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 static jalview.io.gff.GffConstants.CLINICAL_SIGNIFICANCE;
25 import jalview.datamodel.AlignedCodon;
26 import jalview.datamodel.AlignedCodonFrame;
27 import jalview.datamodel.AlignedCodonFrame.SequenceToSequenceMapping;
28 import jalview.datamodel.Alignment;
29 import jalview.datamodel.AlignmentAnnotation;
30 import jalview.datamodel.AlignmentI;
31 import jalview.datamodel.DBRefEntry;
32 import jalview.datamodel.IncompleteCodonException;
33 import jalview.datamodel.Mapping;
34 import jalview.datamodel.Sequence;
35 import jalview.datamodel.SequenceFeature;
36 import jalview.datamodel.SequenceGroup;
37 import jalview.datamodel.SequenceI;
38 import jalview.datamodel.features.SequenceFeatures;
39 import jalview.io.gff.SequenceOntologyI;
40 import jalview.schemes.ResidueProperties;
41 import jalview.util.Comparison;
42 import jalview.util.DBRefUtils;
43 import jalview.util.IntRangeComparator;
44 import jalview.util.MapList;
45 import jalview.util.MappingUtils;
46 import jalview.util.StringUtils;
48 import java.io.UnsupportedEncodingException;
49 import java.net.URLEncoder;
50 import java.util.ArrayList;
51 import java.util.Arrays;
52 import java.util.Collection;
53 import java.util.Collections;
54 import java.util.HashMap;
55 import java.util.HashSet;
56 import java.util.Iterator;
57 import java.util.LinkedHashMap;
58 import java.util.List;
60 import java.util.Map.Entry;
61 import java.util.NoSuchElementException;
63 import java.util.SortedMap;
64 import java.util.TreeMap;
67 * grab bag of useful alignment manipulation operations Expect these to be
68 * refactored elsewhere at some point.
73 public class AlignmentUtils
76 private static final int CODON_LENGTH = 3;
78 private static final String SEQUENCE_VARIANT = "sequence_variant:";
80 private static final String ID = "ID";
83 * A data model to hold the 'normal' base value at a position, and an optional
84 * sequence variant feature
86 static final class DnaVariant
90 SequenceFeature variant;
92 DnaVariant(String nuc)
98 DnaVariant(String nuc, SequenceFeature var)
104 public String getSource()
106 return variant == null ? null : variant.getFeatureGroup();
110 * toString for aid in the debugger only
113 public String toString()
115 return base + ":" + (variant == null ? "" : variant.getDescription());
120 * given an existing alignment, create a new alignment including all, or up to
121 * flankSize additional symbols from each sequence's dataset sequence
127 public static AlignmentI expandContext(AlignmentI core, int flankSize)
129 List<SequenceI> sq = new ArrayList<SequenceI>();
131 for (SequenceI s : core.getSequences())
133 SequenceI newSeq = s.deriveSequence();
134 final int newSeqStart = newSeq.getStart() - 1;
135 if (newSeqStart > maxoffset
136 && newSeq.getDatasetSequence().getStart() < s.getStart())
138 maxoffset = newSeqStart;
144 maxoffset = Math.min(maxoffset, flankSize);
148 * now add offset left and right to create an expanded alignment
150 for (SequenceI s : sq)
153 while (ds.getDatasetSequence() != null)
155 ds = ds.getDatasetSequence();
157 int s_end = s.findPosition(s.getStart() + s.getLength());
158 // find available flanking residues for sequence
159 int ustream_ds = s.getStart() - ds.getStart();
160 int dstream_ds = ds.getEnd() - s_end;
162 // build new flanked sequence
164 // compute gap padding to start of flanking sequence
165 int offset = maxoffset - ustream_ds;
167 // padding is gapChar x ( maxoffset - min(ustream_ds, flank)
170 if (flankSize < ustream_ds)
172 // take up to flankSize residues
173 offset = maxoffset - flankSize;
174 ustream_ds = flankSize;
176 if (flankSize <= dstream_ds)
178 dstream_ds = flankSize - 1;
181 // TODO use Character.toLowerCase to avoid creating String objects?
182 char[] upstream = new String(ds
183 .getSequence(s.getStart() - 1 - ustream_ds, s.getStart() - 1))
184 .toLowerCase().toCharArray();
185 char[] downstream = new String(
186 ds.getSequence(s_end - 1, s_end + dstream_ds)).toLowerCase()
188 char[] coreseq = s.getSequence();
189 char[] nseq = new char[offset + upstream.length + downstream.length
191 char c = core.getGapCharacter();
194 for (; p < offset; p++)
199 System.arraycopy(upstream, 0, nseq, p, upstream.length);
200 System.arraycopy(coreseq, 0, nseq, p + upstream.length,
202 System.arraycopy(downstream, 0, nseq,
203 p + coreseq.length + upstream.length, downstream.length);
204 s.setSequence(new String(nseq));
205 s.setStart(s.getStart() - ustream_ds);
206 s.setEnd(s_end + downstream.length);
208 AlignmentI newAl = new jalview.datamodel.Alignment(
209 sq.toArray(new SequenceI[0]));
210 for (SequenceI s : sq)
212 if (s.getAnnotation() != null)
214 for (AlignmentAnnotation aa : s.getAnnotation())
216 aa.adjustForAlignment(); // JAL-1712 fix
217 newAl.addAnnotation(aa);
221 newAl.setDataset(core.getDataset());
226 * Returns the index (zero-based position) of a sequence in an alignment, or
233 public static int getSequenceIndex(AlignmentI al, SequenceI seq)
237 for (SequenceI alSeq : al.getSequences())
250 * Returns a map of lists of sequences in the alignment, keyed by sequence
251 * name. For use in mapping between different alignment views of the same
254 * @see jalview.datamodel.AlignmentI#getSequencesByName()
256 public static Map<String, List<SequenceI>> getSequencesByName(
259 Map<String, List<SequenceI>> theMap = new LinkedHashMap<String, List<SequenceI>>();
260 for (SequenceI seq : al.getSequences())
262 String name = seq.getName();
265 List<SequenceI> seqs = theMap.get(name);
268 seqs = new ArrayList<SequenceI>();
269 theMap.put(name, seqs);
278 * Build mapping of protein to cDNA alignment. Mappings are made between
279 * sequences where the cDNA translates to the protein sequence. Any new
280 * mappings are added to the protein alignment. Returns true if any mappings
281 * either already exist or were added, else false.
283 * @param proteinAlignment
284 * @param cdnaAlignment
287 public static boolean mapProteinAlignmentToCdna(
288 final AlignmentI proteinAlignment, final AlignmentI cdnaAlignment)
290 if (proteinAlignment == null || cdnaAlignment == null)
295 Set<SequenceI> mappedDna = new HashSet<SequenceI>();
296 Set<SequenceI> mappedProtein = new HashSet<SequenceI>();
299 * First pass - map sequences where cross-references exist. This include
300 * 1-to-many mappings to support, for example, variant cDNA.
302 boolean mappingPerformed = mapProteinToCdna(proteinAlignment,
303 cdnaAlignment, mappedDna, mappedProtein, true);
306 * Second pass - map sequences where no cross-references exist. This only
307 * does 1-to-1 mappings and assumes corresponding sequences are in the same
308 * order in the alignments.
310 mappingPerformed |= mapProteinToCdna(proteinAlignment, cdnaAlignment,
311 mappedDna, mappedProtein, false);
312 return mappingPerformed;
316 * Make mappings between compatible sequences (where the cDNA translation
317 * matches the protein).
319 * @param proteinAlignment
320 * @param cdnaAlignment
322 * a set of mapped DNA sequences (to add to)
323 * @param mappedProtein
324 * a set of mapped Protein sequences (to add to)
326 * if true, only map sequences where xrefs exist
329 protected static boolean mapProteinToCdna(
330 final AlignmentI proteinAlignment, final AlignmentI cdnaAlignment,
331 Set<SequenceI> mappedDna, Set<SequenceI> mappedProtein,
334 boolean mappingExistsOrAdded = false;
335 List<SequenceI> thisSeqs = proteinAlignment.getSequences();
336 for (SequenceI aaSeq : thisSeqs)
338 boolean proteinMapped = false;
339 AlignedCodonFrame acf = new AlignedCodonFrame();
341 for (SequenceI cdnaSeq : cdnaAlignment.getSequences())
344 * Always try to map if sequences have xref to each other; this supports
345 * variant cDNA or alternative splicing for a protein sequence.
