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.ws.SequenceFetcher;
34 import jalview.ws.seqfetcher.ASequenceFetcher;
36 import java.util.ArrayList;
37 import java.util.List;
38 import java.util.Vector;
41 * Functions for cross-referencing sequence databases. user must first specify
42 * if cross-referencing from protein or dna (set dna==true)
50 * A sub-class that ignores Parent attribute when comparing sequence
51 * features. This avoids 'duplicate' CDS features that only
52 * differ in their parent Transcript ids.
54 class MySequenceFeature extends SequenceFeature
56 private SequenceFeature feat;
58 MySequenceFeature(SequenceFeature sf)
64 public boolean equals(Object o)
66 return feat.equals(o, true);
71 * Select just the DNA or protein references for a protein or dna sequence
74 * if true, select references from DNA (i.e. Protein databases), else
75 * DNA database references
77 * a set of references to select from
80 public static DBRefEntry[] findXDbRefs(boolean fromDna, DBRefEntry[] refs)
82 return DBRefUtils.selectRefs(refs, fromDna ? DBRefSource.PROTEINDBS
83 : DBRefSource.DNACODINGDBS);
84 // could attempt to find other cross
85 // refs here - ie PDB xrefs
86 // (not dna, not protein seq)
91 * true if seqs are DNA seqs
93 * @return a list of sequence database cross reference source types
95 public static String[] findSequenceXrefTypes(boolean dna, SequenceI[] seqs)
97 return findSequenceXrefTypes(dna, seqs, null);
101 * Indirect references are references from other sequences from the dataset to
102 * any of the direct DBRefEntrys on the given sequences.
105 * true if seqs are DNA seqs
107 * @return a list of sequence database cross reference source types
109 public static String[] findSequenceXrefTypes(boolean dna,
110 SequenceI[] seqs, AlignmentI dataset)
112 String[] dbrefs = null;
113 List<String> refs = new ArrayList<String>();
114 for (SequenceI seq : seqs)
119 while (dss.getDatasetSequence() != null)
121 dss = dss.getDatasetSequence();
123 DBRefEntry[] rfs = findXDbRefs(dna, dss.getDBRefs());
126 for (DBRefEntry ref : rfs)
128 if (!refs.contains(ref.getSource()))
130 refs.add(ref.getSource());
136 // search for references to this sequence's direct references.
137 DBRefEntry[] lrfs = CrossRef.findXDbRefs(!dna, seq.getDBRefs());
138 List<SequenceI> rseqs = new ArrayList<SequenceI>();
139 CrossRef.searchDatasetXrefs(seq, !dna, lrfs, dataset, rseqs,
140 null); // don't need to specify codon frame for mapping here
141 for (SequenceI rs : rseqs)
143 DBRefEntry[] xrs = findXDbRefs(dna, rs.getDBRefs());
146 for (DBRefEntry ref : xrs)
148 if (!refs.contains(ref.getSource()))
150 refs.add(ref.getSource());
154 // looks like copy and paste - change rfs to xrs?
155 // for (int r = 0; rfs != null && r < rfs.length; r++)
157 // if (!refs.contains(rfs[r].getSource()))
159 // refs.add(rfs[r].getSource());
168 dbrefs = new String[refs.size()];
169 refs.toArray(dbrefs);
174 public static boolean hasCdnaMap(SequenceI[] seqs)
176 // TODO unused - remove?
177 String[] reftypes = findSequenceXrefTypes(false, seqs);
178 for (int s = 0; s < reftypes.length; s++)
180 if (reftypes.equals(DBRefSource.EMBLCDS))
189 public static SequenceI[] getCdnaMap(SequenceI[] seqs)
191 // TODO unused - remove?
