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.DBRefUtils;
32 import jalview.util.MapList;
33 import jalview.ws.SequenceFetcherFactory;
34 import jalview.ws.seqfetcher.ASequenceFetcher;
36 import java.util.ArrayList;
37 import java.util.Iterator;
38 import java.util.List;
41 * Functions for cross-referencing sequence databases.
49 * the dataset of the alignment for which we are searching for
50 * cross-references; in some cases we may resolve xrefs by
51 * searching in the dataset
53 private AlignmentI dataset;
56 * the sequences for which we are seeking cross-references
58 private SequenceI[] fromSeqs;
61 * matcher built from dataset
63 SequenceIdMatcher matcher;
66 * sequences found by cross-ref searches to fromSeqs
68 List<SequenceI> rseqs;
74 * the sequences for which we are seeking cross-references
76 * the containing alignment dataset (may be searched to resolve
79 public CrossRef(SequenceI[] seqs, AlignmentI ds)
82 dataset = ds.getDataset() == null ? ds : ds.getDataset();
86 * Returns a list of distinct database sources for which sequences have either
88 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
89 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
90 * reference from another sequence in the dataset which has a cross-reference
91 * to a direct DBRefEntry on the given sequence</li>
95 * - when true, cross-references *from* dna returned. When false,
96 * cross-references *from* protein are returned
99 public List<String> findXrefSourcesForSequences(boolean dna)
101 List<String> sources = new ArrayList<String>();
102 for (SequenceI seq : fromSeqs)
106 findXrefSourcesForSequence(seq, dna, sources);
113 * Returns a list of distinct database sources for which a sequence has either
115 * <li>a (dna-to-protein or protein-to-dna) cross-reference</li>
116 * <li>an indirect cross-reference - a (dna-to-protein or protein-to-dna)
117 * reference from another sequence in the dataset which has a cross-reference
118 * to a direct DBRefEntry on the given sequence</li>
122 * the sequence whose dbrefs we are searching against
124 * when true, context is DNA - so sources identifying protein
125 * products will be returned.
127 * a list of sources to add matches to
129 void findXrefSourcesForSequence(SequenceI seq, boolean fromDna,
130 List<String> sources)
133 * first find seq's xrefs (dna-to-peptide or peptide-to-dna)
135 DBRefEntry[] rfs = DBRefUtils.selectDbRefs(!fromDna, seq.getDBRefs());
136 addXrefsToSources(rfs, sources);
140 * find sequence's direct (dna-to-dna, peptide-to-peptide) xrefs
142 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(fromDna, seq.getDBRefs());
143 List<SequenceI> foundSeqs = new ArrayList<SequenceI>();
146 * find sequences in the alignment which xref one of these DBRefs
147 * i.e. is xref-ed to a common sequence identifier
149 searchDatasetXrefs(fromDna, seq, lrfs, foundSeqs, null);
152 * add those sequences' (dna-to-peptide or peptide-to-dna) dbref sources
154 for (SequenceI rs : foundSeqs)
156 DBRefEntry[] xrs = DBRefUtils
157 .selectDbRefs(!fromDna, rs.getDBRefs());
158 addXrefsToSources(xrs, sources);
164 * Helper method that adds the source identifiers of some cross-references to
165 * a (non-redundant) list of database sources
170 void addXrefsToSources(DBRefEntry[] xrefs, List<String> sources)
174 for (DBRefEntry ref : xrefs)
177 * avoid duplication e.g. ENSEMBL and Ensembl
179 String source = DBRefUtils.getCanonicalName(ref.getSource());
180 if (!sources.contains(source))
189 * Attempts to find cross-references from the sequences provided in the
190 * constructor to the given source database. Cross-references may be found
192 * <li>in dbrefs on the sequence which hold a mapping to a sequence
194 * <li>provided with a fetched sequence (e.g. ENA translation), or</li>
195 * <li>populated previously after getting cross-references</li>
197 * <li>as other sequences in the alignment which share a dbref identifier with
199 * <li>by fetching from the remote database</li>
201 * The cross-referenced sequences, and mappings to them, are added to the
205 * @return cross-referenced sequences (as dataset sequences)
207 public Alignment findXrefSequences(String source, boolean fromDna)
210 rseqs = new ArrayList<SequenceI>();
211 AlignedCodonFrame cf = new AlignedCodonFrame();
212 matcher = new SequenceIdMatcher(
213 dataset.getSequences());
215 for (SequenceI seq : fromSeqs)
218 while (dss.getDatasetSequence() != null)
220 dss = dss.getDatasetSequence();
222 boolean found = false;
223 DBRefEntry[] xrfs = DBRefUtils
224 .selectDbRefs(!fromDna, dss.getDBRefs());
225 // ENST & ENSP comes in to both Protein and nucleotide, so we need to
228 if ((xrfs == null || xrfs.length == 0) && dataset != null)
231 * found no suitable dbrefs on sequence - look for sequences in the
232 * alignment which share a dbref with this one
234 DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(fromDna,
238 * find sequences (except this one!), of complementary type,
239 * which have a dbref to an accession id for this sequence,
240 * and add them to the results
242 found = searchDatasetXrefs(fromDna, dss, lrfs, rseqs, cf);
244 if (xrfs == null && !found)
247 * no dbref to source on this sequence or matched
248 * complementary sequence in the dataset
252 List<DBRefEntry> sourceRefs = DBRefUtils.searchRefsForSource(xrfs,
254 Iterator<DBRefEntry> refIterator = sourceRefs.iterator();
255 // At this point, if we are retrieving Ensembl, we still don't filter out
256 // ENST when looking for protein crossrefs.