347 * If no xrefs, try to map progressively, assuming that alignments have
348 * mappable sequences in corresponding order. These are not
349 * many-to-many, as that would risk mixing species with similar cDNA
352 if (xrefsOnly && !AlignmentUtils.haveCrossRef(aaSeq, cdnaSeq))
358 * Don't map non-xrefd sequences more than once each. This heuristic
359 * allows us to pair up similar sequences in ordered alignments.
361 if (!xrefsOnly && (mappedProtein.contains(aaSeq)
362 || mappedDna.contains(cdnaSeq)))
366 if (mappingExists(proteinAlignment.getCodonFrames(),
367 aaSeq.getDatasetSequence(), cdnaSeq.getDatasetSequence()))
369 mappingExistsOrAdded = true;
373 MapList map = mapCdnaToProtein(aaSeq, cdnaSeq);
376 acf.addMap(cdnaSeq, aaSeq, map);
377 mappingExistsOrAdded = true;
378 proteinMapped = true;
379 mappedDna.add(cdnaSeq);
380 mappedProtein.add(aaSeq);
386 proteinAlignment.addCodonFrame(acf);
389 return mappingExistsOrAdded;
393 * Answers true if the mappings include one between the given (dataset)
396 public static boolean mappingExists(List<AlignedCodonFrame> mappings,
397 SequenceI aaSeq, SequenceI cdnaSeq)
399 if (mappings != null)
401 for (AlignedCodonFrame acf : mappings)
403 if (cdnaSeq == acf.getDnaForAaSeq(aaSeq))
413 * Builds a mapping (if possible) of a cDNA to a protein sequence.
415 * <li>first checks if the cdna translates exactly to the protein
417 * <li>else checks for translation after removing a STOP codon</li>
418 * <li>else checks for translation after removing a START codon</li>
419 * <li>if that fails, inspect CDS features on the cDNA sequence</li>
421 * Returns null if no mapping is determined.
424 * the aligned protein sequence
426 * the aligned cdna sequence
429 public static MapList mapCdnaToProtein(SequenceI proteinSeq,
433 * Here we handle either dataset sequence set (desktop) or absent (applet).
434 * Use only the char[] form of the sequence to avoid creating possibly large
437 final SequenceI proteinDataset = proteinSeq.getDatasetSequence();
438 char[] aaSeqChars = proteinDataset != null
439 ? proteinDataset.getSequence()
440 : proteinSeq.getSequence();
441 final SequenceI cdnaDataset = cdnaSeq.getDatasetSequence();
442 char[] cdnaSeqChars = cdnaDataset != null ? cdnaDataset.getSequence()
443 : cdnaSeq.getSequence();
444 if (aaSeqChars == null || cdnaSeqChars == null)
450 * cdnaStart/End, proteinStartEnd are base 1 (for dataset sequence mapping)
452 final int mappedLength = CODON_LENGTH * aaSeqChars.length;
453 int cdnaLength = cdnaSeqChars.length;
454 int cdnaStart = cdnaSeq.getStart();
455 int cdnaEnd = cdnaSeq.getEnd();
456 final int proteinStart = proteinSeq.getStart();
457 final int proteinEnd = proteinSeq.getEnd();
460 * If lengths don't match, try ignoring stop codon (if present)
462 if (cdnaLength != mappedLength && cdnaLength > 2)
464 String lastCodon = String.valueOf(cdnaSeqChars,
465 cdnaLength - CODON_LENGTH, CODON_LENGTH).toUpperCase();
466 for (String stop : ResidueProperties.STOP)
468 if (lastCodon.equals(stop))
470 cdnaEnd -= CODON_LENGTH;
471 cdnaLength -= CODON_LENGTH;
478 * If lengths still don't match, try ignoring start codon.
481 if (cdnaLength != mappedLength && cdnaLength > 2
482 && String.valueOf(cdnaSeqChars, 0, CODON_LENGTH).toUpperCase()
483 .equals(ResidueProperties.START))
485 startOffset += CODON_LENGTH;
486 cdnaStart += CODON_LENGTH;
487 cdnaLength -= CODON_LENGTH;
490 if (translatesAs(cdnaSeqChars, startOffset, aaSeqChars))
493 * protein is translation of dna (+/- start/stop codons)
495 MapList map = new MapList(new int[] { cdnaStart, cdnaEnd },
497 { proteinStart, proteinEnd }, CODON_LENGTH, 1);
502 * translation failed - try mapping CDS annotated regions of dna
504 return mapCdsToProtein(cdnaSeq, proteinSeq);
508 * Test whether the given cdna sequence, starting at the given offset,
509 * translates to the given amino acid sequence, using the standard translation
510 * table. Designed to fail fast i.e. as soon as a mismatch position is found.
512 * @param cdnaSeqChars
517 protected static boolean translatesAs(char[] cdnaSeqChars, int cdnaStart,
520 if (cdnaSeqChars == null || aaSeqChars == null)
526 int dnaPos = cdnaStart;
527 for (; dnaPos < cdnaSeqChars.length - 2
528 && aaPos < aaSeqChars.length; dnaPos += CODON_LENGTH, aaPos++)
530 String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH);
531 final String translated = ResidueProperties.codonTranslate(codon);
534 * allow * in protein to match untranslatable in dna
536 final char aaRes = aaSeqChars[aaPos];
537 if ((translated == null || "STOP".equals(translated)) && aaRes == '*')
541 if (translated == null || !(aaRes == translated.charAt(0)))
544 // System.out.println(("Mismatch at " + i + "/" + aaResidue + ": "
545 // + codon + "(" + translated + ") != " + aaRes));
551 * check we matched all of the protein sequence
553 if (aaPos != aaSeqChars.length)
559 * check we matched all of the dna except
560 * for optional trailing STOP codon
562 if (dnaPos == cdnaSeqChars.length)
566 if (dnaPos == cdnaSeqChars.length - CODON_LENGTH)
568 String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH);
569 if ("STOP".equals(ResidueProperties.codonTranslate(codon)))
578 * Align sequence 'seq' to match the alignment of a mapped sequence. Note this
579 * currently assumes that we are aligning cDNA to match protein.
582 * the sequence to be realigned
584 * the alignment whose sequence alignment is to be 'copied'
586 * character string represent a gap in the realigned sequence
587 * @param preserveUnmappedGaps
588 * @param preserveMappedGaps
589 * @return true if the sequence was realigned, false if it could not be
591 public static boolean alignSequenceAs(SequenceI seq, AlignmentI al,
592 String gap, boolean preserveMappedGaps,
593 boolean preserveUnmappedGaps)
596 * Get any mappings from the source alignment to the target (dataset)
599 // TODO there may be one AlignedCodonFrame per dataset sequence, or one with
600 // all mappings. Would it help to constrain this?
601 List<AlignedCodonFrame> mappings = al.getCodonFrame(seq);
602 if (mappings == null || mappings.isEmpty())
608 * Locate the aligned source sequence whose dataset sequence is mapped. We
609 * just take the first match here (as we can't align like more than one
612 SequenceI alignFrom = null;
613 AlignedCodonFrame mapping = null;
614 for (AlignedCodonFrame mp : mappings)
616 alignFrom = mp.findAlignedSequence(seq, al);
617 if (alignFrom != null)
624 if (alignFrom == null)
628 alignSequenceAs(seq, alignFrom, mapping, gap, al.getGapCharacter(),
629 preserveMappedGaps, preserveUnmappedGaps);
634 * Align sequence 'alignTo' the same way as 'alignFrom', using the mapping to
635 * match residues and codons. Flags control whether existing gaps in unmapped
636 * (intron) and mapped (exon) regions are preserved or not. Gaps between
637 * intron and exon are only retained if both flags are set.
