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.api.AlignViewportI;
24 import jalview.datamodel.AlignedCodon;
25 import jalview.datamodel.AlignedCodonFrame;
26 import jalview.datamodel.Alignment;
27 import jalview.datamodel.AlignmentAnnotation;
28 import jalview.datamodel.AlignmentI;
29 import jalview.datamodel.Annotation;
30 import jalview.datamodel.DBRefEntry;
31 import jalview.datamodel.DBRefSource;
32 import jalview.datamodel.FeatureProperties;
33 import jalview.datamodel.GraphLine;
34 import jalview.datamodel.Mapping;
35 import jalview.datamodel.Sequence;
36 import jalview.datamodel.SequenceFeature;
37 import jalview.datamodel.SequenceI;
38 import jalview.schemes.ResidueProperties;
39 import jalview.util.Comparison;
40 import jalview.util.DBRefUtils;
41 import jalview.util.MapList;
42 import jalview.util.ShiftList;
44 import java.util.ArrayList;
45 import java.util.Arrays;
46 import java.util.Comparator;
47 import java.util.List;
51 private static final String STOP_ASTERIX = "*";
53 private static final Comparator<AlignedCodon> comparator = new CodonComparator();
56 * 'final' variables describe the inputs to the translation, which should not
59 final private List<SequenceI> selection;
61 final private String[] seqstring;
63 final private int[] contigs;
65 final private char gapChar;
67 final private AlignmentAnnotation[] annotations;
69 final private int dnaWidth;
71 final private AlignmentI dataset;
74 * Working variables for the translation.
76 * The width of the translation-in-progress protein alignment.
78 private int aaWidth = 0;
81 * This array will be built up so that position i holds the codon positions
82 * e.g. [7, 9, 10] that match column i (base 0) in the aligned translation.
83 * Note this implies a contract that if two codons do not align exactly, their
84 * translated products must occupy different column positions.
86 private AlignedCodon[] alignedCodons;
89 * Constructor given a viewport and the visible contigs.
92 * @param visibleContigs
94 public Dna(AlignViewportI viewport, int[] visibleContigs)
96 this.selection = Arrays.asList(viewport.getSequenceSelection());
97 this.seqstring = viewport.getViewAsString(true);
98 this.contigs = visibleContigs;
99 this.gapChar = viewport.getGapCharacter();
100 this.annotations = viewport.getAlignment().getAlignmentAnnotation();
101 this.dnaWidth = viewport.getAlignment().getWidth();
102 this.dataset = viewport.getAlignment().getDataset();
106 * Test whether codon positions cdp1 should align before, with, or after cdp2.
107 * Returns zero if all positions match (or either argument is null). Returns
108 * -1 if any position in the first codon precedes the corresponding position
109 * in the second codon. Else returns +1 (some position in the second codon
110 * precedes the corresponding position in the first).
112 * Note this is not necessarily symmetric, for example:
114 * <li>compareCodonPos([2,5,6], [3,4,5]) returns -1</li>
115 * <li>compareCodonPos([3,4,5], [2,5,6]) also returns -1</li>
122 public static final int compareCodonPos(AlignedCodon ac1, AlignedCodon ac2)
124 return comparator.compare(ac1, ac2);
125 // return jalview_2_8_2compare(ac1, ac2);
129 * Codon comparison up to Jalview 2.8.2. This rule is sequence order dependent
130 * - see http://issues.jalview.org/browse/JAL-1635
136 private static int jalview_2_8_2compare(AlignedCodon ac1,
139 if (ac1 == null || ac2 == null || (ac1.equals(ac2)))
143 if (ac1.pos1 < ac2.pos1 || ac1.pos2 < ac2.pos2 || ac1.pos3 < ac2.pos3)
145 // one base in cdp1 precedes the corresponding base in the other codon
148 // one base in cdp1 appears after the corresponding base in the other codon.