257 while (refIterator.hasNext())
259 DBRefEntry xref = refIterator.next();
261 // we're only interested in coding cross-references, not
263 if (xref.hasMap() && xref.getMap().getMap().isTripletMap())
265 SequenceI mappedTo = xref.getMap().getTo();
266 if (mappedTo != null)
269 * dbref contains the sequence it maps to; add it to the
270 * results unless we have done so already (could happen if
271 * fetching xrefs for sequences which have xrefs in common)
272 * for example: UNIPROT {P0CE19, P0CE20} -> EMBL {J03321, X06707}
276 * problem: matcher.findIdMatch() is lenient - returns a sequence
277 * with a dbref to the search arg e.g. ENST for ENSP - wrong
278 * but findInDataset() matches ENSP when looking for Uniprot...
280 SequenceI matchInDataset = findInDataset(xref);
281 if (matchInDataset != null && xref.getMap().getTo() != null
282 && matchInDataset != xref.getMap().getTo())
285 .println("Implementation problem (reopen JAL-2154): CrossRef.findInDataset seems to have recovered a different sequence than the one explicitly mapped for xref."
289 + xref.getMap().getTo()
293 /*matcher.findIdMatch(mappedTo);*/
294 if (matchInDataset != null)
296 if (!rseqs.contains(matchInDataset))
298 rseqs.add(matchInDataset);
300 // even if rseqs contained matchInDataset - check mappings between
301 // these seqs are added
302 // need to try harder to only add unique mappings
303 if (xref.getMap().getMap().isTripletMap()
304 && dataset.getMapping(seq, matchInDataset) == null
305 && cf.getMappingBetween(seq, matchInDataset) == null)
307 // materialise a mapping for highlighting between these
311 cf.addMap(dss, matchInDataset, xref.getMap().getMap(),
312 xref.getMap().getMappedFromId());
316 cf.addMap(matchInDataset, dss, xref.getMap().getMap()
317 .getInverse(), xref.getMap().getMappedFromId());
321 refIterator.remove();
324 // TODO: need to determine if this should be a deriveSequence
325 SequenceI rsq = new Sequence(mappedTo);
327 if (xref.getMap().getMap().isTripletMap())
329 // get sense of map correct for adding to product alignment.
332 // map is from dna seq to a protein product
333 cf.addMap(dss, rsq, xref.getMap().getMap(), xref.getMap()
338 // map should be from protein seq to its coding dna
339 cf.addMap(rsq, dss, xref.getMap().getMap().getInverse(),
340 xref.getMap().getMappedFromId());
348 SequenceI matchedSeq = matcher.findIdMatch(xref.getSource() + "|"
349 + xref.getAccessionId());
350 // if there was a match, check it's at least the right type of
352 if (matchedSeq != null && matchedSeq.isProtein() == fromDna)
354 if (constructMapping(seq, matchedSeq, xref, cf, fromDna))
363 // do a bit more work - search for sequences with references matching
364 // xrefs on this sequence.