644 * @param preserveUnmappedGaps
645 * @param preserveMappedGaps
647 public static void alignSequenceAs(SequenceI alignTo, SequenceI alignFrom,
648 AlignedCodonFrame mapping, String myGap, char sourceGap,
649 boolean preserveMappedGaps, boolean preserveUnmappedGaps)
651 // TODO generalise to work for Protein-Protein, dna-dna, dna-protein
653 // aligned and dataset sequence positions, all base zero
657 int basesWritten = 0;
658 char myGapChar = myGap.charAt(0);
659 int ratio = myGap.length();
661 int fromOffset = alignFrom.getStart() - 1;
662 int toOffset = alignTo.getStart() - 1;
663 int sourceGapMappedLength = 0;
664 boolean inExon = false;
665 final int toLength = alignTo.getLength();
666 final int fromLength = alignFrom.getLength();
667 StringBuilder thisAligned = new StringBuilder(2 * toLength);
670 * Traverse the 'model' aligned sequence
672 for (int i = 0; i < fromLength; i++)
674 char sourceChar = alignFrom.getCharAt(i);
675 if (sourceChar == sourceGap)
677 sourceGapMappedLength += ratio;
682 * Found a non-gap character. Locate its mapped region if any.
685 // Note mapping positions are base 1, our sequence positions base 0
686 int[] mappedPos = mapping.getMappedRegion(alignTo, alignFrom,
687 sourceDsPos + fromOffset);
688 if (mappedPos == null)
691 * unmapped position; treat like a gap
693 sourceGapMappedLength += ratio;
694 // System.err.println("Can't align: no codon mapping to residue "
695 // + sourceDsPos + "(" + sourceChar + ")");
700 int mappedCodonStart = mappedPos[0]; // position (1...) of codon start
701 int mappedCodonEnd = mappedPos[mappedPos.length - 1]; // codon end pos
702 StringBuilder trailingCopiedGap = new StringBuilder();
705 * Copy dna sequence up to and including this codon. Optionally, include
706 * gaps before the codon starts (in introns) and/or after the codon starts
709 * Note this only works for 'linear' splicing, not reverse or interleaved.
710 * But then 'align dna as protein' doesn't make much sense otherwise.
712 int intronLength = 0;
713 while (basesWritten + toOffset < mappedCodonEnd
714 && thisSeqPos < toLength)
716 final char c = alignTo.getCharAt(thisSeqPos++);
720 int sourcePosition = basesWritten + toOffset;
721 if (sourcePosition < mappedCodonStart)
724 * Found an unmapped (intron) base. First add in any preceding gaps
727 if (preserveUnmappedGaps && trailingCopiedGap.length() > 0)
729 thisAligned.append(trailingCopiedGap.toString());
730 intronLength += trailingCopiedGap.length();
731 trailingCopiedGap = new StringBuilder();
738 final boolean startOfCodon = sourcePosition == mappedCodonStart;
739 int gapsToAdd = calculateGapsToInsert(preserveMappedGaps,
740 preserveUnmappedGaps, sourceGapMappedLength, inExon,
741 trailingCopiedGap.length(), intronLength, startOfCodon);
742 for (int k = 0; k < gapsToAdd; k++)
744 thisAligned.append(myGapChar);
746 sourceGapMappedLength = 0;
749 thisAligned.append(c);
750 trailingCopiedGap = new StringBuilder();
754 if (inExon && preserveMappedGaps)
756 trailingCopiedGap.append(myGapChar);
758 else if (!inExon && preserveUnmappedGaps)
760 trailingCopiedGap.append(myGapChar);
767 * At end of model aligned sequence. Copy any remaining target sequence, optionally
768 * including (intron) gaps.
770 while (thisSeqPos < toLength)
772 final char c = alignTo.getCharAt(thisSeqPos++);
773 if (c != myGapChar || preserveUnmappedGaps)
775 thisAligned.append(c);
777 sourceGapMappedLength--;
781 * finally add gaps to pad for any trailing source gaps or
782 * unmapped characters
784 if (preserveUnmappedGaps)
786 while (sourceGapMappedLength > 0)
788 thisAligned.append(myGapChar);
789 sourceGapMappedLength--;
794 * All done aligning, set the aligned sequence.
796 alignTo.setSequence(new String(thisAligned));
800 * Helper method to work out how many gaps to insert when realigning.
802 * @param preserveMappedGaps
803 * @param preserveUnmappedGaps
804 * @param sourceGapMappedLength
806 * @param trailingCopiedGap
807 * @param intronLength
808 * @param startOfCodon
811 protected static int calculateGapsToInsert(boolean preserveMappedGaps,
812 boolean preserveUnmappedGaps, int sourceGapMappedLength,
813 boolean inExon, int trailingGapLength, int intronLength,
814 final boolean startOfCodon)
820 * Reached start of codon. Ignore trailing gaps in intron unless we are
821 * preserving gaps in both exon and intron. Ignore them anyway if the
822 * protein alignment introduces a gap at least as large as the intronic
825 if (inExon && !preserveMappedGaps)
827 trailingGapLength = 0;
829 if (!inExon && !(preserveMappedGaps && preserveUnmappedGaps))
831 trailingGapLength = 0;
835 gapsToAdd = Math.max(sourceGapMappedLength, trailingGapLength);
839 if (intronLength + trailingGapLength <= sourceGapMappedLength)
841 gapsToAdd = sourceGapMappedLength - intronLength;
845 gapsToAdd = Math.min(
846 intronLength + trailingGapLength - sourceGapMappedLength,
854 * second or third base of codon; check for any gaps in dna
856 if (!preserveMappedGaps)
858 trailingGapLength = 0;
860 gapsToAdd = Math.max(sourceGapMappedLength, trailingGapLength);
866 * Realigns the given protein to match the alignment of the dna, using codon
867 * mappings to translate aligned codon positions to protein residues.
870 * the alignment whose sequences are realigned by this method
872 * the dna alignment whose alignment we are 'copying'
873 * @return the number of sequences that were realigned
875 public static int alignProteinAsDna(AlignmentI protein, AlignmentI dna)
877 if (protein.isNucleotide() || !dna.isNucleotide())
879 System.err.println("Wrong alignment type in alignProteinAsDna");
882 List<SequenceI> unmappedProtein = new ArrayList<SequenceI>();
883 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons = buildCodonColumnsMap(
884 protein, dna, unmappedProtein);
885 return alignProteinAs(protein, alignedCodons, unmappedProtein);
889 * Realigns the given dna to match the alignment of the protein, using codon
890 * mappings to translate aligned peptide positions to codons.
892 * Always produces a padded CDS alignment.
895 * the alignment whose sequences are realigned by this method
897 * the protein alignment whose alignment we are 'copying'
898 * @return the number of sequences that were realigned
900 public static int alignCdsAsProtein(AlignmentI dna, AlignmentI protein)
902 if (protein.isNucleotide() || !dna.isNucleotide())
904 System.err.println("Wrong alignment type in alignProteinAsDna");
907 // todo: implement this
908 List<AlignedCodonFrame> mappings = protein.getCodonFrames();
909 int alignedCount = 0;
910 int width = 0; // alignment width for padding CDS
911 for (SequenceI dnaSeq : dna.getSequences())
913 if (alignCdsSequenceAsProtein(dnaSeq, protein, mappings,
914 dna.getGapCharacter()))
918 width = Math.max(dnaSeq.getLength(), width);
922 for (SequenceI dnaSeq : dna.getSequences())
924 oldwidth = dnaSeq.getLength();
925 diff = width - oldwidth;
928 dnaSeq.insertCharAt(oldwidth, diff, dna.getGapCharacter());
935 * Helper method to align (if possible) the dna sequence to match the
936 * alignment of a mapped protein sequence. This is currently limited to
937 * handling coding sequence only.