156 public AlignmentI translateCdna()
158 AlignedCodonFrame acf = new AlignedCodonFrame();
160 alignedCodons = new AlignedCodon[dnaWidth];
163 int sSize = selection.size();
164 List<SequenceI> pepseqs = new ArrayList<SequenceI>();
165 for (s = 0; s < sSize; s++)
167 SequenceI newseq = translateCodingRegion(selection.get(s),
168 seqstring[s], acf, pepseqs);
173 SequenceI ds = newseq;
176 while (ds.getDatasetSequence() != null)
178 ds = ds.getDatasetSequence();
180 dataset.addSequence(ds);
185 SequenceI[] newseqs = pepseqs.toArray(new SequenceI[pepseqs.size()]);
186 AlignmentI al = new Alignment(newseqs);
187 // ensure we look aligned.
189 // link the protein translation to the DNA dataset
190 al.setDataset(dataset);
191 translateAlignedAnnotations(al, acf);
192 al.addCodonFrame(acf);
197 * fake the collection of DbRefs with associated exon mappings to identify if
198 * a translation would generate distinct product in the currently selected
205 public static boolean canTranslate(SequenceI[] selection,
208 for (int gd = 0; gd < selection.length; gd++)
210 SequenceI dna = selection[gd];
211 DBRefEntry[] dnarefs = DBRefUtils.selectRefs(dna.getDBRefs(),
212 jalview.datamodel.DBRefSource.DNACODINGDBS);
215 // intersect with pep
216 List<DBRefEntry> mappedrefs = new ArrayList<DBRefEntry>();
217 DBRefEntry[] refs = dna.getDBRefs();
218 for (int d = 0; d < refs.length; d++)
220 if (refs[d].getMap() != null && refs[d].getMap().getMap() != null
221 && refs[d].getMap().getMap().getFromRatio() == 3
222 && refs[d].getMap().getMap().getToRatio() == 1)
224 mappedrefs.add(refs[d]); // add translated protein maps
227 dnarefs = mappedrefs.toArray(new DBRefEntry[mappedrefs.size()]);
228 for (int d = 0; d < dnarefs.length; d++)
230 Mapping mp = dnarefs[d].getMap();
233 for (int vc = 0; vc < viscontigs.length; vc += 2)
235 int[] mpr = mp.locateMappedRange(viscontigs[vc],
250 * Translate nucleotide alignment annotations onto translated amino acid
251 * alignment using codon mapping codons
254 * the translated protein alignment
256 protected void translateAlignedAnnotations(AlignmentI al,
257 AlignedCodonFrame acf)
259 // Can only do this for columns with consecutive codons, or where
260 // annotation is sequence associated.
262 if (annotations != null)
264 for (AlignmentAnnotation annotation : annotations)
267 * Skip hidden or autogenerated annotation. Also (for now), RNA
268 * secondary structure annotation. If we want to show this against
269 * protein we need a smarter way to 'translate' without generating
270 * invalid (unbalanced) structure annotation.
272 if (annotation.autoCalculated || !annotation.visible
273 || annotation.isRNA())
279 Annotation[] anots = (annotation.annotations == null) ? null
280 : new Annotation[aSize];
283 for (int a = 0; a < aSize; a++)
285 // process through codon map.
286 if (a < alignedCodons.length && alignedCodons[a] != null
287 && alignedCodons[a].pos1 == (alignedCodons[a].pos3 - 2))
289 anots[a] = getCodonAnnotation(alignedCodons[a],
290 annotation.annotations);
295 AlignmentAnnotation aa = new AlignmentAnnotation(annotation.label,
296 annotation.description, anots);
297 aa.graph = annotation.graph;
298 aa.graphGroup = annotation.graphGroup;
299 aa.graphHeight = annotation.graphHeight;
300 if (annotation.getThreshold() != null)
302 aa.setThreshold(new GraphLine(annotation.getThreshold()));
304 if (annotation.hasScore)
306 aa.setScore(annotation.getScore());
309 final SequenceI seqRef = annotation.sequenceRef;
312 SequenceI aaSeq = acf.getAaForDnaSeq(seqRef);
315 // aa.compactAnnotationArray(); // throw away alignment annotation
317 aa.setSequenceRef(aaSeq);
319 aa.createSequenceMapping(aaSeq, aaSeq.getStart(), true);
320 aa.adjustForAlignment();
321 aaSeq.addAlignmentAnnotation(aa);
324 al.addAnnotation(aa);
329 private static Annotation getCodonAnnotation(AlignedCodon is,
330 Annotation[] annotations)
332 // Have a look at all the codon positions for annotation and put the first
333 // one found into the translated annotation pos.