365 found = searchDataset(fromDna, dss, xref, rseqs, cf, false);
369 refIterator.remove();
374 * fetch from source database any dbrefs we haven't resolved up to here
376 if (!sourceRefs.isEmpty())
378 retrieveCrossRef(sourceRefs, seq, xrfs, fromDna, cf);
382 Alignment ral = null;
383 if (rseqs.size() > 0)
385 ral = new Alignment(rseqs.toArray(new SequenceI[rseqs.size()]));
388 dataset.addCodonFrame(cf);
394 private void retrieveCrossRef(List<DBRefEntry> sourceRefs, SequenceI seq,
395 DBRefEntry[] xrfs, boolean fromDna, AlignedCodonFrame cf)
397 ASequenceFetcher sftch = SequenceFetcherFactory.getSequenceFetcher();
398 SequenceI[] retrieved = null;
399 SequenceI dss = seq.getDatasetSequence() == null ? seq : seq
400 .getDatasetSequence();
401 // first filter in case we are retrieving crossrefs that have already been
402 // retrieved. this happens for cases where a database record doesn't yield
403 // protein products for CDS
404 removeAlreadyRetrievedSeqs(sourceRefs, fromDna);
405 if (sourceRefs.size() == 0)
407 // no more work to do! We already had all requested sequence records in
413 retrieved = sftch.getSequences(sourceRefs, !fromDna);
414 } catch (Exception e)
417 .println("Problem whilst retrieving cross references for Sequence : "
422 if (retrieved != null)
424 boolean addedXref = false;
425 for (SequenceI retrievedSequence : retrieved)
427 // dataset gets contaminated ccwith non-ds sequences. why ??!
428 // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL->
429 SequenceI retrievedDss = retrievedSequence.getDatasetSequence() == null ? retrievedSequence
430 : retrievedSequence.getDatasetSequence();
431 addedXref |= importCrossRefSeq(cf, dss, retrievedDss);
435 // try again, after looking for matching IDs
436 // shouldn't need to do this unless the dbref mechanism has broken.
437 updateDbrefMappings(seq, xrfs, retrieved, cf, fromDna);
438 for (SequenceI retrievedSequence : retrieved)
440 // dataset gets contaminated ccwith non-ds sequences. why ??!
441 // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL->
442 SequenceI retrievedDss = retrievedSequence.getDatasetSequence() == null ? retrievedSequence
443 : retrievedSequence.getDatasetSequence();
444 addedXref |= importCrossRefSeq(cf, dss, retrievedDss);
451 * Search dataset for sequences with a primary reference contained in
455 * - list of references to filter.
457 * - type of sequence to search for matching primary reference.
459 private void removeAlreadyRetrievedSeqs(List<DBRefEntry> sourceRefs,
462 DBRefEntry[] dbrSourceSet = sourceRefs.toArray(new DBRefEntry[0]);
463 for (SequenceI sq : dataset.getSequences())
465 boolean dupeFound = false;
466 // !fromDna means we are looking only for nucleotide sequences, not
468 if (sq.isProtein() == fromDna)
470 for (DBRefEntry dbr : sq.getPrimaryDBRefs())
472 for (DBRefEntry found : DBRefUtils.searchRefs(dbrSourceSet, dbr))
474 sourceRefs.remove(found);
481 // rebuild the search array from the filtered sourceRefs list
482 dbrSourceSet = sourceRefs.toArray(new DBRefEntry[0]);
488 * process sequence retrieved via a dbref on source sequence to resolve and
492 * @param sourceSequence
493 * @param retrievedSequence
494 * @return true if retrieveSequence was imported
496 private boolean importCrossRefSeq(AlignedCodonFrame cf,
497 SequenceI sourceSequence, SequenceI retrievedSequence)
500 * set when retrievedSequence has been verified as a crossreference for
503 boolean imported = false;
504 DBRefEntry[] dbr = retrievedSequence.getDBRefs();
505 List<SequenceI> newDsSeqs = new ArrayList<SequenceI>();
508 for (DBRefEntry dbref : dbr)
510 SequenceI matched = findInDataset(dbref);
511 if (matched == sourceSequence)
513 // verified retrieved and source sequence cross-reference each other
516 // find any entry where we should put in the sequence being
517 // cross-referenced into the map
518 Mapping map = dbref.getMap();
521 if (map.getTo() != null && map.getMap() != null)
523 if (map.getTo() == sourceSequence)
525 // already called to import once, and most likely this sequence
526 // already imported !
532 * sequence is new to dataset, so save a reference so it can be added.
534 newDsSeqs.add(map.getTo());
539 * there was a matching sequence in dataset, so now, check to see if we can update the map.getTo() sequence to the existing one.