945 static boolean alignCdsSequenceAsProtein(SequenceI cdsSeq,
946 AlignmentI protein, List<AlignedCodonFrame> mappings,
949 SequenceI cdsDss = cdsSeq.getDatasetSequence();
953 .println("alignCdsSequenceAsProtein needs aligned sequence!");
957 List<AlignedCodonFrame> dnaMappings = MappingUtils
958 .findMappingsForSequence(cdsSeq, mappings);
959 for (AlignedCodonFrame mapping : dnaMappings)
961 SequenceI peptide = mapping.findAlignedSequence(cdsSeq, protein);
964 final int peptideLength = peptide.getLength();
965 Mapping map = mapping.getMappingBetween(cdsSeq, peptide);
968 MapList mapList = map.getMap();
969 if (map.getTo() == peptide.getDatasetSequence())
971 mapList = mapList.getInverse();
973 final int cdsLength = cdsDss.getLength();
974 int mappedFromLength = MappingUtils.getLength(mapList
976 int mappedToLength = MappingUtils
977 .getLength(mapList.getToRanges());
978 boolean addStopCodon = (cdsLength == mappedFromLength
979 * CODON_LENGTH + CODON_LENGTH)
980 || (peptide.getDatasetSequence()
981 .getLength() == mappedFromLength - 1);
982 if (cdsLength != mappedToLength && !addStopCodon)
984 System.err.println(String.format(
985 "Can't align cds as protein (length mismatch %d/%d): %s",
986 cdsLength, mappedToLength, cdsSeq.getName()));
990 * pre-fill the aligned cds sequence with gaps
992 char[] alignedCds = new char[peptideLength * CODON_LENGTH
993 + (addStopCodon ? CODON_LENGTH : 0)];
994 Arrays.fill(alignedCds, gapChar);
997 * walk over the aligned peptide sequence and insert mapped
998 * codons for residues in the aligned cds sequence
1000 int copiedBases = 0;
1001 int cdsStart = cdsDss.getStart();
1002 int proteinPos = peptide.getStart() - 1;
1005 for (int col = 0; col < peptideLength; col++)
1007 char residue = peptide.getCharAt(col);
1009 if (Comparison.isGap(residue))
1011 cdsCol += CODON_LENGTH;
1016 int[] codon = mapList.locateInTo(proteinPos, proteinPos);
1019 // e.g. incomplete start codon, X in peptide
1020 cdsCol += CODON_LENGTH;
1024 for (int j = codon[0]; j <= codon[1]; j++)
1026 char mappedBase = cdsDss.getCharAt(j - cdsStart);
1027 alignedCds[cdsCol++] = mappedBase;
1035 * append stop codon if not mapped from protein,
1036 * closing it up to the end of the mapped sequence
1038 if (copiedBases == cdsLength - CODON_LENGTH)
1040 for (int i = alignedCds.length - 1; i >= 0; i--)
1042 if (!Comparison.isGap(alignedCds[i]))
1044 cdsCol = i + 1; // gap just after end of sequence
1048 for (int i = cdsLength - CODON_LENGTH; i < cdsLength; i++)
1050 alignedCds[cdsCol++] = cdsDss.getCharAt(i);
1053 cdsSeq.setSequence(new String(alignedCds));
1062 * Builds a map whose key is an aligned codon position (3 alignment column
1063 * numbers base 0), and whose value is a map from protein sequence to each
1064 * protein's peptide residue for that codon. The map generates an ordering of
1065 * the codons, and allows us to read off the peptides at each position in
1066 * order to assemble 'aligned' protein sequences.
1069 * the protein alignment
1071 * the coding dna alignment
1072 * @param unmappedProtein
1073 * any unmapped proteins are added to this list
1076 protected static Map<AlignedCodon, Map<SequenceI, AlignedCodon>> buildCodonColumnsMap(
1077 AlignmentI protein, AlignmentI dna,
1078 List<SequenceI> unmappedProtein)
1081 * maintain a list of any proteins with no mappings - these will be
1082 * rendered 'as is' in the protein alignment as we can't align them
1084 unmappedProtein.addAll(protein.getSequences());
1086 List<AlignedCodonFrame> mappings = protein.getCodonFrames();
1089 * Map will hold, for each aligned codon position e.g. [3, 5, 6], a map of
1090 * {dnaSequence, {proteinSequence, codonProduct}} at that position. The
1091 * comparator keeps the codon positions ordered.
1093 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons = new TreeMap<AlignedCodon, Map<SequenceI, AlignedCodon>>(
1094 new CodonComparator());
1096 for (SequenceI dnaSeq : dna.getSequences())
1098 for (AlignedCodonFrame mapping : mappings)
1100 SequenceI prot = mapping.findAlignedSequence(dnaSeq, protein);
1103 Mapping seqMap = mapping.getMappingForSequence(dnaSeq);
1104 addCodonPositions(dnaSeq, prot, protein.getGapCharacter(), seqMap,
1106 unmappedProtein.remove(prot);
1112 * Finally add any unmapped peptide start residues (e.g. for incomplete
1113 * codons) as if at the codon position before the second residue
1115 // TODO resolve JAL-2022 so this fudge can be removed
1116 int mappedSequenceCount = protein.getHeight() - unmappedProtein.size();
1117 addUnmappedPeptideStarts(alignedCodons, mappedSequenceCount);
1119 return alignedCodons;
1123 * Scans for any protein mapped from position 2 (meaning unmapped start
1124 * position e.g. an incomplete codon), and synthesizes a 'codon' for it at the
1125 * preceding position in the alignment
1127 * @param alignedCodons
1128 * the codon-to-peptide map
1129 * @param mappedSequenceCount
1130 * the number of distinct sequences in the map
1132 protected static void addUnmappedPeptideStarts(
1133 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons,
1134 int mappedSequenceCount)
1136 // TODO delete this ugly hack once JAL-2022 is resolved
1137 // i.e. we can model startPhase > 0 (incomplete start codon)
1139 List<SequenceI> sequencesChecked = new ArrayList<SequenceI>();
1140 AlignedCodon lastCodon = null;
1141 Map<SequenceI, AlignedCodon> toAdd = new HashMap<SequenceI, AlignedCodon>();
1143 for (Entry<AlignedCodon, Map<SequenceI, AlignedCodon>> entry : alignedCodons
1146 for (Entry<SequenceI, AlignedCodon> sequenceCodon : entry.getValue()
1149 SequenceI seq = sequenceCodon.getKey();
1150 if (sequencesChecked.contains(seq))
1154 sequencesChecked.add(seq);
1155 AlignedCodon codon = sequenceCodon.getValue();
1156 if (codon.peptideCol > 1)
1159 "Problem mapping protein with >1 unmapped start positions: "
1162 else if (codon.peptideCol == 1)
1165 * first position (peptideCol == 0) was unmapped - add it
1167 if (lastCodon != null)
1169 AlignedCodon firstPeptide = new AlignedCodon(lastCodon.pos1,
1170 lastCodon.pos2, lastCodon.pos3,
1171 String.valueOf(seq.getCharAt(0)), 0);
1172 toAdd.put(seq, firstPeptide);
1177 * unmapped residue at start of alignment (no prior column) -
1178 * 'insert' at nominal codon [0, 0, 0]
1180 AlignedCodon firstPeptide = new AlignedCodon(0, 0, 0,
1181 String.valueOf(seq.getCharAt(0)), 0);
1182 toAdd.put(seq, firstPeptide);
1185 if (sequencesChecked.size() == mappedSequenceCount)
1187 // no need to check past first mapped position in all sequences
1191 lastCodon = entry.getKey();
1195 * add any new codons safely after iterating over the map
1197 for (Entry<SequenceI, AlignedCodon> startCodon : toAdd.entrySet())
1199 addCodonToMap(alignedCodons, startCodon.getValue(),
1200 startCodon.getKey());
1205 * Update the aligned protein sequences to match the codon alignments given in
1209 * @param alignedCodons
1210 * an ordered map of codon positions (columns), with sequence/peptide
1211 * values present in each column
1212 * @param unmappedProtein
1215 protected static int alignProteinAs(AlignmentI protein,
1216 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons,
1217 List<SequenceI> unmappedProtein)
1220 * prefill peptide sequences with gaps
1222 int alignedWidth = alignedCodons.size();
1223 char[] gaps = new char[alignedWidth];
1224 Arrays.fill(gaps, protein.getGapCharacter());
1225 Map<SequenceI, char[]> peptides = new HashMap<>();
1226 for (SequenceI seq : protein.getSequences())
1228 if (!unmappedProtein.contains(seq))
1230 peptides.put(seq, Arrays.copyOf(gaps, gaps.length));
1235 * Traverse the codons left to right (as defined by CodonComparator)
1236 * and insert peptides in each column where the sequence is mapped.
1237 * This gives a peptide 'alignment' where residues are aligned if their
1238 * corresponding codons occupy the same columns in the cdna alignment.