335 Annotation annot = null;
336 for (int p = 1; p <= 3; p++)
338 int dnaCol = is.getBaseColumn(p);
339 if (annotations[dnaCol] != null)
343 annot = new Annotation(annotations[dnaCol]);
349 Annotation cpy = new Annotation(annotations[dnaCol]);
350 if (annot.colour == null)
352 annot.colour = cpy.colour;
354 if (annot.description == null || annot.description.length() == 0)
356 annot.description = cpy.description;
358 if (annot.displayCharacter == null)
360 annot.displayCharacter = cpy.displayCharacter;
362 if (annot.secondaryStructure == 0)
364 annot.secondaryStructure = cpy.secondaryStructure;
366 annot.value += cpy.value;
373 annot.value /= contrib;
379 * Translate a na sequence
382 * sequence displayed under viscontigs visible columns
384 * ORF read in some global alignment reference frame
386 * Definition of global ORF alignment reference frame
388 * @return sequence ready to be added to alignment.
390 protected SequenceI translateCodingRegion(SequenceI selection,
391 String seqstring, AlignedCodonFrame acf,
392 List<SequenceI> proteinSeqs)
394 List<int[]> skip = new ArrayList<int[]>();
395 int skipint[] = null;
396 ShiftList vismapping = new ShiftList(); // map from viscontigs to seqstring
399 int[] scontigs = new int[contigs.length];
401 for (vc = 0; vc < contigs.length; vc += 2)
405 vismapping.addShift(npos, contigs[vc]);
410 vismapping.addShift(npos, contigs[vc] - contigs[vc - 1] + 1);
412 scontigs[vc] = contigs[vc];
413 scontigs[vc + 1] = contigs[vc + 1];
416 // allocate a roughly sized buffer for the protein sequence
417 StringBuilder protein = new StringBuilder(seqstring.length() / 2);
418 String seq = seqstring.replace('U', 'T').replace('u', 'T');
419 char codon[] = new char[3];
420 int cdp[] = new int[3];
426 for (npos = 0, nend = seq.length(); npos < nend; npos++)
428 if (!Comparison.isGap(seq.charAt(npos)))
430 cdp[rf] = npos; // store position
431 codon[rf++] = seq.charAt(npos); // store base
436 * Filled up a reading frame...
438 AlignedCodon alignedCodon = new AlignedCodon(cdp[0], cdp[1],
440 String aa = ResidueProperties.codonTranslate(new String(codon));
442 final String gapString = String.valueOf(gapChar);
448 skipint = new int[] { alignedCodon.pos1,
454 skipint[1] = alignedCodon.pos3; // cdp[2];
461 skipint[0] = vismapping.shift(skipint[0]);
462 skipint[1] = vismapping.shift(skipint[1]);
463 for (vc = 0; vc < scontigs.length;)
465 if (scontigs[vc + 1] < skipint[0])
467 // before skipint starts
471 if (scontigs[vc] > skipint[1])
473 // finished editing so
476 // Edit the contig list to include the skipped region which did
479 // from : s1 e1 s2 e2 s3 e3
480 // to s: s1 e1 s2 k0 k1 e2 s3 e3
481 // list increases by one unless one boundary (s2==k0 or e2==k1)
482 // matches, and decreases by one if skipint intersects whole
484 if (scontigs[vc] <= skipint[0])
486 if (skipint[0] == scontigs[vc])
488 // skipint at start of contig
489 // shift the start of this contig
490 if (scontigs[vc + 1] > skipint[1])
492 scontigs[vc] = skipint[1];
497 if (scontigs[vc + 1] == skipint[1])
500 t = new int[scontigs.length - 2];
503 System.arraycopy(scontigs, 0, t, 0, vc - 1);
505 if (vc + 2 < t.length)
507 System.arraycopy(scontigs, vc + 2, t, vc,
514 // truncate contig to before the skipint region
515 scontigs[vc + 1] = skipint[0] - 1;
522 // scontig starts before start of skipint
523 if (scontigs[vc + 1] < skipint[1])
525 // skipint truncates end of scontig
526 scontigs[vc + 1] = skipint[0] - 1;
531 // divide region to new contigs
532 t = new int[scontigs.length + 2];
533 System.arraycopy(scontigs, 0, t, 0, vc + 1);
534 t[vc + 1] = skipint[0];
535 t[vc + 2] = skipint[1];
536 System.arraycopy(scontigs, vc + 1, t, vc + 3,
537 scontigs.length - (vc + 1));
547 if (aa.equals("STOP"))
553 boolean findpos = true;
557 * Compare this codon's base positions with those currently aligned to
558 * this column in the translation.