544 // compare ms with dss and replace with dss in mapping
545 // if map is congruent
546 SequenceI ms = map.getTo();
547 // TODO findInDataset requires exact sequence match but
548 // 'congruent' test is only for the mapped part
549 // maybe not a problem in practice since only ENA provide a
550 // mapping and it is to the full protein translation of CDS
551 // matcher.findIdMatch(map.getTo());
552 // TODO addendum: if matched is shorter than getTo, this will fail
553 // - when it should really succeed.
554 int sf = map.getMap().getToLowest();
555 int st = map.getMap().getToHighest();
556 SequenceI mappedrg = ms.getSubSequence(sf, st);
557 if (mappedrg.getLength() > 0
558 && ms.getSequenceAsString().equals(
559 matched.getSequenceAsString()))
562 * sequences were a match,
564 String msg = "Mapping updated from " + ms.getName()
565 + " to retrieved crossreference "
567 System.out.println(msg);
569 DBRefEntry[] toRefs = map.getTo().getDBRefs();
573 * transfer database refs
575 for (DBRefEntry ref : toRefs)
577 matched.addDBRef(ref); // add or update mapping
583 * give the reverse reference the inverse mapping
584 * (if it doesn't have one already)
586 setReverseMapping(matched, dbref, cf);
589 * copy sequence features as well, avoiding
590 * duplication (e.g. same variation from two
593 SequenceFeature[] sfs = ms.getSequenceFeatures();
596 for (SequenceFeature feat : sfs)
599 * make a flyweight feature object which ignores Parent
600 * attribute in equality test; this avoids creating many
601 * otherwise duplicate exon features on genomic sequence
603 SequenceFeature newFeature = new SequenceFeature(
607 public boolean equals(Object o)
609 return super.equals(o, true);
612 matched.addSequenceFeature(newFeature);
617 cf.addMap(retrievedSequence, map.getTo(), map.getMap());
618 } catch (Exception e)
621 .println("Exception when consolidating Mapped sequence set...");
622 e.printStackTrace(System.err);
630 retrievedSequence.updatePDBIds();
631 rseqs.add(retrievedSequence);
632 if (dataset.findIndex(retrievedSequence) == -1)
634 dataset.addSequence(retrievedSequence);
635 matcher.add(retrievedSequence);
637 for (SequenceI newToSeq : newDsSeqs)
640 if (dataset.findIndex(newToSeq) == -1)
642 dataset.addSequence(newToSeq);
643 matcher.add(newToSeq);
650 * Sets the inverse sequence mapping in the corresponding dbref of the mapped
651 * to sequence (if any). This is used after fetching a cross-referenced
652 * sequence, if the fetched sequence has a mapping to the original sequence,
653 * to set the mapping in the original sequence's dbref.
656 * the sequence mapped from
660 void setReverseMapping(SequenceI mapFrom, DBRefEntry dbref,
661 AlignedCodonFrame mappings)
663 SequenceI mapTo = dbref.getMap().getTo();
668 DBRefEntry[] dbrefs = mapTo.getDBRefs();
673 for (DBRefEntry toRef : dbrefs)
675 if (toRef.hasMap() && mapFrom == toRef.getMap().getTo())
678 * found the reverse dbref; update its mapping if null
680 if (toRef.getMap().getMap() == null)
682 MapList inverse = dbref.getMap().getMap().getInverse();
683 toRef.getMap().setMap(inverse);
684 mappings.addMap(mapTo, mapFrom, inverse);
691 * Returns null or the first sequence in the dataset which is identical to
692 * xref.mapTo, and has a) a primary dbref matching xref, or if none found, the
693 * first one with an ID source|xrefacc
696 * with map and mapped-to sequence
699 SequenceI findInDataset(DBRefEntry xref)
701 if (xref == null || !xref.hasMap() || xref.getMap().getTo() == null)
705 SequenceI mapsTo = xref.getMap().getTo();
706 String name = xref.getAccessionId();
707 String name2 = xref.getSource() + "|" + name;
708 SequenceI dss = mapsTo.getDatasetSequence() == null ? mapsTo : mapsTo
709 .getDatasetSequence();
710 // first check ds if ds is directly referenced
711 if (dataset.findIndex(dss) > -1)
715 DBRefEntry template = new DBRefEntry(xref.getSource(), null,
716 xref.getAccessionId());
718 * remember the first ID match - in case we don't find a match to template
720 SequenceI firstIdMatch = null;
721 for (SequenceI seq : dataset.getSequences())
723 // first check primary refs.
