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.AlignedCodon;
24 import jalview.datamodel.AlignedCodonFrame;
25 import jalview.datamodel.Alignment;
26 import jalview.datamodel.AlignmentAnnotation;
27 import jalview.datamodel.AlignmentI;
28 import jalview.datamodel.DBRefEntry;
29 import jalview.datamodel.DBRefSource;
30 import jalview.datamodel.FeatureProperties;
31 import jalview.datamodel.Mapping;
32 import jalview.datamodel.SearchResults;
33 import jalview.datamodel.Sequence;
34 import jalview.datamodel.SequenceGroup;
35 import jalview.datamodel.SequenceI;
36 import jalview.schemes.ResidueProperties;
37 import jalview.util.DBRefUtils;
38 import jalview.util.MapList;
39 import jalview.util.MappingUtils;
41 import java.util.ArrayList;
42 import java.util.Arrays;
43 import java.util.Collection;
44 import java.util.HashMap;
45 import java.util.HashSet;
46 import java.util.Iterator;
47 import java.util.LinkedHashMap;
48 import java.util.LinkedHashSet;
49 import java.util.List;
51 import java.util.Map.Entry;
53 import java.util.TreeMap;
56 * grab bag of useful alignment manipulation operations Expect these to be
57 * refactored elsewhere at some point.
62 public class AlignmentUtils
66 * given an existing alignment, create a new alignment including all, or up to
67 * flankSize additional symbols from each sequence's dataset sequence
73 public static AlignmentI expandContext(AlignmentI core, int flankSize)
75 List<SequenceI> sq = new ArrayList<SequenceI>();
77 for (SequenceI s : core.getSequences())
79 SequenceI newSeq = s.deriveSequence();
80 final int newSeqStart = newSeq.getStart() - 1;
81 if (newSeqStart > maxoffset
82 && newSeq.getDatasetSequence().getStart() < s.getStart())
84 maxoffset = newSeqStart;
90 maxoffset = Math.min(maxoffset, flankSize);
94 * now add offset left and right to create an expanded alignment
96 for (SequenceI s : sq)
99 while (ds.getDatasetSequence() != null)
101 ds = ds.getDatasetSequence();
103 int s_end = s.findPosition(s.getStart() + s.getLength());
104 // find available flanking residues for sequence
105 int ustream_ds = s.getStart() - ds.getStart();
106 int dstream_ds = ds.getEnd() - s_end;
108 // build new flanked sequence
110 // compute gap padding to start of flanking sequence
111 int offset = maxoffset - ustream_ds;
113 // padding is gapChar x ( maxoffset - min(ustream_ds, flank)
116 if (flankSize < ustream_ds)
118 // take up to flankSize residues
119 offset = maxoffset - flankSize;
120 ustream_ds = flankSize;
122 if (flankSize <= dstream_ds)
124 dstream_ds = flankSize - 1;
127 // TODO use Character.toLowerCase to avoid creating String objects?
128 char[] upstream = new String(ds.getSequence(s.getStart() - 1
129 - ustream_ds, s.getStart() - 1)).toLowerCase().toCharArray();
130 char[] downstream = new String(ds.getSequence(s_end - 1, s_end
131 + dstream_ds)).toLowerCase().toCharArray();
132 char[] coreseq = s.getSequence();
133 char[] nseq = new char[offset + upstream.length + downstream.length
135 char c = core.getGapCharacter();
138 for (; p < offset; p++)
143 System.arraycopy(upstream, 0, nseq, p, upstream.length);
144 System.arraycopy(coreseq, 0, nseq, p + upstream.length,
146 System.arraycopy(downstream, 0, nseq, p + coreseq.length
147 + upstream.length, downstream.length);
148 s.setSequence(new String(nseq));
149 s.setStart(s.getStart() - ustream_ds);
150 s.setEnd(s_end + downstream.length);
152 AlignmentI newAl = new jalview.datamodel.Alignment(
153 sq.toArray(new SequenceI[0]));
154 for (SequenceI s : sq)
156 if (s.getAnnotation() != null)
158 for (AlignmentAnnotation aa : s.getAnnotation())
160 aa.adjustForAlignment(); // JAL-1712 fix
161 newAl.addAnnotation(aa);
165 newAl.setDataset(core.getDataset());
170 * Returns the index (zero-based position) of a sequence in an alignment, or
177 public static int getSequenceIndex(AlignmentI al, SequenceI seq)
181 for (SequenceI alSeq : al.getSequences())
194 * Returns a map of lists of sequences in the alignment, keyed by sequence
195 * name. For use in mapping between different alignment views of the same
198 * @see jalview.datamodel.AlignmentI#getSequencesByName()
200 public static Map<String, List<SequenceI>> getSequencesByName(
203 Map<String, List<SequenceI>> theMap = new LinkedHashMap<String, List<SequenceI>>();
204 for (SequenceI seq : al.getSequences())
206 String name = seq.getName();
209 List<SequenceI> seqs = theMap.get(name);
212 seqs = new ArrayList<SequenceI>();
213 theMap.put(name, seqs);
222 * Build mapping of protein to cDNA alignment. Mappings are made between
223 * sequences where the cDNA translates to the protein sequence. Any new
224 * mappings are added to the protein alignment. Returns true if any mappings
225 * either already exist or were added, else false.