1241 for (AlignedCodon codon : alignedCodons.keySet())
1243 final Map<SequenceI, AlignedCodon> columnResidues = alignedCodons
1245 for (Entry<SequenceI, AlignedCodon> entry : columnResidues.entrySet())
1247 char residue = entry.getValue().product.charAt(0);
1248 peptides.get(entry.getKey())[column] = residue;
1254 * and finally set the constructed sequences
1256 for (Entry<SequenceI, char[]> entry : peptides.entrySet())
1258 entry.getKey().setSequence(new String(entry.getValue()));
1265 * Populate the map of aligned codons by traversing the given sequence
1266 * mapping, locating the aligned positions of mapped codons, and adding those
1267 * positions and their translation products to the map.
1270 * the aligned sequence we are mapping from
1272 * the sequence to be aligned to the codons
1274 * the gap character in the dna sequence
1276 * a mapping to a sequence translation
1277 * @param alignedCodons
1278 * the map we are building up
1280 static void addCodonPositions(SequenceI dna, SequenceI protein,
1281 char gapChar, Mapping seqMap,
1282 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons)
1284 Iterator<AlignedCodon> codons = seqMap.getCodonIterator(dna, gapChar);
1287 * add codon positions, and their peptide translations, to the alignment
1288 * map, while remembering the first codon mapped
1290 while (codons.hasNext())
1294 AlignedCodon codon = codons.next();
1295 addCodonToMap(alignedCodons, codon, protein);
1296 } catch (IncompleteCodonException e)
1298 // possible incomplete trailing codon - ignore
1299 } catch (NoSuchElementException e)
1301 // possibly peptide lacking STOP
1307 * Helper method to add a codon-to-peptide entry to the aligned codons map
1309 * @param alignedCodons
1313 protected static void addCodonToMap(
1314 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons,
1315 AlignedCodon codon, SequenceI protein)
1317 Map<SequenceI, AlignedCodon> seqProduct = alignedCodons.get(codon);
1318 if (seqProduct == null)
1320 seqProduct = new HashMap<SequenceI, AlignedCodon>();
1321 alignedCodons.put(codon, seqProduct);
1323 seqProduct.put(protein, codon);
1327 * Returns true if a cDNA/Protein mapping either exists, or could be made,
1328 * between at least one pair of sequences in the two alignments. Currently,
1331 * <li>One alignment must be nucleotide, and the other protein</li>
1332 * <li>At least one pair of sequences must be already mapped, or mappable</li>
1333 * <li>Mappable means the nucleotide translation matches the protein
1335 * <li>The translation may ignore start and stop codons if present in the
1343 public static boolean isMappable(AlignmentI al1, AlignmentI al2)
1345 if (al1 == null || al2 == null)
1351 * Require one nucleotide and one protein
1353 if (al1.isNucleotide() == al2.isNucleotide())
1357 AlignmentI dna = al1.isNucleotide() ? al1 : al2;
1358 AlignmentI protein = dna == al1 ? al2 : al1;
1359 List<AlignedCodonFrame> mappings = protein.getCodonFrames();
1360 for (SequenceI dnaSeq : dna.getSequences())
1362 for (SequenceI proteinSeq : protein.getSequences())
1364 if (isMappable(dnaSeq, proteinSeq, mappings))
1374 * Returns true if the dna sequence is mapped, or could be mapped, to the
1382 protected static boolean isMappable(SequenceI dnaSeq,
1383 SequenceI proteinSeq, List<AlignedCodonFrame> mappings)
1385 if (dnaSeq == null || proteinSeq == null)
1390 SequenceI dnaDs = dnaSeq.getDatasetSequence() == null ? dnaSeq
1391 : dnaSeq.getDatasetSequence();
1392 SequenceI proteinDs = proteinSeq.getDatasetSequence() == null
1394 : proteinSeq.getDatasetSequence();
1396 for (AlignedCodonFrame mapping : mappings)
1398 if (proteinDs == mapping.getAaForDnaSeq(dnaDs))
1408 * Just try to make a mapping (it is not yet stored), test whether
1411 return mapCdnaToProtein(proteinDs, dnaDs) != null;
1415 * Finds any reference annotations associated with the sequences in
1416 * sequenceScope, that are not already added to the alignment, and adds them
1417 * to the 'candidates' map. Also populates a lookup table of annotation
1418 * labels, keyed by calcId, for use in constructing tooltips or the like.
1420 * @param sequenceScope
1421 * the sequences to scan for reference annotations
1422 * @param labelForCalcId
1423 * (optional) map to populate with label for calcId
1425 * map to populate with annotations for sequence
1427 * the alignment to check for presence of annotations
1429 public static void findAddableReferenceAnnotations(
1430 List<SequenceI> sequenceScope, Map<String, String> labelForCalcId,
1431 final Map<SequenceI, List<AlignmentAnnotation>> candidates,
1434 if (sequenceScope == null)
1440 * For each sequence in scope, make a list of any annotations on the
1441 * underlying dataset sequence which are not already on the alignment.
1443 * Add to a map of { alignmentSequence, <List of annotations to add> }
1445 for (SequenceI seq : sequenceScope)
1447 SequenceI dataset = seq.getDatasetSequence();
1448 if (dataset == null)
1452 AlignmentAnnotation[] datasetAnnotations = dataset.getAnnotation();
1453 if (datasetAnnotations == null)
1457 final List<AlignmentAnnotation> result = new ArrayList<AlignmentAnnotation>();
1458 for (AlignmentAnnotation dsann : datasetAnnotations)
1461 * Find matching annotations on the alignment. If none is found, then
1462 * add this annotation to the list of 'addable' annotations for this
1465 final Iterable<AlignmentAnnotation> matchedAlignmentAnnotations = al
1466 .findAnnotations(seq, dsann.getCalcId(), dsann.label);
1467 if (!matchedAlignmentAnnotations.iterator().hasNext())
1470 if (labelForCalcId != null)
1472 labelForCalcId.put(dsann.getCalcId(), dsann.label);
1477 * Save any addable annotations for this sequence
1479 if (!result.isEmpty())
1481 candidates.put(seq, result);
1487 * Adds annotations to the top of the alignment annotations, in the same order
1488 * as their related sequences.
1490 * @param annotations
1491 * the annotations to add
1493 * the alignment to add them to
1494 * @param selectionGroup
1495 * current selection group (or null if none)
1497 public static void addReferenceAnnotations(
1498 Map<SequenceI, List<AlignmentAnnotation>> annotations,
1499 final AlignmentI alignment, final SequenceGroup selectionGroup)
1501 for (SequenceI seq : annotations.keySet())
1503 for (AlignmentAnnotation ann : annotations.get(seq))
1505 AlignmentAnnotation copyAnn = new AlignmentAnnotation(ann);
1507 int endRes = ann.annotations.length;
1508 if (selectionGroup != null)
1510 startRes = selectionGroup.getStartRes();
1511 endRes = selectionGroup.getEndRes();
1513 copyAnn.restrict(startRes, endRes);
1516 * Add to the sequence (sets copyAnn.datasetSequence), unless the
1517 * original annotation is already on the sequence.
1519 if (!seq.hasAnnotation(ann))
1521 seq.addAlignmentAnnotation(copyAnn);
1524 copyAnn.adjustForAlignment();
1525 // add to the alignment and set visible
1526 alignment.addAnnotation(copyAnn);
1527 copyAnn.visible = true;
1533 * Set visibility of alignment annotations of specified types (labels), for
1534 * specified sequences. This supports controls like "Show all secondary
1535 * structure", "Hide all Temp factor", etc.
1537 * @al the alignment to scan for annotations
1539 * the types (labels) of annotations to be updated
1540 * @param forSequences
1541 * if not null, only annotations linked to one of these sequences are
1542 * in scope for update; if null, acts on all sequence annotations
1544 * if this flag is true, 'types' is ignored (label not checked)
1546 * if true, set visibility on, else set off
1548 public static void showOrHideSequenceAnnotations(AlignmentI al,
1549 Collection<String> types, List<SequenceI> forSequences,
1550 boolean anyType, boolean doShow)
1552 AlignmentAnnotation[] anns = al.getAlignmentAnnotation();
1555 for (AlignmentAnnotation aa : anns)
1557 if (anyType || types.contains(aa.label))
1559 if ((aa.sequenceRef != null) && (forSequences == null
1560 || forSequences.contains(aa.sequenceRef)))
1562 aa.visible = doShow;
1570 * Returns true if either sequence has a cross-reference to the other
1576 public static boolean haveCrossRef(SequenceI seq1, SequenceI seq2)
1578 // Note: moved here from class CrossRef as the latter class has dependencies
1579 // not availability to the applet's classpath
1580 return hasCrossRef(seq1, seq2) || hasCrossRef(seq2, seq1);
1584 * Returns true if seq1 has a cross-reference to seq2. Currently this assumes
1585 * that sequence name is structured as Source|AccessionId.