724 List<DBRefEntry> match = DBRefUtils.searchRefs(seq.getPrimaryDBRefs()
725 .toArray(new DBRefEntry[0]), template);
726 if (match != null && match.size() == 1 && sameSequence(seq, dss))
731 * clumsy alternative to using SequenceIdMatcher which currently
732 * returns sequences with a dbref to the matched accession id
733 * which we don't want
735 if (firstIdMatch == null
736 && (name.equals(seq.getName()) || seq.getName().startsWith(
739 if (sameSequence(seq, dss))
749 * Answers true if seq1 and seq2 contain exactly the same characters (ignoring
750 * case), else false. This method compares the lengths, then each character in
751 * turn, in order to 'fail fast'. For case-sensitive comparison, it would be
752 * possible to use Arrays.equals(seq1.getSequence(), seq2.getSequence()).
758 // TODO move to Sequence / SequenceI
759 static boolean sameSequence(SequenceI seq1, SequenceI seq2)
765 if (seq1 == null || seq2 == null)
769 char[] c1 = seq1.getSequence();
770 char[] c2 = seq2.getSequence();
771 if (c1.length != c2.length)
775 for (int i = 0; i < c1.length; i++)
777 int diff = c1[i] - c2[i];
779 * same char or differ in case only ('a'-'A' == 32)
781 if (diff != 0 && diff != 32 && diff != -32)
790 * Updates any empty mappings in the cross-references with one to a compatible
791 * retrieved sequence if found, and adds any new mappings to the
799 void updateDbrefMappings(SequenceI mapFrom, DBRefEntry[] xrefs,
800 SequenceI[] retrieved, AlignedCodonFrame acf, boolean fromDna)
802 SequenceIdMatcher idMatcher = new SequenceIdMatcher(retrieved);
803 for (DBRefEntry xref : xrefs)
807 String targetSeqName = xref.getSource() + "|"
808 + xref.getAccessionId();
809 SequenceI[] matches = idMatcher.findAllIdMatches(targetSeqName);
814 for (SequenceI seq : matches)
816 constructMapping(mapFrom, seq, xref, acf, fromDna);
823 * Tries to make a mapping between sequences. If successful, adds the mapping
824 * to the dbref and the mappings collection and answers true, otherwise
825 * answers false. The following methods of making are mapping are tried in
828 * <li>if 'mapTo' holds a mapping to 'mapFrom', take the inverse; this is, for
829 * example, the case after fetching EMBL cross-references for a Uniprot
831 * <li>else check if the dna translates exactly to the protein (give or take
832 * start and stop codons></li>
833 * <li>else try to map based on CDS features on the dna sequence</li>
842 boolean constructMapping(SequenceI mapFrom, SequenceI mapTo,
843 DBRefEntry xref, AlignedCodonFrame mappings, boolean fromDna)
845 MapList mapping = null;
846 SequenceI dsmapFrom = mapFrom.getDatasetSequence() == null ? mapFrom
847 : mapFrom.getDatasetSequence();
848 SequenceI dsmapTo = mapTo.getDatasetSequence() == null ? mapTo
849 : mapTo.getDatasetSequence();
851 * look for a reverse mapping, if found make its inverse.
852 * Note - we do this on dataset sequences only.
854 if (dsmapTo.getDBRefs() != null)
856 for (DBRefEntry dbref : dsmapTo.getDBRefs())
858 String name = dbref.getSource() + "|" + dbref.getAccessionId();
859 if (dbref.hasMap() && dsmapFrom.getName().startsWith(name))
862 * looks like we've found a map from 'mapTo' to 'mapFrom'
863 * - invert it to make the mapping the other way
865 MapList reverse = dbref.getMap().getMap().getInverse();
866 xref.setMap(new Mapping(dsmapTo, reverse));
867 mappings.addMap(mapFrom, dsmapTo, reverse);
875 mapping = AlignmentUtils.mapCdnaToProtein(mapTo, mapFrom);
879 mapping = AlignmentUtils.mapCdnaToProtein(mapFrom, mapTo);
882 mapping = mapping.getInverse();
889 xref.setMap(new Mapping(mapTo, mapping));
892 * and add a reverse DbRef with the inverse mapping
894 if (mapFrom.getDatasetSequence() != null && false)
895 // && mapFrom.getDatasetSequence().getSourceDBRef() != null)
897 // possible need to search primary references... except, why doesn't xref
898 // == getSourceDBRef ??