227 * @param proteinAlignment
228 * @param cdnaAlignment
231 public static boolean mapProteinToCdna(
232 final AlignmentI proteinAlignment,
233 final AlignmentI cdnaAlignment)
235 if (proteinAlignment == null || cdnaAlignment == null)
240 Set<SequenceI> mappedDna = new HashSet<SequenceI>();
241 Set<SequenceI> mappedProtein = new HashSet<SequenceI>();
244 * First pass - map sequences where cross-references exist. This include
245 * 1-to-many mappings to support, for example, variant cDNA.
247 boolean mappingPerformed = mapProteinToCdna(proteinAlignment,
248 cdnaAlignment, mappedDna, mappedProtein, true);
251 * Second pass - map sequences where no cross-references exist. This only
252 * does 1-to-1 mappings and assumes corresponding sequences are in the same
253 * order in the alignments.
255 mappingPerformed |= mapProteinToCdna(proteinAlignment, cdnaAlignment,
256 mappedDna, mappedProtein, false);
257 return mappingPerformed;
261 * Make mappings between compatible sequences (where the cDNA translation
262 * matches the protein).
264 * @param proteinAlignment
265 * @param cdnaAlignment
267 * a set of mapped DNA sequences (to add to)
268 * @param mappedProtein
269 * a set of mapped Protein sequences (to add to)
271 * if true, only map sequences where xrefs exist
274 protected static boolean mapProteinToCdna(
275 final AlignmentI proteinAlignment,
276 final AlignmentI cdnaAlignment, Set<SequenceI> mappedDna,
277 Set<SequenceI> mappedProtein, boolean xrefsOnly)
279 boolean mappingPerformed = false;
280 List<SequenceI> thisSeqs = proteinAlignment.getSequences();
281 for (SequenceI aaSeq : thisSeqs)
283 boolean proteinMapped = false;
284 AlignedCodonFrame acf = new AlignedCodonFrame();
286 for (SequenceI cdnaSeq : cdnaAlignment.getSequences())
289 * Always try to map if sequences have xref to each other; this supports
290 * variant cDNA or alternative splicing for a protein sequence.
292 * If no xrefs, try to map progressively, assuming that alignments have
293 * mappable sequences in corresponding order. These are not
294 * many-to-many, as that would risk mixing species with similar cDNA
297 if (xrefsOnly && !AlignmentUtils.haveCrossRef(aaSeq, cdnaSeq))
303 * Don't map non-xrefd sequences more than once each. This heuristic
304 * allows us to pair up similar sequences in ordered alignments.
307 && (mappedProtein.contains(aaSeq) || mappedDna
312 if (!mappingExists(proteinAlignment.getCodonFrames(),
313 aaSeq.getDatasetSequence(), cdnaSeq.getDatasetSequence()))
315 MapList map = mapProteinToCdna(aaSeq, cdnaSeq);
318 acf.addMap(cdnaSeq, aaSeq, map);
319 mappingPerformed = true;
320 proteinMapped = true;
321 mappedDna.add(cdnaSeq);
322 mappedProtein.add(aaSeq);
328 proteinAlignment.addCodonFrame(acf);
331 return mappingPerformed;
335 * Answers true if the mappings include one between the given (dataset)
338 public static boolean mappingExists(Set<AlignedCodonFrame> set,
339 SequenceI aaSeq, SequenceI cdnaSeq)
343 for (AlignedCodonFrame acf : set)
345 if (cdnaSeq == acf.getDnaForAaSeq(aaSeq))
355 * Build a mapping (if possible) of a protein to a cDNA sequence. The cDNA
356 * must be three times the length of the protein, possibly after ignoring
357 * start and/or stop codons, and must translate to the protein. Returns null
358 * if no mapping is determined.
364 public static MapList mapProteinToCdna(SequenceI proteinSeq,
368 * Here we handle either dataset sequence set (desktop) or absent (applet).
