X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FCrossRef.java;h=1783f376c2a2f0a735e90114babd3430932c760c;hb=b1f3e3bac931d7f309f60e50b18c435c65586de8;hp=71b0aa0b9adb5cd92d39b5f7908fa3dcc349e4ed;hpb=655b78299307682a4c7a6e5af0ed4618cbc9c924;p=jalview.git diff --git a/src/jalview/analysis/CrossRef.java b/src/jalview/analysis/CrossRef.java index 71b0aa0..1783f37 100644 --- a/src/jalview/analysis/CrossRef.java +++ b/src/jalview/analysis/CrossRef.java @@ -24,6 +24,7 @@ import jalview.datamodel.AlignedCodonFrame; import jalview.datamodel.Alignment; import jalview.datamodel.AlignmentI; import jalview.datamodel.DBRefEntry; +import jalview.datamodel.DBRefSource; import jalview.datamodel.Mapping; import jalview.datamodel.Sequence; import jalview.datamodel.SequenceFeature; @@ -98,7 +99,7 @@ public class CrossRef */ public List findXrefSourcesForSequences(boolean dna) { - List sources = new ArrayList(); + List sources = new ArrayList<>(); for (SequenceI seq : fromSeqs) { if (seq != null) @@ -106,6 +107,16 @@ public class CrossRef findXrefSourcesForSequence(seq, dna, sources); } } + sources.remove(DBRefSource.EMBL); // hack to prevent EMBL xrefs resulting in + // redundant datasets + if (dna) + { + sources.remove(DBRefSource.ENSEMBL); // hack to prevent Ensembl and + // EnsemblGenomes xref option shown + // from cdna panel + sources.remove(DBRefSource.ENSEMBLGENOMES); + } + // redundant datasets return sources; } @@ -132,15 +143,15 @@ public class CrossRef /* * first find seq's xrefs (dna-to-peptide or peptide-to-dna) */ - DBRefEntry[] rfs = DBRefUtils.selectDbRefs(!fromDna, seq.getDBRefs()); + List rfs = DBRefUtils.selectDbRefs(!fromDna, seq.getDBRefs()); addXrefsToSources(rfs, sources); if (dataset != null) { /* * find sequence's direct (dna-to-dna, peptide-to-peptide) xrefs */ - DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(fromDna, seq.getDBRefs()); - List foundSeqs = new ArrayList(); + List lrfs = DBRefUtils.selectDbRefs(fromDna, seq.getDBRefs()); + List foundSeqs = new ArrayList<>(); /* * find sequences in the alignment which xref one of these DBRefs @@ -153,8 +164,8 @@ public class CrossRef */ for (SequenceI rs : foundSeqs) { - DBRefEntry[] xrs = DBRefUtils - .selectDbRefs(!fromDna, rs.getDBRefs()); + List xrs = DBRefUtils.selectDbRefs(!fromDna, + rs.getDBRefs()); addXrefsToSources(xrs, sources); } } @@ -167,7 +178,7 @@ public class CrossRef * @param xrefs * @param sources */ - void addXrefsToSources(DBRefEntry[] xrefs, List sources) + void addXrefsToSources(List xrefs, List sources) { if (xrefs != null) { @@ -207,10 +218,9 @@ public class CrossRef public Alignment findXrefSequences(String source, boolean fromDna) { - rseqs = new ArrayList(); + rseqs = new ArrayList<>(); AlignedCodonFrame cf = new AlignedCodonFrame(); - matcher = new SequenceIdMatcher( - dataset.getSequences()); + matcher = new SequenceIdMatcher(dataset.getSequences()); for (SequenceI seq : fromSeqs) { @@ -220,18 +230,18 @@ public class CrossRef dss = dss.getDatasetSequence(); } boolean found = false; - DBRefEntry[] xrfs = DBRefUtils - .selectDbRefs(!fromDna, dss.getDBRefs()); + List xrfs = DBRefUtils.selectDbRefs(!fromDna, + dss.getDBRefs()); // ENST & ENSP comes in to both Protein and nucleotide, so we need to // filter them // out later. - if ((xrfs == null || xrfs.length == 0) && dataset != null) + if ((xrfs == null || xrfs.size() == 0) && dataset != null) { /* * found no suitable dbrefs on sequence - look for sequences in the * alignment which share a dbref with this one */ - DBRefEntry[] lrfs = DBRefUtils.selectDbRefs(fromDna, + List lrfs = DBRefUtils.selectDbRefs(fromDna, seq.getDBRefs()); /* @@ -281,13 +291,10 @@ public class CrossRef if (matchInDataset != null && xref.getMap().getTo() != null && matchInDataset != xref.getMap().getTo()) { - System.err - .println("Implementation problem (reopen JAL-2154): CrossRef.findInDataset seems to have recovered a different sequence than the one explicitly mapped for xref." - + "Found:" - + matchInDataset - + "\nExpected:" - + xref.getMap().getTo() - + "\nFor xref:" + System.err.println( + "Implementation problem (reopen JAL-2154): CrossRef.findInDataset seems to have recovered a different sequence than the one explicitly mapped for xref." + + "Found:" + matchInDataset + "\nExpected:" + + xref.getMap().getTo() + "\nFor xref:" + xref); } /*matcher.findIdMatch(mappedTo);*/ @@ -313,8 +320,9 @@ public class CrossRef } else { - cf.addMap(matchInDataset, dss, xref.getMap().getMap() - .getInverse(), xref.getMap().getMappedFromId()); + cf.addMap(matchInDataset, dss, + xref.getMap().getMap().getInverse(), + xref.getMap().getMappedFromId()); } } @@ -330,8 +338,8 @@ public class CrossRef if (fromDna) { // map is from dna seq to a protein product - cf.addMap(dss, rsq, xref.getMap().getMap(), xref.getMap() - .getMappedFromId()); + cf.addMap(dss, rsq, xref.getMap().getMap(), + xref.getMap().getMappedFromId()); } else { @@ -345,8 +353,8 @@ public class CrossRef if (!found) { - SequenceI matchedSeq = matcher.findIdMatch(xref.getSource() + "|" - + xref.getAccessionId()); + SequenceI matchedSeq = matcher.findIdMatch( + xref.getSource() + "|" + xref.getAccessionId()); // if there was a match, check it's at least the right type of // molecule! if (matchedSeq != null && matchedSeq.isProtein() == fromDna) @@ -362,7 +370,7 @@ public class CrossRef { // do a bit more work - search for sequences with references matching // xrefs on this sequence. - found = searchDataset(fromDna, dss, xref, rseqs, cf, false); + found = searchDataset(fromDna, dss, xref, rseqs, cf, false, DBRefUtils.SEARCH_MODE_FULL); } if (found) { @@ -392,12 +400,12 @@ public class CrossRef } private void retrieveCrossRef(List sourceRefs, SequenceI seq, - DBRefEntry[] xrfs, boolean fromDna, AlignedCodonFrame cf) + List xrfs, boolean fromDna, AlignedCodonFrame cf) { ASequenceFetcher sftch = SequenceFetcherFactory.getSequenceFetcher(); SequenceI[] retrieved = null; - SequenceI dss = seq.getDatasetSequence() == null ? seq : seq - .getDatasetSequence(); + SequenceI dss = seq.getDatasetSequence() == null ? seq + : seq.getDatasetSequence(); // first filter in case we are retrieving crossrefs that have already been // retrieved. this happens for cases where a database record doesn't yield // protein products for CDS @@ -413,8 +421,8 @@ public class CrossRef retrieved = sftch.getSequences(sourceRefs, !fromDna); } catch (Exception e) { - System.err - .println("Problem whilst retrieving cross references for Sequence : " + System.err.println( + "Problem whilst retrieving cross references for Sequence : " + seq.getName()); e.printStackTrace(); } @@ -422,14 +430,24 @@ public class CrossRef if (retrieved != null) { boolean addedXref = false; + List newDsSeqs = new ArrayList<>(), + doNotAdd = new ArrayList<>(); + for (SequenceI retrievedSequence : retrieved) { // dataset gets contaminated ccwith non-ds sequences. why ??! // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL-> - SequenceI retrievedDss = retrievedSequence.getDatasetSequence() == null ? retrievedSequence - : retrievedSequence.getDatasetSequence(); - addedXref |= importCrossRefSeq(cf, dss, retrievedDss); + SequenceI retrievedDss = retrievedSequence + .getDatasetSequence() == null ? retrievedSequence + : retrievedSequence.getDatasetSequence(); + addedXref |= importCrossRefSeq(cf, newDsSeqs, doNotAdd, dss, + retrievedDss); } + // JBPNote: What assumptions are made for dbref structures on + // retrieved sequences ? + // addedXref will be true means importCrossRefSeq found + // sequences with dbrefs with mappings to sequences congruent with dss + if (!addedXref) { // try again, after looking for matching IDs @@ -439,9 +457,20 @@ public class CrossRef { // dataset gets contaminated ccwith non-ds sequences. why ??! // try: Ensembl -> Nuc->Ensembl, Nuc->Uniprot-->Protein->EMBL-> - SequenceI retrievedDss = retrievedSequence.getDatasetSequence() == null ? retrievedSequence - : retrievedSequence.getDatasetSequence(); - addedXref |= importCrossRefSeq(cf, dss, retrievedDss); + SequenceI retrievedDss = retrievedSequence + .getDatasetSequence() == null ? retrievedSequence + : retrievedSequence.getDatasetSequence(); + addedXref |= importCrossRefSeq(cf, newDsSeqs, doNotAdd, dss, + retrievedDss); + } + } + for (SequenceI newToSeq : newDsSeqs) + { + if (!doNotAdd.contains(newToSeq) + && dataset.findIndex(newToSeq) == -1) + { + dataset.addSequence(newToSeq); + matcher.add(newToSeq); } } } @@ -459,19 +488,24 @@ public class CrossRef private void removeAlreadyRetrievedSeqs(List sourceRefs, boolean fromDna) { - DBRefEntry[] dbrSourceSet = sourceRefs.toArray(new DBRefEntry[0]); - for (SequenceI sq : dataset.getSequences()) + List dbrSourceSet = new ArrayList<>(sourceRefs); + List dsSeqs = dataset.getSequences(); + for (int ids = 0, nds = dsSeqs.size(); ids < nds; ids++) { + SequenceI sq = dsSeqs.get(ids); boolean dupeFound = false; // !fromDna means we are looking only for nucleotide sequences, not // protein if (sq.isProtein() == fromDna) { - for (DBRefEntry dbr : sq.getPrimaryDBRefs()) + List sqdbrefs = sq.getPrimaryDBRefs(); + for (int idb = 0, ndb = sqdbrefs.size(); idb < ndb; idb++) { - for (DBRefEntry found : DBRefUtils.searchRefs(dbrSourceSet, dbr)) + DBRefEntry dbr = sqdbrefs.get(idb); + List searchrefs = DBRefUtils.searchRefs(dbrSourceSet, dbr, DBRefUtils.SEARCH_MODE_FULL); + for (int isr = 0, nsr = searchrefs.size(); isr < nsr; isr++) { - sourceRefs.remove(found); + sourceRefs.remove(searchrefs.get(isr)); dupeFound = true; } } @@ -479,14 +513,17 @@ public class CrossRef if (dupeFound) { // rebuild the search array from the filtered sourceRefs list - dbrSourceSet = sourceRefs.toArray(new DBRefEntry[0]); + dbrSourceSet.clear(); + dbrSourceSet.addAll(sourceRefs); } } } /** * process sequence retrieved via a dbref on source sequence to resolve and - * transfer data + * transfer data JBPNote: as of 2022-02-03 - this assumes retrievedSequence + * has dbRefs with Mapping references to a sequence congruent with + * sourceSequence * * @param cf * @param sourceSequence @@ -494,6 +531,7 @@ public class CrossRef * @return true if retrieveSequence was imported */ private boolean importCrossRefSeq(AlignedCodonFrame cf, + List newDsSeqs, List doNotAdd, SequenceI sourceSequence, SequenceI retrievedSequence) { /** @@ -501,12 +539,14 @@ public class CrossRef * sourceSequence */ boolean imported = false; - DBRefEntry[] dbr = retrievedSequence.getDBRefs(); - List newDsSeqs = new ArrayList(); + List dbr = retrievedSequence.getDBRefs(); if (dbr != null) { - for (DBRefEntry dbref : dbr) + for (int ib = 0, nb = dbr.size(); ib < nb; ib++) { + + DBRefEntry dbref = dbr.get(ib); + // matched will return null if the dbref has no map SequenceI matched = findInDataset(dbref); if (matched == sourceSequence) { @@ -518,9 +558,10 @@ public class CrossRef Mapping map = dbref.getMap(); if (map != null) { - if (map.getTo() != null && map.getMap() != null) + SequenceI ms = map.getTo(); + if (ms != null && map.getMap() != null) { - if (map.getTo() == sourceSequence) + if (ms == sourceSequence) { // already called to import once, and most likely this sequence // already imported ! @@ -531,7 +572,7 @@ public class CrossRef /* * sequence is new to dataset, so save a reference so it can be added. */ - newDsSeqs.add(map.getTo()); + newDsSeqs.add(ms); continue; } @@ -543,7 +584,6 @@ public class CrossRef { // compare ms with dss and replace with dss in mapping // if map is congruent - SequenceI ms = map.getTo(); // TODO findInDataset requires exact sequence match but // 'congruent' test is only for the mapped part // maybe not a problem in practice since only ENA provide a @@ -554,9 +594,8 @@ public class CrossRef int sf = map.getMap().getToLowest(); int st = map.getMap().getToHighest(); SequenceI mappedrg = ms.getSubSequence(sf, st); - if (mappedrg.getLength() > 0 - && ms.getSequenceAsString().equals( - matched.getSequenceAsString())) + if (mappedrg.getLength() > 0 && ms.getSequenceAsString() + .equals(matched.getSequenceAsString())) { /* * sequences were a match, @@ -566,7 +605,7 @@ public class CrossRef + matched.getName(); System.out.println(msg); - DBRefEntry[] toRefs = map.getTo().getDBRefs(); + List toRefs = map.getTo().getDBRefs(); if (toRefs != null) { /* @@ -574,9 +613,15 @@ public class CrossRef */ for (DBRefEntry ref : toRefs) { + if (dbref.getSrcAccString() + .equals(ref.getSrcAccString())) + { + continue; // avoid overwriting the ref on source sequence + } matched.addDBRef(ref); // add or update mapping } } + doNotAdd.add(map.getTo()); map.setTo(matched); /* @@ -590,35 +635,31 @@ public class CrossRef * duplication (e.g. same variation from two * transcripts) */ - SequenceFeature[] sfs = ms.getSequenceFeatures(); - if (sfs != null) + List sfs = ms.getFeatures() + .getAllFeatures(); + for (SequenceFeature feat : sfs) { - for (SequenceFeature feat : sfs) + /* + * make a flyweight feature object which ignores Parent + * attribute in equality test; this avoids creating many + * otherwise duplicate exon features on genomic sequence + */ + SequenceFeature newFeature = new SequenceFeature(feat) { - /* - * make a flyweight feature object which ignores Parent - * attribute in equality test; this avoids creating many - * otherwise duplicate exon features on genomic sequence - */ - SequenceFeature newFeature = new SequenceFeature( - feat) + @Override + public boolean equals(Object o) { - @Override - public boolean equals(Object o) - { - return super.equals(o, true); - } - }; - matched.addSequenceFeature(newFeature); - } + return super.equals(o, true); + } + }; + matched.addSequenceFeature(newFeature); } - } cf.addMap(retrievedSequence, map.getTo(), map.getMap()); } catch (Exception e) { - System.err - .println("Exception when consolidating Mapped sequence set..."); + System.err.println( + "Exception when consolidating Mapped sequence set..."); e.printStackTrace(System.err); } } @@ -634,18 +675,10 @@ public class CrossRef dataset.addSequence(retrievedSequence); matcher.add(retrievedSequence); } - for (SequenceI newToSeq : newDsSeqs) - { - - if (dataset.findIndex(newToSeq) == -1) - { - dataset.addSequence(newToSeq); - matcher.add(newToSeq); - } - } } return imported; } + /** * Sets the inverse sequence mapping in the corresponding dbref of the mapped * to sequence (if any). This is used after fetching a cross-referenced @@ -665,7 +698,7 @@ public class CrossRef { return; } - DBRefEntry[] dbrefs = mapTo.getDBRefs(); + List dbrefs = mapTo.getDBRefs(); if (dbrefs == null) { return; @@ -691,7 +724,7 @@ public class CrossRef * Returns null or the first sequence in the dataset which is identical to * xref.mapTo, and has a) a primary dbref matching xref, or if none found, the * first one with an ID source|xrefacc - * + * JBPNote: Could refactor this to AlignmentI/DatasetI * @param xref * with map and mapped-to sequence * @return @@ -705,8 +738,8 @@ public class CrossRef SequenceI mapsTo = xref.getMap().getTo(); String name = xref.getAccessionId(); String name2 = xref.getSource() + "|" + name; - SequenceI dss = mapsTo.getDatasetSequence() == null ? mapsTo : mapsTo - .getDatasetSequence(); + SequenceI dss = mapsTo.getDatasetSequence() == null ? mapsTo + : mapsTo.getDatasetSequence(); // first check ds if ds is directly referenced if (dataset.findIndex(dss) > -1) { @@ -721,8 +754,8 @@ public class CrossRef for (SequenceI seq : dataset.getSequences()) { // first check primary refs. - List match = DBRefUtils.searchRefs(seq.getPrimaryDBRefs() - .toArray(new DBRefEntry[0]), template); + List match = DBRefUtils.searchRefs( + seq.getPrimaryDBRefs(), template, DBRefUtils.SEARCH_MODE_FULL); if (match != null && match.size() == 1 && sameSequence(seq, dss)) { return seq; @@ -732,9 +765,8 @@ public class CrossRef * returns sequences with a dbref to the matched accession id * which we don't want */ - if (firstIdMatch == null - && (name.equals(seq.getName()) || seq.getName().startsWith( - name2))) + if (firstIdMatch == null && (name.equals(seq.getName()) + || seq.getName().startsWith(name2))) { if (sameSequence(seq, dss)) { @@ -766,15 +798,15 @@ public class CrossRef { return false; } - char[] c1 = seq1.getSequence(); - char[] c2 = seq2.getSequence(); - if (c1.length != c2.length) + + if (seq1.getLength() != seq2.getLength()) { return false; } - for (int i = 0; i < c1.length; i++) + int length = seq1.getLength(); + for (int i = 0; i < length; i++) { - int diff = c1[i] - c2[i]; + int diff = seq1.getCharAt(i) - seq2.getCharAt(i); /* * same char or differ in case only ('a'-'A' == 32) */ @@ -790,13 +822,13 @@ public class CrossRef * Updates any empty mappings in the cross-references with one to a compatible * retrieved sequence if found, and adds any new mappings to the * AlignedCodonFrame - * + * JBPNote: TODO: this