1591 public static boolean hasCrossRef(SequenceI seq1, SequenceI seq2)
1593 if (seq1 == null || seq2 == null)
1597 String name = seq2.getName();
1598 final DBRefEntry[] xrefs = seq1.getDBRefs();
1601 for (DBRefEntry xref : xrefs)
1603 String xrefName = xref.getSource() + "|" + xref.getAccessionId();
1604 // case-insensitive test, consistent with DBRefEntry.equalRef()
1605 if (xrefName.equalsIgnoreCase(name))
1615 * Constructs an alignment consisting of the mapped (CDS) regions in the given
1616 * nucleotide sequences, and updates mappings to match. The CDS sequences are
1617 * added to the original alignment's dataset, which is shared by the new
1618 * alignment. Mappings from nucleotide to CDS, and from CDS to protein, are
1619 * added to the alignment dataset.
1622 * aligned nucleotide (dna or cds) sequences
1624 * the alignment dataset the sequences belong to
1626 * (optional) to restrict results to CDS that map to specified
1628 * @return an alignment whose sequences are the cds-only parts of the dna
1629 * sequences (or null if no mappings are found)
1631 public static AlignmentI makeCdsAlignment(SequenceI[] dna,
1632 AlignmentI dataset, SequenceI[] products)
1634 if (dataset == null || dataset.getDataset() != null)
1636 throw new IllegalArgumentException(
1637 "IMPLEMENTATION ERROR: dataset.getDataset() must be null!");
1639 List<SequenceI> foundSeqs = new ArrayList<SequenceI>();
1640 List<SequenceI> cdsSeqs = new ArrayList<SequenceI>();
1641 List<AlignedCodonFrame> mappings = dataset.getCodonFrames();
1642 HashSet<SequenceI> productSeqs = null;
1643 if (products != null)
1645 productSeqs = new HashSet<SequenceI>();
1646 for (SequenceI seq : products)
1648 productSeqs.add(seq.getDatasetSequence() == null ? seq
1649 : seq.getDatasetSequence());
1654 * Construct CDS sequences from mappings on the alignment dataset.
1656 * - find the protein product(s) mapped to from each dna sequence
1657 * - if the mapping covers the whole dna sequence (give or take start/stop
1658 * codon), take the dna as the CDS sequence
1659 * - else search dataset mappings for a suitable dna sequence, i.e. one
1660 * whose whole sequence is mapped to the protein
1661 * - if no sequence found, construct one from the dna sequence and mapping
1662 * (and add it to dataset so it is found if this is repeated)
1664 for (SequenceI dnaSeq : dna)
1666 SequenceI dnaDss = dnaSeq.getDatasetSequence() == null ? dnaSeq
1667 : dnaSeq.getDatasetSequence();
1669 List<AlignedCodonFrame> seqMappings = MappingUtils
1670 .findMappingsForSequence(dnaSeq, mappings);
1671 for (AlignedCodonFrame mapping : seqMappings)
1673 List<Mapping> mappingsFromSequence = mapping
1674 .getMappingsFromSequence(dnaSeq);
1676 for (Mapping aMapping : mappingsFromSequence)
1678 MapList mapList = aMapping.getMap();
1679 if (mapList.getFromRatio() == 1)
1682 * not a dna-to-protein mapping (likely dna-to-cds)
1688 * skip if mapping is not to one of the target set of proteins
1690 SequenceI proteinProduct = aMapping.getTo();
1691 if (productSeqs != null && !productSeqs.contains(proteinProduct))
1697 * try to locate the CDS from the dataset mappings;
1698 * guard against duplicate results (for the case that protein has
1699 * dbrefs to both dna and cds sequences)
1701 SequenceI cdsSeq = findCdsForProtein(mappings, dnaSeq,
1702 seqMappings, aMapping);
1705 if (!foundSeqs.contains(cdsSeq))
1707 foundSeqs.add(cdsSeq);
1708 SequenceI derivedSequence = cdsSeq.deriveSequence();
1709 cdsSeqs.add(derivedSequence);
1710 if (!dataset.getSequences().contains(cdsSeq))
1712 dataset.addSequence(cdsSeq);
1719 * didn't find mapped CDS sequence - construct it and add
1720 * its dataset sequence to the dataset
1722 cdsSeq = makeCdsSequence(dnaSeq.getDatasetSequence(), aMapping,
1723 dataset).deriveSequence();
1724 // cdsSeq has a name constructed as CDS|<dbref>
1725 // <dbref> will be either the accession for the coding sequence,
1726 // marked in the /via/ dbref to the protein product accession
1727 // or it will be the original nucleotide accession.
1728 SequenceI cdsSeqDss = cdsSeq.getDatasetSequence();
1730 cdsSeqs.add(cdsSeq);
1732 if (!dataset.getSequences().contains(cdsSeqDss))
1734 // check if this sequence is a newly created one
1735 // so needs adding to the dataset
1736 dataset.addSequence(cdsSeqDss);
1740 * add a mapping from CDS to the (unchanged) mapped to range
1742 List<int[]> cdsRange = Collections
1743 .singletonList(new int[]
1744 { 1, cdsSeq.getLength() });
1745 MapList cdsToProteinMap = new MapList(cdsRange,
1746 mapList.getToRanges(), mapList.getFromRatio(),
1747 mapList.getToRatio());
1748 AlignedCodonFrame cdsToProteinMapping = new AlignedCodonFrame();
1749 cdsToProteinMapping.addMap(cdsSeqDss, proteinProduct,
1753 * guard against duplicating the mapping if repeating this action
1755 if (!mappings.contains(cdsToProteinMapping))
1757 mappings.add(cdsToProteinMapping);
1760 propagateDBRefsToCDS(cdsSeqDss, dnaSeq.getDatasetSequence(),
1761 proteinProduct, aMapping);
1763 * add another mapping from original 'from' range to CDS
1765 AlignedCodonFrame dnaToCdsMapping = new AlignedCodonFrame();
1766 MapList dnaToCdsMap = new MapList(mapList.getFromRanges(),
1768 dnaToCdsMapping.addMap(dnaSeq.getDatasetSequence(), cdsSeqDss,
1770 if (!mappings.contains(dnaToCdsMapping))
1772 mappings.add(dnaToCdsMapping);
1776 * add DBRef with mapping from protein to CDS
1777 * (this enables Get Cross-References from protein alignment)
1778 * This is tricky because we can't have two DBRefs with the
1779 * same source and accession, so need a different accession for
1780 * the CDS from the dna sequence
1783 // specific use case:
1784 // Genomic contig ENSCHR:1, contains coding regions for ENSG01,
1785 // ENSG02, ENSG03, with transcripts and products similarly named.
1786 // cannot add distinct dbrefs mapping location on ENSCHR:1 to ENSG01
1788 // JBPNote: ?? can't actually create an example that demonstrates we
1790 // synthesize an xref.
1792 for (DBRefEntry primRef : dnaDss.getPrimaryDBRefs())
1794 // creates a complementary cross-reference to the source sequence's
1795 // primary reference.
1797 DBRefEntry cdsCrossRef = new DBRefEntry(primRef.getSource(),
1798 primRef.getSource() + ":" + primRef.getVersion(),
1799 primRef.getAccessionId());
1801 .setMap(new Mapping(dnaDss, new MapList(dnaToCdsMap)));
1802 cdsSeqDss.addDBRef(cdsCrossRef);
1804 // problem here is that the cross-reference is synthesized -
1805 // cdsSeq.getName() may be like 'CDS|dnaaccession' or
1807 // assuming cds version same as dna ?!?