560 final int compareCodonPos = compareCodonPos(alignedCodon,
561 alignedCodons[aspos]);
562 switch (compareCodonPos)
567 * This codon should precede the mapped positions - need to insert a
568 * gap in all prior sequences.
570 insertAAGap(aspos, proteinSeqs);
577 * This codon belongs after the aligned codons at aspos. Prefix it
578 * with a gap and try the next position.
587 * Exact match - codon 'belongs' at this translated position.
594 if (alignedCodons[aspos] == null)
596 // mark this column as aligning to this aligned reading frame
597 alignedCodons[aspos] = alignedCodon;
599 else if (!alignedCodons[aspos].equals(alignedCodon))
601 throw new IllegalStateException(
602 "Tried to coalign " + alignedCodons[aspos].toString()
603 + " with " + alignedCodon.toString());
605 if (aspos >= aaWidth)
607 // update maximum alignment width
610 // ready for next translated reading frame alignment position (if any)
616 SequenceI newseq = new Sequence(selection.getName(),
620 final String errMsg = "trimming contigs for incomplete terminal codon.";
621 System.err.println(errMsg);
622 // map and trim contigs to ORF region
623 vc = scontigs.length - 1;
624 lastnpos = vismapping.shift(lastnpos); // place npos in context of
625 // whole dna alignment (rather
626 // than visible contigs)
627 // incomplete ORF could be broken over one or two visible contig
629 while (vc >= 0 && scontigs[vc] > lastnpos)
631 if (vc > 0 && scontigs[vc - 1] > lastnpos)
637 // correct last interval in list.
638 scontigs[vc] = lastnpos;
642 if (vc > 0 && (vc + 1) < scontigs.length)
644 // truncate map list to just vc elements
645 int t[] = new int[vc + 1];
646 System.arraycopy(scontigs, 0, t, 0, vc + 1);
654 if (scontigs != null)
657 // map scontigs to actual sequence positions on selection
658 for (vc = 0; vc < scontigs.length; vc += 2)
660 scontigs[vc] = selection.findPosition(scontigs[vc]); // not from 1!
661 scontigs[vc + 1] = selection.findPosition(scontigs[vc + 1]); // exclusive
662 if (scontigs[vc + 1] == selection.getEnd())
667 // trim trailing empty intervals.
668 if ((vc + 2) < scontigs.length)
670 int t[] = new int[vc + 2];
671 System.arraycopy(scontigs, 0, t, 0, vc + 2);
675 * delete intervals in scontigs which are not translated. 1. map skip
676 * into sequence position intervals 2. truncate existing ranges and add
677 * new ranges to exclude untranslated regions. if (skip.size()>0) {
678 * Vector narange = new Vector(); for (vc=0; vc<scontigs.length; vc++) {
679 * narange.addElement(new int[] {scontigs[vc]}); } int sint=0,iv[]; vc =
680 * 0; while (sint<skip.size()) { skipint = (int[]) skip.elementAt(sint);
681 * do { iv = (int[]) narange.elementAt(vc); if (iv[0]>=skipint[0] &&
682 * iv[0]<=skipint[1]) { if (iv[0]==skipint[0]) { // delete beginning of
683 * range } else { // truncate range and create new one if necessary iv =
684 * (int[]) narange.elementAt(vc+1); if (iv[0]<=skipint[1]) { // truncate
685 * range iv[0] = skipint[1]; } else { } } } else if (iv[0]<skipint[0]) {
686 * iv = (int[]) narange.elementAt(vc+1); } } while (iv[0]) } }
688 MapList map = new MapList(scontigs, new int[] { 1, resSize }, 3, 1);
690 transferCodedFeatures(selection, newseq, map);
693 * Construct a dataset sequence for our new peptide.