369 * Use only the char[] form of the sequence to avoid creating possibly large
372 final SequenceI proteinDataset = proteinSeq.getDatasetSequence();
373 char[] aaSeqChars = proteinDataset != null ? proteinDataset
374 .getSequence() : proteinSeq.getSequence();
375 final SequenceI cdnaDataset = cdnaSeq.getDatasetSequence();
376 char[] cdnaSeqChars = cdnaDataset != null ? cdnaDataset.getSequence()
377 : cdnaSeq.getSequence();
378 if (aaSeqChars == null || cdnaSeqChars == null)
384 * cdnaStart/End, proteinStartEnd are base 1 (for dataset sequence mapping)
386 final int mappedLength = 3 * aaSeqChars.length;
387 int cdnaLength = cdnaSeqChars.length;
389 int cdnaEnd = cdnaLength;
390 final int proteinStart = 1;
391 final int proteinEnd = aaSeqChars.length;
394 * If lengths don't match, try ignoring stop codon.
396 if (cdnaLength != mappedLength && cdnaLength > 2)
398 String lastCodon = String.valueOf(cdnaSeqChars, cdnaLength - 3, 3)
400 for (String stop : ResidueProperties.STOP)
402 if (lastCodon.equals(stop))
412 * If lengths still don't match, try ignoring start codon.
414 if (cdnaLength != mappedLength
416 && String.valueOf(cdnaSeqChars, 0, 3).toUpperCase()
418 ResidueProperties.START))
424 if (cdnaLength != mappedLength)
428 if (!translatesAs(cdnaSeqChars, cdnaStart - 1, aaSeqChars))
432 MapList map = new MapList(new int[]
433 { cdnaStart, cdnaEnd }, new int[]
434 { proteinStart, proteinEnd }, 3, 1);
439 * Test whether the given cdna sequence, starting at the given offset,
440 * translates to the given amino acid sequence, using the standard translation
441 * table. Designed to fail fast i.e. as soon as a mismatch position is found.
443 * @param cdnaSeqChars
448 protected static boolean translatesAs(char[] cdnaSeqChars, int cdnaStart,
451 if (cdnaSeqChars == null || aaSeqChars == null)
457 for (int i = cdnaStart; i < cdnaSeqChars.length - 2
458 && aaResidue < aaSeqChars.length; i += 3, aaResidue++)
460 String codon = String.valueOf(cdnaSeqChars, i, 3);
461 final String translated = ResidueProperties.codonTranslate(
464 * ? allow X in protein to match untranslatable in dna ?
466 final char aaRes = aaSeqChars[aaResidue];
467 if ((translated == null || "STOP".equals(translated)) && aaRes == 'X')
471 if (translated == null
472 || !(aaRes == translated.charAt(0)))
475 // System.out.println(("Mismatch at " + i + "/" + aaResidue + ": "
476 // + codon + "(" + translated + ") != " + aaRes));
480 // fail if we didn't match all of the aa sequence
481 return (aaResidue == aaSeqChars.length);
485 * Align sequence 'seq' to match the alignment of a mapped sequence. Note this
486 * currently assumes that we are aligning cDNA to match protein.
489 * the sequence to be realigned
491 * the alignment whose sequence alignment is to be 'copied'
493 * character string represent a gap in the realigned sequence
494 * @param preserveUnmappedGaps
495 * @param preserveMappedGaps
496 * @return true if the sequence was realigned, false if it could not be
498 public static boolean alignSequenceAs(SequenceI seq, AlignmentI al,
499 String gap, boolean preserveMappedGaps,
500 boolean preserveUnmappedGaps)
503 * Get any mappings from the source alignment to the target (dataset) sequence.
505 // TODO there may be one AlignedCodonFrame per dataset sequence, or one with
506 // all mappings. Would it help to constrain this?
507 List<AlignedCodonFrame> mappings = al.getCodonFrame(seq);
508 if (mappings == null || mappings.isEmpty())
514 * Locate the aligned source sequence whose dataset sequence is mapped. We
515 * just take the first match here (as we can't align cDNA like more than one
518 SequenceI alignFrom = null;
519 AlignedCodonFrame mapping = null;
520 for (AlignedCodonFrame mp : mappings)
522 alignFrom = mp.findAlignedSequence(seq.getDatasetSequence(), al);
523 if (alignFrom != null)
530 if (alignFrom == null)
534 alignSequenceAs(seq, alignFrom, mapping, gap, al.getGapCharacter(),
535 preserveMappedGaps, preserveUnmappedGaps);
540 * Align sequence 'alignTo' the same way as 'alignFrom', using the mapping to
541 * match residues and codons. Flags control whether existing gaps in unmapped
542 * (intron) and mapped (exon) regions are preserved or not. Gaps linking intro
543 * and exon are only retained if both flags are set.