192 Vector cseqs = new Vector();
193 for (int s = 0; s < seqs.length; s++)
195 DBRefEntry[] cdna = findXDbRefs(true, seqs[s].getDBRefs());
196 for (int c = 0; c < cdna.length; c++)
198 if (cdna[c].getSource().equals(DBRefSource.EMBLCDS))
201 .println("TODO: unimplemented sequence retrieval for coding region sequence.");
202 // TODO: retrieve CDS dataset sequences
203 // need global dataset sequence retriever/resolver to reuse refs
204 // and construct Mapping entry.
205 // insert gaps in CDS according to peptide gaps.
206 // add gapped sequence to cseqs
210 if (cseqs.size() > 0)
212 SequenceI[] rsqs = new SequenceI[cseqs.size()];
213 cseqs.copyInto(rsqs);
226 public static Alignment findXrefSequences(SequenceI[] seqs, boolean dna,
229 return findXrefSequences(seqs, dna, source, null);
235 * sequences whose xrefs are being retrieved
237 * true if sequences are nucleotide
240 * alignment to search for product sequences.
241 * @return products (as dataset sequences)
243 public static Alignment findXrefSequences(SequenceI[] seqs, boolean dna,
244 String source, AlignmentI dataset)
246 List<SequenceI> rseqs = new ArrayList<SequenceI>();
247 AlignedCodonFrame cf = new AlignedCodonFrame();
248 for (SequenceI seq : seqs)
251 while (dss.getDatasetSequence() != null)
253 dss = dss.getDatasetSequence();
255 boolean found = false;
256 DBRefEntry[] xrfs = CrossRef.findXDbRefs(dna, dss.getDBRefs());
257 if ((xrfs == null || xrfs.length == 0) && dataset != null)
259 System.out.println("Attempting to find ds Xrefs refs.");
260 // FIXME should be dss not seq here?
261 DBRefEntry[] lrfs = CrossRef.findXDbRefs(!dna, seq.getDBRefs());
262 // less ambiguous would be a 'find primary dbRefEntry' method.
263 // filter for desired source xref here
264 found = CrossRef.searchDatasetXrefs(dss, !dna, lrfs, dataset,
267 for (int r = 0; xrfs != null && r < xrfs.length; r++)
269 DBRefEntry xref = xrfs[r];
270 if (source != null && !source.equals(xref.getSource()))
276 if (xref.getMap().getTo() != null)
278 SequenceI rsq = new Sequence(xref.getMap().getTo());
280 if (xref.getMap().getMap().getFromRatio() != xref
281 .getMap().getMap().getToRatio())
283 // get sense of map correct for adding to product alignment.
286 // map is from dna seq to a protein product
287 cf.addMap(dss, rsq, xref.getMap().getMap());
291 // map should be from protein seq to its coding dna
292 cf.addMap(rsq, dss, xref.getMap().getMap().getInverse());
300 // do a bit more work - search for sequences with references matching
301 // xrefs on this sequence.
304 found |= searchDataset(dss, xref, dataset, rseqs, cf); // ,false,!dna);
307 xrfs[r] = null; // we've recovered seqs for this one.
314 if (xrfs != null && xrfs.length > 0)
316 // Try and get the sequence reference...
318 * Ideal world - we ask for a sequence fetcher implementation here if
319 * (jalview.io.RunTimeEnvironment.getSequenceFetcher()) (
321 ASequenceFetcher sftch = new SequenceFetcher();
322 SequenceI[] retrieved = null;
324 for (int r = 0; r < xrfs.length; r++)
326 // filter out any irrelevant or irretrievable references
328 || ((source != null && !source.equals(xrfs[r]
329 .getSource())) || !sftch.isFetchable(xrfs[r]
339 .println("Attempting to retrieve cross referenced sequences.");
340 DBRefEntry[] t = new DBRefEntry[l];
342 for (int r = 0; r < xrfs.length; r++)
352 retrieved = sftch.getSequences(xrfs, !dna);
353 // problem here is we don't know which of xrfs resulted in which
355 } catch (Exception e)
358 .println("Problem whilst retrieving cross references for Sequence : "
363 if (retrieved != null)
365 List<SequenceFeature> copiedFeatures = new ArrayList<SequenceFeature>();
366 CrossRef me = new CrossRef();
367 for (int rs = 0; rs < retrieved.length; rs++)
369 // TODO: examine each sequence for 'redundancy'
370 DBRefEntry[] dbr = retrieved[rs].getDBRefs();
371 if (dbr != null && dbr.length > 0)
373 for (int di = 0; di < dbr.length; di++)
375 // find any entry where we should put in the sequence being
376 // cross-referenced into the map
377 Mapping map = dbr[di].getMap();
380 if (map.getTo() != null && map.getMap() != null)
382 // should search the local dataset to find any existing
383 // candidates for To !