695 SequenceI rseq = newseq.deriveSequence();
698 * Store a mapping (between the dataset sequences for the two
701 // SIDE-EFFECT: acf stores the aligned sequence reseq; to remove!
702 acf.addMap(selection, rseq, map);
706 // register the mapping somehow
712 * Insert a gap into the aligned proteins and the codon mapping array.
718 protected void insertAAGap(int pos, List<SequenceI> proteinSeqs)
721 for (SequenceI seq : proteinSeqs)
723 seq.insertCharAt(pos, gapChar);
726 checkCodonFrameWidth();
732 * Shift from [pos] to the end one to the right, and null out [pos]
734 System.arraycopy(alignedCodons, pos, alignedCodons, pos + 1,
735 alignedCodons.length - pos - 1);
736 alignedCodons[pos] = null;
741 * Check the codons array can accommodate a single insertion, if not resize
744 protected void checkCodonFrameWidth()
746 if (alignedCodons[alignedCodons.length - 1] != null)
749 * arraycopy insertion would bump a filled slot off the end, so expand.
751 AlignedCodon[] c = new AlignedCodon[alignedCodons.length + 10];
752 System.arraycopy(alignedCodons, 0, c, 0, alignedCodons.length);
758 * Given a peptide newly translated from a dna sequence, copy over and set any
759 * features on the peptide from the DNA.
765 private static void transferCodedFeatures(SequenceI dna, SequenceI pep,
768 DBRefEntry[] dnarefs = DBRefUtils.selectRefs(dna.getDBRefs(),
769 DBRefSource.DNACODINGDBS);
772 // intersect with pep
773 for (int d = 0; d < dnarefs.length; d++)
775 Mapping mp = dnarefs[d].getMap();
781 for (SequenceFeature sf : dna.getFeatures().getAllFeatures())
783 if (FeatureProperties.isCodingFeature(null, sf.getType()))
785 // if (map.intersectsFrom(sf[f].begin, sf[f].end))
794 * Returns an alignment consisting of the reversed (and optionally
795 * complemented) sequences set in this object's constructor
800 public AlignmentI reverseCdna(boolean complement)
802 int sSize = selection.size();
803 List<SequenceI> reversed = new ArrayList<SequenceI>();
804 for (int s = 0; s < sSize; s++)
806 SequenceI newseq = reverseSequence(selection.get(s).getName(),
807 seqstring[s], complement);
811 reversed.add(newseq);
815 SequenceI[] newseqs = reversed.toArray(new SequenceI[reversed.size()]);
816 AlignmentI al = new Alignment(newseqs);
817 ((Alignment) al).createDatasetAlignment();
822 * Returns a reversed, and optionally complemented, sequence. The new
823 * sequence's name is the original name with "|rev" or "|revcomp" appended.
824 * aAcCgGtT and DNA ambiguity codes are complemented, any other characters are
831 public static SequenceI reverseSequence(String seqName, String sequence,
834 String newName = seqName + "|rev" + (complement ? "comp" : "");
835 char[] originalSequence = sequence.toCharArray();
836 int length = originalSequence.length;
837 char[] reversedSequence = new char[length];
839 for (int i = 0; i < length; i++)
841 char c = complement ? getComplement(originalSequence[i])
842 : originalSequence[i];
843 reversedSequence[length - i - 1] = c;
844 if (!Comparison.isGap(c))
849 SequenceI reversed = new Sequence(newName, reversedSequence, 1, bases);
854 * Returns dna complement (preserving case) for aAcCgGtTuU. Ambiguity codes
855 * are treated as on http://reverse-complement.com/. Anything else is left
861 public static char getComplement(char c)