1809 DBRefEntry proteinToCdsRef = new DBRefEntry(primRef.getSource(),
1810 primRef.getVersion(), cdsSeq.getName());
1812 proteinToCdsRef.setMap(
1813 new Mapping(cdsSeqDss, cdsToProteinMap.getInverse()));
1814 proteinProduct.addDBRef(proteinToCdsRef);
1818 * transfer any features on dna that overlap the CDS
1820 transferFeatures(dnaSeq, cdsSeq, dnaToCdsMap, null,
1821 SequenceOntologyI.CDS);
1826 AlignmentI cds = new Alignment(
1827 cdsSeqs.toArray(new SequenceI[cdsSeqs.size()]));
1828 cds.setDataset(dataset);
1834 * A helper method that finds a CDS sequence in the alignment dataset that is
1835 * mapped to the given protein sequence, and either is, or has a mapping from,
1836 * the given dna sequence.
1839 * set of all mappings on the dataset
1841 * a dna (or cds) sequence we are searching from
1842 * @param seqMappings
1843 * the set of mappings involving dnaSeq
1845 * an initial candidate from seqMappings
1848 static SequenceI findCdsForProtein(List<AlignedCodonFrame> mappings,
1849 SequenceI dnaSeq, List<AlignedCodonFrame> seqMappings,
1853 * TODO a better dna-cds-protein mapping data representation to allow easy
1854 * navigation; until then this clunky looping around lists of mappings
1856 SequenceI seqDss = dnaSeq.getDatasetSequence() == null ? dnaSeq
1857 : dnaSeq.getDatasetSequence();
1858 SequenceI proteinProduct = aMapping.getTo();
1861 * is this mapping from the whole dna sequence (i.e. CDS)?
1862 * allowing for possible stop codon on dna but not peptide
1864 int mappedFromLength = MappingUtils
1865 .getLength(aMapping.getMap().getFromRanges());
1866 int dnaLength = seqDss.getLength();
1867 if (mappedFromLength == dnaLength
1868 || mappedFromLength == dnaLength - CODON_LENGTH)
1874 * looks like we found the dna-to-protein mapping; search for the
1875 * corresponding cds-to-protein mapping
1877 List<AlignedCodonFrame> mappingsToPeptide = MappingUtils
1878 .findMappingsForSequence(proteinProduct, mappings);
1879 for (AlignedCodonFrame acf : mappingsToPeptide)
1881 for (SequenceToSequenceMapping map : acf.getMappings())
1883 Mapping mapping = map.getMapping();
1884 if (mapping != aMapping
1885 && mapping.getMap().getFromRatio() == CODON_LENGTH
1886 && proteinProduct == mapping.getTo()
1887 && seqDss != map.getFromSeq())
1889 mappedFromLength = MappingUtils
1890 .getLength(mapping.getMap().getFromRanges());
1891 if (mappedFromLength == map.getFromSeq().getLength())
1894 * found a 3:1 mapping to the protein product which covers
1895 * the whole dna sequence i.e. is from CDS; finally check it
1896 * is from the dna start sequence
1898 SequenceI cdsSeq = map.getFromSeq();
1899 List<AlignedCodonFrame> dnaToCdsMaps = MappingUtils
1900 .findMappingsForSequence(cdsSeq, seqMappings);
1901 if (!dnaToCdsMaps.isEmpty())
1913 * Helper method that makes a CDS sequence as defined by the mappings from the
1914 * given sequence i.e. extracts the 'mapped from' ranges (which may be on
1915 * forward or reverse strand).
1920 * - existing dataset. We check for sequences that look like the CDS
1921 * we are about to construct, if one exists already, then we will
1922 * just return that one.
1923 * @return CDS sequence (as a dataset sequence)
1925 static SequenceI makeCdsSequence(SequenceI seq, Mapping mapping,
1928 char[] seqChars = seq.getSequence();
1929 List<int[]> fromRanges = mapping.getMap().getFromRanges();
1930 int cdsWidth = MappingUtils.getLength(fromRanges);
1931 char[] newSeqChars = new char[cdsWidth];
1934 for (int[] range : fromRanges)
1936 if (range[0] <= range[1])
1938 // forward strand mapping - just copy the range
1939 int length = range[1] - range[0] + 1;
1940 System.arraycopy(seqChars, range[0] - 1, newSeqChars, newPos,
1946 // reverse strand mapping - copy and complement one by one
1947 for (int i = range[0]; i >= range[1]; i--)
1949 newSeqChars[newPos++] = Dna.getComplement(seqChars[i - 1]);
1955 * assign 'from id' held in the mapping if set (e.g. EMBL protein_id),
1956 * else generate a sequence name
1958 String mapFromId = mapping.getMappedFromId();
1959 String seqId = "CDS|" + (mapFromId != null ? mapFromId : seq.getName());
1960 SequenceI newSeq = new Sequence(seqId, newSeqChars, 1, newPos);
1961 if (dataset != null)
1963 SequenceI[] matches = dataset.findSequenceMatch(newSeq.getName());
1964 if (matches != null)
1966 boolean matched = false;
1967 for (SequenceI mtch : matches)
1969 if (mtch.getStart() != newSeq.getStart())
1973 if (mtch.getEnd() != newSeq.getEnd())
1977 if (!Arrays.equals(mtch.getSequence(), newSeq.getSequence()))
1989 "JAL-2154 regression: warning - found (and ignnored a duplicate CDS sequence):"
1995 // newSeq.setDescription(mapFromId);
2001 * add any DBRefEntrys to cdsSeq from contig that have a Mapping congruent to
2002 * the given mapping.
2007 * @return list of DBRefEntrys added.
2009 public static List<DBRefEntry> propagateDBRefsToCDS(SequenceI cdsSeq,
2010 SequenceI contig, SequenceI proteinProduct, Mapping mapping)
2013 // gather direct refs from contig congrent with mapping
2014 List<DBRefEntry> direct = new ArrayList<DBRefEntry>();
2015 HashSet<String> directSources = new HashSet<String>();
2016 if (contig.getDBRefs() != null)
2018 for (DBRefEntry dbr : contig.getDBRefs())
2020 if (dbr.hasMap() && dbr.getMap().getMap().isTripletMap())
2022 MapList map = dbr.getMap().getMap();
2023 // check if map is the CDS mapping
2024 if (mapping.getMap().equals(map))
2027 directSources.add(dbr.getSource());
2032 DBRefEntry[] onSource = DBRefUtils.selectRefs(
2033 proteinProduct.getDBRefs(),
2034 directSources.toArray(new String[0]));
2035 List<DBRefEntry> propagated = new ArrayList<DBRefEntry>();
2037 // and generate appropriate mappings
2038 for (DBRefEntry cdsref : direct)
2040 // clone maplist and mapping
2041 MapList cdsposmap = new MapList(
2042 Arrays.asList(new int[][]
2043 { new int[] { cdsSeq.getStart(), cdsSeq.getEnd() } }),
2044 cdsref.getMap().getMap().getToRanges(), 3, 1);
2045 Mapping cdsmap = new Mapping(cdsref.getMap().getTo(),
2046 cdsref.getMap().getMap());
2049 DBRefEntry newref = new DBRefEntry(cdsref.getSource(),
2050 cdsref.getVersion(), cdsref.getAccessionId(),
2051 new Mapping(cdsmap.getTo(), cdsposmap));
2053 // and see if we can map to the protein product for this mapping.
2054 // onSource is the filtered set of accessions on protein that we are
2055 // tranferring, so we assume accession is the same.
2056 if (cdsmap.getTo() == null && onSource != null)
2058 List<DBRefEntry> sourceRefs = DBRefUtils.searchRefs(onSource,
2059 cdsref.getAccessionId());
2060 if (sourceRefs != null)
2062 for (DBRefEntry srcref : sourceRefs)
2064 if (srcref.getSource().equalsIgnoreCase(cdsref.getSource()))
2066 // we have found a complementary dbref on the protein product, so
2067 // update mapping's getTo
2068 newref.getMap().setTo(proteinProduct);
2073 cdsSeq.addDBRef(newref);
2074 propagated.add(newref);
2080 * Transfers co-located features on 'fromSeq' to 'toSeq', adjusting the
2081 * feature start/end ranges, optionally omitting specified feature types.
2082 * Returns the number of features copied.
