, boolean fromDna)
826 SequenceIdMatcher idMatcher = new SequenceIdMatcher(retrieved);
827 for (DBRefEntry xref : xrefs)
831 String targetSeqName = xref.getSource() + "|"
832 + xref.getAccessionId();
833 SequenceI[] matches = idMatcher.findAllIdMatches(targetSeqName);
838 for (SequenceI seq : matches)
840 constructMapping(mapFrom, seq, xref, acf, fromDna);
847 * Tries to make a mapping between sequences. If successful, adds the mapping
848 * to the dbref and the mappings collection and answers true, otherwise
849 * answers false. The following methods of making are mapping are tried in
852 * <li>if 'mapTo' holds a mapping to 'mapFrom', take the inverse; this is, for
853 * example, the case after fetching EMBL cross-references for a Uniprot
855 * <li>else check if the dna translates exactly to the protein (give or take
856 * start and stop codons></li>
857 * <li>else try to map based on CDS features on the dna sequence</li>
866 boolean constructMapping(SequenceI mapFrom, SequenceI mapTo,
867 DBRefEntry xref, AlignedCodonFrame mappings, boolean fromDna)
869 MapList mapping = null;
870 SequenceI dsmapFrom = mapFrom.getDatasetSequence() == null ? mapFrom
871 : mapFrom.getDatasetSequence();
872 SequenceI dsmapTo = mapTo.getDatasetSequence() == null ? mapTo
873 : mapTo.getDatasetSequence();
875 * look for a reverse mapping, if found make its inverse.
876 * Note - we do this on dataset sequences only.
878 if (dsmapTo.getDBRefs() != null)
880 for (DBRefEntry dbref : dsmapTo.getDBRefs())
882 String name = dbref.getSource() + "|" + dbref.getAccessionId();
883 if (dbref.hasMap() && dsmapFrom.getName().startsWith(name))
886 * looks like we've found a map from 'mapTo' to 'mapFrom'
887 * - invert it to make the mapping the other way
889 MapList reverse = dbref.getMap().getMap().getInverse();
890 xref.setMap(new Mapping(dsmapTo, reverse));
891 mappings.addMap(mapFrom, dsmapTo, reverse);
899 mapping = AlignmentUtils.mapCdnaToProtein(mapTo, mapFrom);
903 mapping = AlignmentUtils.mapCdnaToProtein(mapFrom, mapTo);
906 mapping = mapping.getInverse();
913 xref.setMap(new Mapping(mapTo, mapping));
916 * and add a reverse DbRef with the inverse mapping
918 if (mapFrom.getDatasetSequence() != null && false)
919 // && mapFrom.getDatasetSequence().getSourceDBRef() != null)
921 // possible need to search primary references... except, why doesn't xref
922 // == getSourceDBRef ??
923 // DBRefEntry dbref = new DBRefEntry(mapFrom.getDatasetSequence()
924 // .getSourceDBRef());
925 // dbref.setMap(new Mapping(mapFrom.getDatasetSequence(), mapping
927 // mapTo.addDBRef(dbref);
932 AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping);
933 mappings.addMap(mapFrom, mapTo, mapping);
937 mappings.addMap(mapTo, mapFrom, mapping.getInverse());
944 * find references to lrfs in the cross-reference set of each sequence in
945 * dataset (that is not equal to sequenceI) Identifies matching DBRefEntry
946 * based on source and accession string only - Map and Version are nulled.
949 * - true if context was searching from Dna sequences, false if
950 * context was searching from Protein sequences
954 * @return true if matches were found.