550 * @param preserveUnmappedGaps
551 * @param preserveMappedGaps
553 public static void alignSequenceAs(SequenceI alignTo,
555 AlignedCodonFrame mapping, String myGap, char sourceGap,
556 boolean preserveMappedGaps, boolean preserveUnmappedGaps)
558 // TODO generalise to work for Protein-Protein, dna-dna, dna-protein
559 final char[] thisSeq = alignTo.getSequence();
560 final char[] thatAligned = alignFrom.getSequence();
561 StringBuilder thisAligned = new StringBuilder(2 * thisSeq.length);
563 // aligned and dataset sequence positions, all base zero
567 int basesWritten = 0;
568 char myGapChar = myGap.charAt(0);
569 int ratio = myGap.length();
572 * Traverse the aligned protein sequence.
574 int sourceGapMappedLength = 0;
575 boolean inExon = false;
576 for (char sourceChar : thatAligned)
578 if (sourceChar == sourceGap)
580 sourceGapMappedLength += ratio;
585 * Found a residue. Locate its mapped codon (start) position.
588 // Note mapping positions are base 1, our sequence positions base 0
589 int[] mappedPos = mapping.getMappedRegion(alignTo, alignFrom,
591 if (mappedPos == null)
594 * Abort realignment if unmapped protein. Or could ignore it??
596 System.err.println("Can't align: no codon mapping to residue "
597 + sourceDsPos + "(" + sourceChar + ")");
601 int mappedCodonStart = mappedPos[0]; // position (1...) of codon start
602 int mappedCodonEnd = mappedPos[mappedPos.length - 1]; // codon end pos
603 StringBuilder trailingCopiedGap = new StringBuilder();
606 * Copy dna sequence up to and including this codon. Optionally, include
607 * gaps before the codon starts (in introns) and/or after the codon starts
610 * Note this only works for 'linear' splicing, not reverse or interleaved.
611 * But then 'align dna as protein' doesn't make much sense otherwise.
613 int intronLength = 0;
614 while (basesWritten < mappedCodonEnd && thisSeqPos < thisSeq.length)
616 final char c = thisSeq[thisSeqPos++];
621 if (basesWritten < mappedCodonStart)
624 * Found an unmapped (intron) base. First add in any preceding gaps
627 if (preserveUnmappedGaps && trailingCopiedGap.length() > 0)
629 thisAligned.append(trailingCopiedGap.toString());
630 intronLength += trailingCopiedGap.length();
631 trailingCopiedGap = new StringBuilder();
638 final boolean startOfCodon = basesWritten == mappedCodonStart;
639 int gapsToAdd = calculateGapsToInsert(preserveMappedGaps,
640 preserveUnmappedGaps, sourceGapMappedLength, inExon,
641 trailingCopiedGap.length(), intronLength, startOfCodon);
642 for (int i = 0; i < gapsToAdd; i++)
644 thisAligned.append(myGapChar);
646 sourceGapMappedLength = 0;
649 thisAligned.append(c);
650 trailingCopiedGap = new StringBuilder();
654 if (inExon && preserveMappedGaps)
656 trailingCopiedGap.append(myGapChar);
658 else if (!inExon && preserveUnmappedGaps)
660 trailingCopiedGap.append(myGapChar);
667 * At end of protein sequence. Copy any remaining dna sequence, optionally
668 * including (intron) gaps. We do not copy trailing gaps in protein.
670 while (thisSeqPos < thisSeq.length)
672 final char c = thisSeq[thisSeqPos++];
673 if (c != myGapChar || preserveUnmappedGaps)
675 thisAligned.append(c);
680 * All done aligning, set the aligned sequence.
682 alignTo.setSequence(new String(thisAligned));
686 * Helper method to work out how many gaps to insert when realigning.
688 * @param preserveMappedGaps
689 * @param preserveUnmappedGaps
690 * @param sourceGapMappedLength
692 * @param trailingCopiedGap
693 * @param intronLength
694 * @param startOfCodon
697 protected static int calculateGapsToInsert(boolean preserveMappedGaps,
698 boolean preserveUnmappedGaps, int sourceGapMappedLength,
699 boolean inExon, int trailingGapLength,
700 int intronLength, final boolean startOfCodon)
706 * Reached start of codon. Ignore trailing gaps in intron unless we are
707 * preserving gaps in both exon and intron. Ignore them anyway if the
708 * protein alignment introduces a gap at least as large as the intronic
711 if (inExon && !preserveMappedGaps)
713 trailingGapLength = 0;
715 if (!inExon && !(preserveMappedGaps && preserveUnmappedGaps))
717 trailingGapLength = 0;
721 gapsToAdd = Math.max(sourceGapMappedLength, trailingGapLength);
725 if (intronLength + trailingGapLength <= sourceGapMappedLength)
727 gapsToAdd = sourceGapMappedLength - intronLength;
731 gapsToAdd = Math.min(intronLength + trailingGapLength
732 - sourceGapMappedLength, trailingGapLength);
739 * second or third base of codon; check for any gaps in dna
741 if (!preserveMappedGaps)
743 trailingGapLength = 0;
745 gapsToAdd = Math.max(sourceGapMappedLength, trailingGapLength);
751 * Returns a list of sequences mapped from the given sequences and aligned
752 * (gapped) in the same way. For example, the cDNA for aligned protein, where
753 * a single gap in protein generates three gaps in cDNA.