386 // compare ms with dss and replace with dss in mapping
387 // if map is congruent
388 SequenceI ms = map.getTo();
389 int sf = map.getMap().getToLowest();
390 int st = map.getMap().getToHighest();
391 SequenceI mappedrg = ms.getSubSequence(sf, st);
392 SequenceI loc = dss.getSubSequence(sf, st);
393 if (mappedrg.getLength() > 0
394 && mappedrg.getSequenceAsString().equals(
395 loc.getSequenceAsString()))
397 String msg = "Mapping updated from "
399 + " to retrieved crossreference "
401 System.out.println(msg);
402 // method to update all refs of existing To on
403 // retrieved sequence with dss and merge any props
407 * copy sequence features as well, avoiding
408 * duplication (e.g. from 2 transcripts)
410 SequenceFeature[] sfs = ms
411 .getSequenceFeatures();
414 for (SequenceFeature feat : sfs)
417 * we override the equality test here (but not
418 * elsewhere) to ignore Parent attribute
419 * TODO not quite working yet!
422 .contains(me.new MySequenceFeature(
425 dss.addSequenceFeature(feat);
426 copiedFeatures.add(feat);
430 cf.addMap(retrieved[rs].getDatasetSequence(),
433 } catch (Exception e)
436 .println("Exception when consolidating Mapped sequence set...");
437 e.printStackTrace(System.err);
443 retrieved[rs].updatePDBIds();
444 rseqs.add(retrieved[rs]);
452 Alignment ral = null;
453 if (rseqs.size() > 0)
455 SequenceI[] rsqs = new SequenceI[rseqs.size()];
457 ral = new Alignment(rsqs);
458 if (cf != null && !cf.isEmpty())
460 ral.addCodonFrame(cf);
467 * find references to lrfs in the cross-reference set of each sequence in
468 * dataset (that is not equal to sequenceI) Identifies matching DBRefEntry
469 * based on source and accession string only - Map and Version are nulled.
475 * @return true if matches were found.
477 private static boolean searchDatasetXrefs(SequenceI sequenceI,
478 boolean dna, DBRefEntry[] lrfs, AlignmentI dataset,
479 List<SequenceI> rseqs, AlignedCodonFrame cf)
481 boolean found = false;
486 for (int i = 0; i < lrfs.length; i++)
488 DBRefEntry xref = new DBRefEntry(lrfs[i]);
490 xref.setVersion(null);
492 found = searchDataset(sequenceI, xref, dataset, rseqs, cf, false, dna);
498 * search a given sequence dataset for references matching cross-references to
505 * set of unique sequences
507 * @return true if one or more unique sequences were found and added
509 public static boolean searchDataset(SequenceI sequenceI, DBRefEntry xrf,
510 AlignmentI dataset, List<SequenceI> rseqs, AlignedCodonFrame cf)
512 return searchDataset(sequenceI, xrf, dataset, rseqs, cf, true, false);
516 * TODO: generalise to different protein classifications Search dataset for
517 * DBRefEntrys matching the given one (xrf) and add the associated sequence to
525 * - search all references or only subset
527 * search dna or protein xrefs (if direct=false)
528 * @return true if relationship found and sequence added.