relies on sequence IDs like UNIPROT|ACCESSION - which do not always happen. * @param mapFrom * @param xrefs * @param retrieved * @param acf */ - void updateDbrefMappings(SequenceI mapFrom, DBRefEntry[] xrefs, + void updateDbrefMappings(SequenceI mapFrom, List xrefs, SequenceI[] retrieved, AlignedCodonFrame acf, boolean fromDna) { SequenceIdMatcher idMatcher = new SequenceIdMatcher(retrieved); @@ -905,7 +937,7 @@ public class CrossRef if (fromDna) { - AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping); + // AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping); mappings.addMap(mapFrom, mapTo, mapping); } else @@ -930,20 +962,22 @@ public class CrossRef * @return true if matches were found. */ private boolean searchDatasetXrefs(boolean fromDna, SequenceI sequenceI, - DBRefEntry[] lrfs, List foundSeqs, AlignedCodonFrame cf) + List lrfs, List foundSeqs, + AlignedCodonFrame cf) { boolean found = false; if (lrfs == null) { return false; } - for (int i = 0; i < lrfs.length; i++) + for (int i = 0, n = lrfs.size(); i < n; i++) { - DBRefEntry xref = new DBRefEntry(lrfs[i]); - // add in wildcards - xref.setVersion(null); - xref.setMap(null); - found |= searchDataset(fromDna, sequenceI, xref, foundSeqs, cf, false); +// DBRefEntry xref = new DBRefEntry(lrfs.get(i)); +// // add in wildcards +// xref.setVersion(null); +// xref.setMap(null); + found |= searchDataset(fromDna, sequenceI, lrfs.get(i), foundSeqs, cf, + false, DBRefUtils.SEARCH_MODE_NO_MAP_NO_VERSION); } return found; } @@ -974,11 +1008,12 @@ public class CrossRef * sequenceI or all the returned sequences (eg a genomic reference * associated with a locus and one or more transcripts) * + * @param mode SEARCH_MODE_FULL for all; SEARCH_MODE_NO_MAP_NO_VERSION optional * @return true if relationship found and sequence added. */ boolean searchDataset(boolean fromDna, SequenceI fromSeq, DBRefEntry xrf, List foundSeqs, AlignedCodonFrame mappings, - boolean direct) + boolean direct, int mode) { boolean found = false; if (dataset == null) @@ -990,8 +1025,8 @@ public class CrossRef System.err.println("Empty dataset sequence set - NO VECTOR"); return false; } - List ds; - synchronized (ds = dataset.getSequences()) + List ds = dataset.getSequences(); + synchronized (ds) { for (SequenceI nxt : ds) { @@ -999,10 +1034,9 @@ public class CrossRef { if (nxt.getDatasetSequence() != null) { - System.err - .println("Implementation warning: CrossRef initialised with a dataset alignment with non-dataset sequences in it! (" - + nxt.getDisplayId(true) - + " has ds reference " + System.err.println( + "Implementation warning: CrossRef initialised with a dataset alignment with non-dataset sequences in it! (" + + nxt.getDisplayId(true) + " has ds reference " + nxt.getDatasetSequence().getDisplayId(true) + ")"); } @@ -1024,13 +1058,13 @@ public class CrossRef } // look for direct or indirect references in common - DBRefEntry[] poss = nxt.getDBRefs(); + List poss = nxt.getDBRefs(); List cands = null; // todo: indirect specifies we select either direct references to nxt // that match xrf which is indirect to sequenceI, or indirect // references to nxt that match xrf which is direct to sequenceI - cands = DBRefUtils.searchRefs(poss, xrf); + cands = DBRefUtils.searchRefs(poss, xrf, mode); // else // { // poss = DBRefUtils.selectDbRefs(nxt.isProtein()!fromDna, poss);