756 * @param gapCharacter
760 public static List<SequenceI> getAlignedTranslation(
761 List<SequenceI> sequences, char gapCharacter,
762 Set<AlignedCodonFrame> mappings)
764 List<SequenceI> alignedSeqs = new ArrayList<SequenceI>();
766 for (SequenceI seq : sequences)
768 List<SequenceI> mapped = getAlignedTranslation(seq, gapCharacter,
770 alignedSeqs.addAll(mapped);
776 * Returns sequences aligned 'like' the source sequence, as mapped by the
777 * given mappings. Normally we expect zero or one 'mapped' sequences, but this
778 * will support 1-to-many as well.
781 * @param gapCharacter
785 protected static List<SequenceI> getAlignedTranslation(SequenceI seq,
786 char gapCharacter, Set<AlignedCodonFrame> mappings)
788 List<SequenceI> result = new ArrayList<SequenceI>();
789 for (AlignedCodonFrame mapping : mappings)
791 if (mapping.involvesSequence(seq))
793 SequenceI mapped = getAlignedTranslation(seq, gapCharacter, mapping);
804 * Returns the translation of 'seq' (as held in the mapping) with
805 * corresponding alignment (gaps).
808 * @param gapCharacter
812 protected static SequenceI getAlignedTranslation(SequenceI seq,
813 char gapCharacter, AlignedCodonFrame mapping)
815 String gap = String.valueOf(gapCharacter);
816 boolean toDna = false;
818 SequenceI mapTo = mapping.getDnaForAaSeq(seq);
821 // mapping is from protein to nucleotide
823 // should ideally get gap count ratio from mapping
824 gap = String.valueOf(new char[]
825 { gapCharacter, gapCharacter, gapCharacter });
829 // mapping is from nucleotide to protein
830 mapTo = mapping.getAaForDnaSeq(seq);
833 StringBuilder newseq = new StringBuilder(seq.getLength()
836 int residueNo = 0; // in seq, base 1
837 int[] phrase = new int[fromRatio];
838 int phraseOffset = 0;
840 boolean first = true;
841 final Sequence alignedSeq = new Sequence("", "");
843 for (char c : seq.getSequence())
845 if (c == gapCharacter)
848 if (gapWidth >= fromRatio)
856 phrase[phraseOffset++] = residueNo + 1;
857 if (phraseOffset == fromRatio)
860 * Have read a whole codon (or protein residue), now translate: map
861 * source phrase to positions in target sequence add characters at
862 * these positions to newseq Note mapping positions are base 1, our
863 * sequence positions base 0.
865 SearchResults sr = new SearchResults();
866 for (int pos : phrase)
868 mapping.markMappedRegion(seq, pos, sr);
870 newseq.append(sr.getCharacters());
874 // Hack: Copy sequence dataset, name and description from
875 // SearchResults.match[0].sequence
876 // TODO? carry over sequence names from original 'complement'
878 SequenceI mappedTo = sr.getResultSequence(0);
879 alignedSeq.setName(mappedTo.getName());
880 alignedSeq.setDescription(mappedTo.getDescription());
881 alignedSeq.setDatasetSequence(mappedTo);
888 alignedSeq.setSequence(newseq.toString());
893 * Realigns the given protein to match the alignment of the dna, using codon
894 * mappings to translate aligned codon positions to protein residues.
897 * the alignment whose sequences are realigned by this method
899 * the dna alignment whose alignment we are 'copying'
900 * @return the number of sequences that were realigned
902 public static int alignProteinAsDna(AlignmentI protein, AlignmentI dna)
904 Set<AlignedCodonFrame> mappings = protein.getCodonFrames();
907 * Map will hold, for each aligned codon position e.g. [3, 5, 6], a map of
908 * {dnaSequence, {proteinSequence, codonProduct}} at that position. The
909 * comparator keeps the codon positions ordered.