899 // DBRefEntry dbref = new DBRefEntry(mapFrom.getDatasetSequence()
900 // .getSourceDBRef());
901 // dbref.setMap(new Mapping(mapFrom.getDatasetSequence(), mapping
903 // mapTo.addDBRef(dbref);
908 AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping);
909 mappings.addMap(mapFrom, mapTo, mapping);
913 mappings.addMap(mapTo, mapFrom, mapping.getInverse());
920 * find references to lrfs in the cross-reference set of each sequence in
921 * dataset (that is not equal to sequenceI) Identifies matching DBRefEntry
922 * based on source and accession string only - Map and Version are nulled.
925 * - true if context was searching from Dna sequences, false if
926 * context was searching from Protein sequences
930 * @return true if matches were found.
932 private boolean searchDatasetXrefs(boolean fromDna, SequenceI sequenceI,
933 DBRefEntry[] lrfs, List<SequenceI> foundSeqs, AlignedCodonFrame cf)
935 boolean found = false;
940 for (int i = 0; i < lrfs.length; i++)
942 DBRefEntry xref = new DBRefEntry(lrfs[i]);
944 xref.setVersion(null);
946 found |= searchDataset(fromDna, sequenceI, xref, foundSeqs, cf, false);
952 * Searches dataset for DBRefEntrys matching the given one (xrf) and adds the
953 * associated sequence to rseqs
956 * true if context was searching for refs *from* dna sequence, false
957 * if context was searching for refs *from* protein sequence
959 * a sequence to ignore (start point of search)
961 * a cross-reference to try to match
963 * result list to add to
965 * a set of sequence mappings to add to
967 * - indicates the type of relationship between returned sequences,
968 * xrf, and sequenceI that is required.
970 * <li>direct implies xrf is a primary reference for sequenceI AND
971 * the sequences to be located (eg a uniprot ID for a protein
972 * sequence, and a uniprot ref on a transcript sequence).</li>
973 * <li>indirect means xrf is a cross reference with respect to
974 * sequenceI or all the returned sequences (eg a genomic reference
975 * associated with a locus and one or more transcripts)</li>
977 * @return true if relationship found and sequence added.
979 boolean searchDataset(boolean fromDna, SequenceI fromSeq, DBRefEntry xrf,
980 List<SequenceI> foundSeqs, AlignedCodonFrame mappings,
983 boolean found = false;
988 if (dataset.getSequences() == null)
990 System.err.println("Empty dataset sequence set - NO VECTOR");
994 synchronized (ds = dataset.getSequences())
996 for (SequenceI nxt : ds)
1000 if (nxt.getDatasetSequence() != null)
1003 .println("Implementation warning: CrossRef initialised with a dataset alignment with non-dataset sequences in it! ("
1004 + nxt.getDisplayId(true)
1005 + " has ds reference "
1006 + nxt.getDatasetSequence().getDisplayId(true)
1009 if (nxt == fromSeq || nxt == fromSeq.getDatasetSequence())
1014 * only look at same molecule type if 'direct', or
1015 * complementary type if !direct
1018 boolean isDna = !nxt.isProtein();
1019 if (direct ? (isDna != fromDna) : (isDna == fromDna))
1021 // skip this sequence because it is wrong molecule type
1026 // look for direct or indirect references in common
1027 DBRefEntry[] poss = nxt.getDBRefs();
1028 List<DBRefEntry> cands = null;
1030 // todo: indirect specifies we select either direct references to nxt
1031 // that match xrf which is indirect to sequenceI, or indirect
1032 // references to nxt that match xrf which is direct to sequenceI
1033 cands = DBRefUtils.searchRefs(poss, xrf);
1036 // poss = DBRefUtils.selectDbRefs(nxt.isProtein()!fromDna, poss);
1037 // cands = DBRefUtils.searchRefs(poss, xrf);
1039 if (!cands.isEmpty())
1041 if (foundSeqs.contains(nxt))
1047 if (mappings != null && !direct)
1050 * if the matched sequence has mapped dbrefs to
1051 * protein product / cdna, add equivalent mappings to
1052 * our source sequence
1054 for (DBRefEntry candidate : cands)
1056 Mapping mapping = candidate.getMap();
1057 if (mapping != null)
1059 MapList map = mapping.getMap();
1060 if (mapping.getTo() != null
1061 && map.getFromRatio() != map.getToRatio())
1064 * add a mapping, as from dna to peptide sequence
1066 if (map.getFromRatio() == 3)
1068 mappings.addMap(nxt, fromSeq, map);
1072 mappings.addMap(nxt, fromSeq, map.getInverse());