956 private boolean searchDatasetXrefs(boolean fromDna, SequenceI sequenceI,
957 List<DBRefEntry> lrfs, List<SequenceI> foundSeqs,
958 AlignedCodonFrame cf)
960 boolean found = false;
965 for (int i = 0, n = lrfs.size(); i < n; i++)
967 // DBRefEntry xref = new DBRefEntry(lrfs.get(i));
968 // // add in wildcards
969 // xref.setVersion(null);
970 // xref.setMap(null);
971 found |= searchDataset(fromDna, sequenceI, lrfs.get(i), foundSeqs, cf,
972 false, DBRefUtils.SEARCH_MODE_NO_MAP_NO_VERSION);
978 * Searches dataset for DBRefEntrys matching the given one (xrf) and adds the
979 * associated sequence to rseqs
982 * true if context was searching for refs *from* dna sequence, false
983 * if context was searching for refs *from* protein sequence
985 * a sequence to ignore (start point of search)
987 * a cross-reference to try to match
989 * result list to add to
991 * a set of sequence mappings to add to
993 * - indicates the type of relationship between returned sequences,
994 * xrf, and sequenceI that is required.
996 * <li>direct implies xrf is a primary reference for sequenceI AND
997 * the sequences to be located (eg a uniprot ID for a protein
998 * sequence, and a uniprot ref on a transcript sequence).</li>
999 * <li>indirect means xrf is a cross reference with respect to
1000 * sequenceI or all the returned sequences (eg a genomic reference
1001 * associated with a locus and one or more transcripts)</li>
1003 * @param mode SEARCH_MODE_FULL for all; SEARCH_MODE_NO_MAP_NO_VERSION optional
1004 * @return true if relationship found and sequence added.
1006 boolean searchDataset(boolean fromDna, SequenceI fromSeq, DBRefEntry xrf,
1007 List<SequenceI> foundSeqs, AlignedCodonFrame mappings,
1008 boolean direct, int mode)
1010 boolean found = false;
1011 if (dataset == null)
1015 if (dataset.getSequences() == null)
1017 System.err.println("Empty dataset sequence set - NO VECTOR");
1020 List<SequenceI> ds = dataset.getSequences();
1023 for (SequenceI nxt : ds)
1027 if (nxt.getDatasetSequence() != null)
1030 "Implementation warning: CrossRef initialised with a dataset alignment with non-dataset sequences in it! ("
1031 + nxt.getDisplayId(true) + " has ds reference "
1032 + nxt.getDatasetSequence().getDisplayId(true)
1035 if (nxt == fromSeq || nxt == fromSeq.getDatasetSequence())
1040 * only look at same molecule type if 'direct', or
1041 * complementary type if !direct
1044 boolean isDna = !nxt.isProtein();
1045 if (direct ? (isDna != fromDna) : (isDna == fromDna))
1047 // skip this sequence because it is wrong molecule type
1052 // look for direct or indirect references in common
1053 List<DBRefEntry> poss = nxt.getDBRefs();
1054 List<DBRefEntry> cands = null;
1056 // todo: indirect specifies we select either direct references to nxt
1057 // that match xrf which is indirect to sequenceI, or indirect
1058 // references to nxt that match xrf which is direct to sequenceI
1059 cands = DBRefUtils.searchRefs(poss, xrf, mode);
1062 // poss = DBRefUtils.selectDbRefs(nxt.isProtein()!fromDna, poss);
1063 // cands = DBRefUtils.searchRefs(poss, xrf);
1065 if (!cands.isEmpty())
1067 if (foundSeqs.contains(nxt))
1073 if (mappings != null && !direct)
1076 * if the matched sequence has mapped dbrefs to
1077 * protein product / cdna, add equivalent mappings to
1078 * our source sequence
1080 for (DBRefEntry candidate : cands)
1082 Mapping mapping = candidate.getMap();
1083 if (mapping != null)
1085 MapList map = mapping.getMap();
1086 if (mapping.getTo() != null
1087 && map.getFromRatio() != map.getToRatio())
1090 * add a mapping, as from dna to peptide sequence
1092 if (map.getFromRatio() == 3)
1094 mappings.addMap(nxt, fromSeq, map);
1098 mappings.addMap(nxt, fromSeq, map.getInverse());