911 Map<AlignedCodon, Map<SequenceI, String>> alignedCodons = new TreeMap<AlignedCodon, Map<SequenceI, String>>(
912 new CodonComparator());
913 for (SequenceI dnaSeq : dna.getSequences())
915 for (AlignedCodonFrame mapping : mappings)
917 Mapping seqMap = mapping.getMappingForSequence(dnaSeq);
918 SequenceI prot = mapping.findAlignedSequence(
919 dnaSeq.getDatasetSequence(), protein);
922 addCodonPositions(dnaSeq, prot, protein.getGapCharacter(),
923 seqMap, alignedCodons);
927 return alignProteinAs(protein, alignedCodons);
931 * Update the aligned protein sequences to match the codon alignments given in
935 * @param alignedCodons
936 * an ordered map of codon positions (columns), with sequence/peptide
937 * values present in each column
940 protected static int alignProteinAs(AlignmentI protein,
941 Map<AlignedCodon, Map<SequenceI, String>> alignedCodons)
944 * Prefill aligned sequences with gaps before inserting aligned protein
947 int alignedWidth = alignedCodons.size();
948 char[] gaps = new char[alignedWidth];
949 Arrays.fill(gaps, protein.getGapCharacter());
950 String allGaps = String.valueOf(gaps);
951 for (SequenceI seq : protein.getSequences())
953 seq.setSequence(allGaps);
957 for (AlignedCodon codon : alignedCodons.keySet())
959 final Map<SequenceI, String> columnResidues = alignedCodons.get(codon);
960 for (Entry<SequenceI, String> entry : columnResidues
963 // place translated codon at its column position in sequence
964 entry.getKey().getSequence()[column] = entry.getValue().charAt(0);
972 * Populate the map of aligned codons by traversing the given sequence
973 * mapping, locating the aligned positions of mapped codons, and adding those
974 * positions and their translation products to the map.
977 * the aligned sequence we are mapping from
979 * the sequence to be aligned to the codons
981 * the gap character in the dna sequence
983 * a mapping to a sequence translation
984 * @param alignedCodons
985 * the map we are building up
987 static void addCodonPositions(SequenceI dna, SequenceI protein,
990 Map<AlignedCodon, Map<SequenceI, String>> alignedCodons)
992 Iterator<AlignedCodon> codons = seqMap.getCodonIterator(dna, gapChar);
993 while (codons.hasNext())
995 AlignedCodon codon = codons.next();
996 Map<SequenceI, String> seqProduct = alignedCodons.get(codon);
997 if (seqProduct == null)
999 seqProduct = new HashMap<SequenceI, String>();
1000 alignedCodons.put(codon, seqProduct);
1002 seqProduct.put(protein, codon.product);
1007 * Returns true if a cDNA/Protein mapping either exists, or could be made,
1008 * between at least one pair of sequences in the two alignments. Currently,
1011 * <li>One alignment must be nucleotide, and the other protein</li>
1012 * <li>At least one pair of sequences must be already mapped, or mappable</li>
1013 * <li>Mappable means the nucleotide translation matches the protein sequence</li>
1014 * <li>The translation may ignore start and stop codons if present in the
1022 public static boolean isMappable(AlignmentI al1, AlignmentI al2)
1024 if (al1 == null || al2 == null)
1030 * Require one nucleotide and one protein
1032 if (al1.isNucleotide() == al2.isNucleotide())
1036 AlignmentI dna = al1.isNucleotide() ? al1 : al2;
1037 AlignmentI protein = dna == al1 ? al2 : al1;
1038 Set<AlignedCodonFrame> mappings = protein.getCodonFrames();
1039 for (SequenceI dnaSeq : dna.getSequences())
1041 for (SequenceI proteinSeq : protein.getSequences())
1043 if (isMappable(dnaSeq, proteinSeq, mappings))
1053 * Returns true if the dna sequence is mapped, or could be mapped, to the
1061 protected static boolean isMappable(SequenceI dnaSeq,
1062 SequenceI proteinSeq,
1063 Set<AlignedCodonFrame> mappings)
1065 if (dnaSeq == null || proteinSeq == null)
1070 SequenceI dnaDs = dnaSeq.getDatasetSequence() == null ? dnaSeq : dnaSeq.getDatasetSequence();
1071 SequenceI proteinDs = proteinSeq.getDatasetSequence() == null ? proteinSeq
1072 : proteinSeq.getDatasetSequence();
1077 for (AlignedCodonFrame mapping : mappings) {
1078 if ( proteinDs == mapping.getAaForDnaSeq(dnaDs)) {
1084 * Just try to make a mapping (it is not yet stored), test whether
1087 return mapProteinToCdna(proteinDs, dnaDs) != null;
1091 * Finds any reference annotations associated with the sequences in
1092 * sequenceScope, that are not already added to the alignment, and adds them
1093 * to the 'candidates' map. Also populates a lookup table of annotation
1094 * labels, keyed by calcId, for use in constructing tooltips or the like.
