X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FCrossRef.java;h=87f854a4fc9eb764886f999dcc82d14173fc5d80;hb=dc82563dc0d6752da31ade9031035854e5c33409;hp=0e7662b4a9566b35e6bc3eb6a6eabc2b13112882;hpb=cd1dfe669b5d4d155cdb5f4a97554ae91ab10cf9;p=jalview.git diff --git a/src/jalview/analysis/CrossRef.java b/src/jalview/analysis/CrossRef.java index 0e7662b..87f854a 100644 --- a/src/jalview/analysis/CrossRef.java +++ b/src/jalview/analysis/CrossRef.java @@ -20,6 +20,10 @@ */ package jalview.analysis; +import java.util.ArrayList; +import java.util.Iterator; +import java.util.List; + import jalview.datamodel.AlignedCodonFrame; import jalview.datamodel.Alignment; import jalview.datamodel.AlignmentI; @@ -31,12 +35,7 @@ import jalview.datamodel.SequenceFeature; import jalview.datamodel.SequenceI; import jalview.util.DBRefUtils; import jalview.util.MapList; -import jalview.ws.SequenceFetcherFactory; -import jalview.ws.seqfetcher.ASequenceFetcher; - -import java.util.ArrayList; -import java.util.Iterator; -import java.util.List; +import jalview.ws.SequenceFetcher; /** * Functions for cross-referencing sequence databases. @@ -143,14 +142,16 @@ 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 lrfs = DBRefUtils.selectDbRefs(fromDna, + seq.getDBRefs()); List foundSeqs = new ArrayList<>(); /* @@ -164,7 +165,7 @@ public class CrossRef */ for (SequenceI rs : foundSeqs) { - DBRefEntry[] xrs = DBRefUtils.selectDbRefs(!fromDna, + List xrs = DBRefUtils.selectDbRefs(!fromDna, rs.getDBRefs()); addXrefsToSources(xrs, sources); } @@ -178,7 +179,7 @@ public class CrossRef * @param xrefs * @param sources */ - void addXrefsToSources(DBRefEntry[] xrefs, List sources) + void addXrefsToSources(List xrefs, List sources) { if (xrefs != null) { @@ -230,18 +231,18 @@ public class CrossRef dss = dss.getDatasetSequence(); } boolean found = false; - DBRefEntry[] xrfs = DBRefUtils.selectDbRefs(!fromDna, + 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()); /* @@ -370,7 +371,8 @@ 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) { @@ -400,9 +402,8 @@ 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(); @@ -418,7 +419,7 @@ public class CrossRef } try { - retrieved = sftch.getSequences(sourceRefs, !fromDna); + retrieved = SequenceFetcher.getInstance().getSequences(sourceRefs, !fromDna); } catch (Exception e) { System.err.println( @@ -443,6 +444,11 @@ public class CrossRef 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 @@ -483,19 +489,25 @@ 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; } } @@ -503,14 +515,17 @@ 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 @@ -526,11 +541,14 @@ public class CrossRef * sourceSequence */ boolean imported = false; - DBRefEntry[] dbr = retrievedSequence.getDBRefs(); + 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) { @@ -542,9 +560,10 @@ 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 ! @@ -555,7 +574,7 @@ 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; } @@ -567,7 +586,6 @@ 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 @@ -589,7 +607,7 @@ public class CrossRef + matched.getName(); System.out.println(msg); - DBRefEntry[] toRefs = map.getTo().getDBRefs(); + List toRefs = map.getTo().getDBRefs(); if (toRefs != null) { /* @@ -682,7 +700,7 @@ public class CrossRef { return; } - DBRefEntry[] dbrefs = mapTo.getDBRefs(); + List dbrefs = mapTo.getDBRefs(); if (dbrefs == null) { return; @@ -707,7 +725,8 @@ 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 + * first one with an ID source|xrefacc JBPNote: Could refactor this to + * AlignmentI/DatasetI * * @param xref * with map and mapped-to sequence @@ -738,8 +757,8 @@ 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; @@ -805,14 +824,15 @@ 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 + * 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); @@ -921,7 +941,7 @@ public class CrossRef if (fromDna) { - AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping); + // AlignmentUtils.computeProteinFeatures(mapFrom, mapTo, mapping); mappings.addMap(mapFrom, mapTo, mapping); } else @@ -946,7 +966,7 @@ public class CrossRef * @return true if matches were found. */ private boolean searchDatasetXrefs(boolean fromDna, SequenceI sequenceI, - DBRefEntry[] lrfs, List foundSeqs, + List lrfs, List foundSeqs, AlignedCodonFrame cf) { boolean found = false; @@ -954,14 +974,14 @@ public class CrossRef { 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; } @@ -992,11 +1012,13 @@ 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) @@ -1041,13 +1063,13 @@ 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);