530 public static boolean searchDataset(SequenceI sequenceI, DBRefEntry xrf,
531 AlignmentI dataset, List<SequenceI> rseqs, AlignedCodonFrame cf,
532 boolean direct, boolean dna)
534 boolean found = false;
535 SequenceI[] typer = new SequenceI[1];
540 if (dataset.getSequences() == null)
542 System.err.println("Empty dataset sequence set - NO VECTOR");
546 synchronized (ds = dataset.getSequences())
548 for (SequenceI nxt : ds)
552 if (nxt.getDatasetSequence() != null)
555 .println("Implementation warning: getProducts passed a dataset alignment without dataset sequences in it!");
557 if (nxt != sequenceI && nxt != sequenceI.getDatasetSequence())
559 // check if this is the correct sequence type
562 boolean isDna = jalview.util.Comparison.isNucleotide(typer);
563 if ((direct && isDna == dna) || (!direct && isDna != dna))
565 // skip this sequence because it is same molecule type
570 // look for direct or indirect references in common
571 DBRefEntry[] poss = nxt.getDBRefs(), cands = null;
574 cands = jalview.util.DBRefUtils.searchRefs(poss, xrf);
578 poss = CrossRef.findXDbRefs(dna, poss); //
579 cands = jalview.util.DBRefUtils.searchRefs(poss, xrf);
583 if (!rseqs.contains(nxt))
586 boolean foundmap = cf != null;
587 // don't search if we aren't given a codon map object
588 for (int r = 0; foundmap && r < cands.length; r++)
590 if (cands[r].hasMap())
592 if (cands[r].getMap().getTo() != null
593 && cands[r].getMap().getMap().getFromRatio() != cands[r]
594 .getMap().getMap().getToRatio())
597 // get sense of map correct for adding to product
601 // map is from dna seq to a protein product
602 cf.addMap(sequenceI, nxt, cands[r].getMap()
607 // map should be from protein seq to its coding dna
608 cf.addMap(nxt, sequenceI, cands[r].getMap()
609 .getMap().getInverse());
614 // TODO: add mapping between sequences if necessary
627 * precalculate different products that can be found for seqs in dataset and
634 * - don't actually build lists - just get types
635 * @return public static Object[] buildXProductsList(boolean dna, SequenceI[]
636 * seqs, AlignmentI dataset, boolean fake) { String types[] =
637 * jalview.analysis.CrossRef.findSequenceXrefTypes( dna, seqs,
638 * dataset); if (types != null) { System.out.println("Xref Types for:
639 * "+(dna ? "dna" : "prot")); for (int t = 0; t < types.length; t++) {
640 * System.out.println("Type: " + types[t]); SequenceI[] prod =
641 * jalview.analysis.CrossRef.findXrefSequences(seqs, dna, types[t]);
642 * System.out.println("Found " + ((prod == null) ? "no" : "" +
643 * prod.length) + " products"); if (prod!=null) { for (int p=0;
644 * p<prod.length; p++) { System.out.println("Prod "+p+":
645 * "+prod[p].getDisplayId(true)); } } } } else {
646 * System.out.println("Trying getProducts for
647 * "+al.getSequenceAt(0).getDisplayId(true));
648 * System.out.println("Search DS Xref for: "+(dna ? "dna" : "prot"));
649 * // have a bash at finding the products amongst all the retrieved
650 * sequences. SequenceI[] prod =
651 * jalview.analysis.CrossRef.findXrefSequences(al
652 * .getSequencesArray(), dna, null, ds); System.out.println("Found " +
653 * ((prod == null) ? "no" : "" + prod.length) + " products"); if
654 * (prod!=null) { // select non-equivalent sequences from dataset list
655 * for (int p=0; p<prod.length; p++) { System.out.println("Prod "+p+":
656 * "+prod[p].getDisplayId(true)); } } } }