1096 * @param sequenceScope
1097 * the sequences to scan for reference annotations
1098 * @param labelForCalcId
1099 * (optional) map to populate with label for calcId
1101 * map to populate with annotations for sequence
1103 * the alignment to check for presence of annotations
1105 public static void findAddableReferenceAnnotations(
1106 List<SequenceI> sequenceScope, Map<String, String> labelForCalcId,
1107 final Map<SequenceI, List<AlignmentAnnotation>> candidates,
1110 if (sequenceScope == null)
1116 * For each sequence in scope, make a list of any annotations on the
1117 * underlying dataset sequence which are not already on the alignment.
1119 * Add to a map of { alignmentSequence, <List of annotations to add> }
1121 for (SequenceI seq : sequenceScope)
1123 SequenceI dataset = seq.getDatasetSequence();
1124 if (dataset == null)
1128 AlignmentAnnotation[] datasetAnnotations = dataset.getAnnotation();
1129 if (datasetAnnotations == null)
1133 final List<AlignmentAnnotation> result = new ArrayList<AlignmentAnnotation>();
1134 for (AlignmentAnnotation dsann : datasetAnnotations)
1137 * Find matching annotations on the alignment. If none is found, then
1138 * add this annotation to the list of 'addable' annotations for this
1141 final Iterable<AlignmentAnnotation> matchedAlignmentAnnotations = al
1142 .findAnnotations(seq, dsann.getCalcId(),
1144 if (!matchedAlignmentAnnotations.iterator().hasNext())
1147 if (labelForCalcId != null)
1149 labelForCalcId.put(dsann.getCalcId(), dsann.label);
1154 * Save any addable annotations for this sequence
1156 if (!result.isEmpty())
1158 candidates.put(seq, result);
1164 * Adds annotations to the top of the alignment annotations, in the same order
1165 * as their related sequences.
1167 * @param annotations
1168 * the annotations to add
1170 * the alignment to add them to
1171 * @param selectionGroup
1172 * current selection group (or null if none)
1174 public static void addReferenceAnnotations(
1175 Map<SequenceI, List<AlignmentAnnotation>> annotations,
1176 final AlignmentI alignment, final SequenceGroup selectionGroup)
1178 for (SequenceI seq : annotations.keySet())
1180 for (AlignmentAnnotation ann : annotations.get(seq))
1182 AlignmentAnnotation copyAnn = new AlignmentAnnotation(ann);
1184 int endRes = ann.annotations.length;
1185 if (selectionGroup != null)
1187 startRes = selectionGroup.getStartRes();
1188 endRes = selectionGroup.getEndRes();
1190 copyAnn.restrict(startRes, endRes);
1193 * Add to the sequence (sets copyAnn.datasetSequence), unless the
1194 * original annotation is already on the sequence.
1196 if (!seq.hasAnnotation(ann))
1198 seq.addAlignmentAnnotation(copyAnn);
1201 copyAnn.adjustForAlignment();
1202 // add to the alignment and set visible
1203 alignment.addAnnotation(copyAnn);
1204 copyAnn.visible = true;
1210 * Set visibility of alignment annotations of specified types (labels), for
1211 * specified sequences. This supports controls like
1212 * "Show all secondary structure", "Hide all Temp factor", etc.
1214 * @al the alignment to scan for annotations
1216 * the types (labels) of annotations to be updated
1217 * @param forSequences
1218 * if not null, only annotations linked to one of these sequences are
1219 * in scope for update; if null, acts on all sequence annotations
1221 * if this flag is true, 'types' is ignored (label not checked)
1223 * if true, set visibility on, else set off
1225 public static void showOrHideSequenceAnnotations(AlignmentI al,
1226 Collection<String> types, List<SequenceI> forSequences,
1227 boolean anyType, boolean doShow)
1229 for (AlignmentAnnotation aa : al
1230 .getAlignmentAnnotation())
1232 if (anyType || types.contains(aa.label))
1234 if ((aa.sequenceRef != null)
1235 && (forSequences == null || forSequences
1236 .contains(aa.sequenceRef)))
1238 aa.visible = doShow;
1245 * Returns true if either sequence has a cross-reference to the other
1251 public static boolean haveCrossRef(SequenceI seq1, SequenceI seq2)
1253 // Note: moved here from class CrossRef as the latter class has dependencies
1254 // not availability to the applet's classpath
1255 return hasCrossRef(seq1, seq2) || hasCrossRef(seq2, seq1);
1259 * Returns true if seq1 has a cross-reference to seq2. Currently this assumes
1260 * that sequence name is structured as Source|AccessionId.
1266 public static boolean hasCrossRef(SequenceI seq1, SequenceI seq2)
1268 if (seq1 == null || seq2 == null)
1272 String name = seq2.getName();
1273 final DBRefEntry[] xrefs = seq1.getDBRef();
1276 for (DBRefEntry xref : xrefs)
1278 String xrefName = xref.getSource() + "|" + xref.getAccessionId();
1279 // case-insensitive test, consistent with DBRefEntry.equalRef()
1280 if (xrefName.equalsIgnoreCase(name))
1290 * Constructs an alignment consisting of the mapped exon regions in the given
1291 * nucleotide sequences, and updates mappings to match.
1294 * aligned dna sequences
1296 * from dna to protein; these are replaced with new mappings
1297 * @return an alignment whose sequences are the exon-only parts of the dna
1298 * sequences (or null if no exons are found)
1300 public static AlignmentI makeExonAlignment(SequenceI[] dna,
1301 Set<AlignedCodonFrame> mappings)
1303 Set<AlignedCodonFrame> newMappings = new LinkedHashSet<AlignedCodonFrame>();
1304 List<SequenceI> exonSequences = new ArrayList<SequenceI>();
1306 for (SequenceI dnaSeq : dna)
1308 final SequenceI ds = dnaSeq.getDatasetSequence();
1309 List<AlignedCodonFrame> seqMappings = MappingUtils
1310 .findMappingsForSequence(ds, mappings);
1311 for (AlignedCodonFrame acf : seqMappings)
1313 AlignedCodonFrame newMapping = new AlignedCodonFrame();
1314 final List<SequenceI> mappedExons = makeExonSequences(ds, acf,
1316 if (!mappedExons.isEmpty())
1318 exonSequences.addAll(mappedExons);
1319 newMappings.add(newMapping);
1323 AlignmentI al = new Alignment(
1324 exonSequences.toArray(new SequenceI[exonSequences.size()]));
1325 al.setDataset(null);
1328 * Replace the old mappings with the new ones
1331 mappings.addAll(newMappings);
1337 * Helper method to make exon-only sequences and populate their mappings to
1340 * For example, if ggCCaTTcGAg has mappings [3, 4, 6, 7, 9, 10] to protein
1341 * then generate a sequence CCTTGA with mapping [1, 6] to the same protein
1344 * Typically eukaryotic dna will include exons encoding for a single peptide
1345 * sequence i.e. return a single result. Bacterial dna may have overlapping
1346 * exon mappings coding for multiple peptides so return multiple results
1347 * (example EMBL KF591215).
1350 * a dna dataset sequence
1352 * containing one or more mappings of the sequence to protein
1354 * the new mapping to populate, from the exon-only sequences to their
1355 * mapped protein sequences
1358 protected static List<SequenceI> makeExonSequences(SequenceI dnaSeq,
1359 AlignedCodonFrame mapping, AlignedCodonFrame newMapping)
1361 List<SequenceI> exonSequences = new ArrayList<SequenceI>();
1362 List<Mapping> seqMappings = mapping.getMappingsForSequence(dnaSeq);
1363 final char[] dna = dnaSeq.getSequence();
1364 for (Mapping seqMapping : seqMappings)
1366 StringBuilder newSequence = new StringBuilder(dnaSeq.getLength());
1369 * Get the codon regions as { [2, 5], [7, 12], [14, 14] etc }
1371 final List<int[]> dnaExonRanges = seqMapping.getMap().getFromRanges();
1372 for (int[] range : dnaExonRanges)
1374 for (int pos = range[0]; pos <= range[1]; pos++)
1376 newSequence.append(dna[pos - 1]);
1380 SequenceI exon = new Sequence(dnaSeq.getName(),
1381 newSequence.toString());
1384 * Locate any xrefs to CDS database on the protein product and attach to
1385 * the CDS sequence. Also add as a sub-token of the sequence name.
1387 // default to "CDS" if we can't locate an actual gene id
1388 String cdsAccId = FeatureProperties
1389 .getCodingFeature(DBRefSource.EMBL);
1390 DBRefEntry[] cdsRefs = DBRefUtils.selectRefs(seqMapping.getTo()
1391 .getDBRef(), DBRefSource.CODINGDBS);
1392 if (cdsRefs != null)
1394 for (DBRefEntry cdsRef : cdsRefs)
1396 exon.addDBRef(new DBRefEntry(cdsRef));
1397 cdsAccId = cdsRef.getAccessionId();
1400 exon.setName(exon.getName() + "|" + cdsAccId);
1401 exon.createDatasetSequence();
1404 * Build new mappings - from the same protein regions, but now to
1407 List<int[]> exonRange = new ArrayList<int[]>();
1408 exonRange.add(new int[]
1409 { 1, newSequence.length() });
1410 MapList map = new MapList(exonRange, seqMapping.getMap()
1413 newMapping.addMap(exon.getDatasetSequence(), seqMapping.getTo(), map);
1414 MapList cdsToDnaMap = new MapList(dnaExonRanges, exonRange, 1, 1);
1415 newMapping.addMap(dnaSeq, exon.getDatasetSequence(), cdsToDnaMap);
1417 exonSequences.add(exon);
1419 return exonSequences;