2087 * the mapping from 'fromSeq' to 'toSeq'
2089 * if not null, only features of this type are copied (including
2090 * subtypes in the Sequence Ontology)
2093 public static int transferFeatures(SequenceI fromSeq, SequenceI toSeq,
2094 MapList mapping, String select, String... omitting)
2096 SequenceI copyTo = toSeq;
2097 while (copyTo.getDatasetSequence() != null)
2099 copyTo = copyTo.getDatasetSequence();
2103 * get features, optionally restricted by an ontology term
2105 List<SequenceFeature> sfs = select == null ? fromSeq.getFeatures()
2106 .getPositionalFeatures() : fromSeq.getFeatures()
2107 .getFeaturesByOntology(select);
2110 for (SequenceFeature sf : sfs)
2112 String type = sf.getType();
2113 boolean omit = false;
2114 for (String toOmit : omitting)
2116 if (type.equals(toOmit))
2127 * locate the mapped range - null if either start or end is
2128 * not mapped (no partial overlaps are calculated)
2130 int start = sf.getBegin();
2131 int end = sf.getEnd();
2132 int[] mappedTo = mapping.locateInTo(start, end);
2134 * if whole exon range doesn't map, try interpreting it
2135 * as 5' or 3' exon overlapping the CDS range
2137 if (mappedTo == null)
2139 mappedTo = mapping.locateInTo(end, end);
2140 if (mappedTo != null)
2143 * end of exon is in CDS range - 5' overlap
2144 * to a range from the start of the peptide
2149 if (mappedTo == null)
2151 mappedTo = mapping.locateInTo(start, start);
2152 if (mappedTo != null)
2155 * start of exon is in CDS range - 3' overlap
2156 * to a range up to the end of the peptide
2158 mappedTo[1] = toSeq.getLength();
2161 if (mappedTo != null)
2163 int newBegin = Math.min(mappedTo[0], mappedTo[1]);
2164 int newEnd = Math.max(mappedTo[0], mappedTo[1]);
2165 SequenceFeature copy = new SequenceFeature(sf, newBegin, newEnd,
2166 sf.getFeatureGroup(), sf.getScore());
2167 copyTo.addSequenceFeature(copy);
2175 * Returns a mapping from dna to protein by inspecting sequence features of
2176 * type "CDS" on the dna.
2182 public static MapList mapCdsToProtein(SequenceI dnaSeq,
2183 SequenceI proteinSeq)
2185 List<int[]> ranges = findCdsPositions(dnaSeq);
2186 int mappedDnaLength = MappingUtils.getLength(ranges);
2188 int proteinLength = proteinSeq.getLength();
2189 int proteinStart = proteinSeq.getStart();
2190 int proteinEnd = proteinSeq.getEnd();
2193 * incomplete start codon may mean X at start of peptide
2194 * we ignore both for mapping purposes
2196 if (proteinSeq.getCharAt(0) == 'X')
2198 // todo JAL-2022 support startPhase > 0
2202 List<int[]> proteinRange = new ArrayList<int[]>();
2205 * dna length should map to protein (or protein plus stop codon)
2207 int codesForResidues = mappedDnaLength / CODON_LENGTH;
2208 if (codesForResidues == (proteinLength + 1))
2210 // assuming extra codon is for STOP and not in peptide
2213 if (codesForResidues == proteinLength)
2215 proteinRange.add(new int[] { proteinStart, proteinEnd });
2216 return new MapList(ranges, proteinRange, CODON_LENGTH, 1);
2222 * Returns a list of CDS ranges found (as sequence positions base 1), i.e. of
2223 * start/end positions of sequence features of type "CDS" (or a sub-type of
2224 * CDS in the Sequence Ontology). The ranges are sorted into ascending start
2225 * position order, so this method is only valid for linear CDS in the same
2226 * sense as the protein product.
2231 public static List<int[]> findCdsPositions(SequenceI dnaSeq)
2233 List<int[]> result = new ArrayList<int[]>();
2235 List<SequenceFeature> sfs = dnaSeq.getFeatures().getFeaturesByOntology(
2236 SequenceOntologyI.CDS);
2241 SequenceFeatures.sortFeatures(sfs, true);
2244 for (SequenceFeature sf : sfs)
2249 phase = Integer.parseInt(sf.getPhase());
2250 } catch (NumberFormatException e)
2255 * phase > 0 on first codon means 5' incomplete - skip to the start
2256 * of the next codon; example ENST00000496384
2258 int begin = sf.getBegin();
2259 int end = sf.getEnd();
2260 if (result.isEmpty())
2265 // shouldn't happen!
2267 .println("Error: start phase extends beyond start CDS in "
2268 + dnaSeq.getName());
2271 result.add(new int[] { begin, end });
2275 * remove 'startPhase' positions (usually 0) from the first range
2276 * so we begin at the start of a complete codon
2278 if (!result.isEmpty())
2280 // TODO JAL-2022 correctly model start phase > 0
2281 result.get(0)[0] += startPhase;
2285 * Finally sort ranges by start position. This avoids a dependency on
2286 * keeping features in order on the sequence (if they are in order anyway,
2287 * the sort will have almost no work to do). The implicit assumption is CDS
2288 * ranges are assembled in order. Other cases should not use this method,
2289 * but instead construct an explicit mapping for CDS (e.g. EMBL parsing).
2291 Collections.sort(result, IntRangeComparator.ASCENDING);
2296 * Maps exon features from dna to protein, and computes variants in peptide
2297 * product generated by variants in dna, and adds them as sequence_variant
2298 * features on the protein sequence. Returns the number of variant features
2303 * @param dnaToProtein
2305 public static int computeProteinFeatures(SequenceI dnaSeq,
2306 SequenceI peptide, MapList dnaToProtein)
2308 while (dnaSeq.getDatasetSequence() != null)
2310 dnaSeq = dnaSeq.getDatasetSequence();
2312 while (peptide.getDatasetSequence() != null)
2314 peptide = peptide.getDatasetSequence();
2317 transferFeatures(dnaSeq, peptide, dnaToProtein, SequenceOntologyI.EXON);
2320 * compute protein variants from dna variants and codon mappings;
2321 * NB - alternatively we could retrieve this using the REST service e.g.
2322 * http://rest.ensembl.org/overlap/translation
2323 * /ENSP00000288602?feature=transcript_variation;content-type=text/xml
2324 * which would be a bit slower but possibly more reliable
2328 * build a map with codon variations for each potentially varying peptide
2330 LinkedHashMap<Integer, List<DnaVariant>[]> variants = buildDnaVariantsMap(
2331 dnaSeq, dnaToProtein);
2334 * scan codon variations, compute peptide variants and add to peptide sequence
2337 for (Entry<Integer, List<DnaVariant>[]> variant : variants.entrySet())
2339 int peptidePos = variant.getKey();
2340 List<DnaVariant>[] codonVariants = variant.getValue();
2341 count += computePeptideVariants(peptide, peptidePos, codonVariants);
2348 * Computes non-synonymous peptide variants from codon variants and adds them
2349 * as sequence_variant features on the protein sequence (one feature per
2350 * allele variant). Selected attributes (variant id, clinical significance)
2351 * are copied over to the new features.
2354 * the protein sequence
2356 * the position to compute peptide variants for
2357 * @param codonVariants
2358 * a list of dna variants per codon position
2359 * @return the number of features added
2361 static int computePeptideVariants(SequenceI peptide, int peptidePos,
2362 List<DnaVariant>[] codonVariants)
2364 String residue = String.valueOf(peptide.getCharAt(peptidePos - 1));
2366 String base1 = codonVariants[0].get(0).base;
2367 String base2 = codonVariants[1].get(0).base;
2368 String base3 = codonVariants[2].get(0).base;
2371 * variants in first codon base
2373 for (DnaVariant var : codonVariants[0])
2375 if (var.variant != null)
2377 String alleles = (String) var.variant.getValue("alleles");
2378 if (alleles != null)
2380 for (String base : alleles.split(","))
2382 String codon = base + base2 + base3;
2383 if (addPeptideVariant(peptide, peptidePos, residue, var, codon))
2393 * variants in second codon base
2395 for (DnaVariant var : codonVariants[1])
2397 if (var.variant != null)
2399 String alleles = (String) var.variant.getValue("alleles");
2400 if (alleles != null)
2402 for (String base : alleles.split(","))
2404 String codon = base1 + base + base3;
2405 if (addPeptideVariant(peptide, peptidePos, residue, var, codon))
2415 * variants in third codon base
2417 for (DnaVariant var : codonVariants[2])
2419 if (var.variant != null)
2421 String alleles = (String) var.variant.getValue("alleles");
2422 if (alleles != null)
2424 for (String base : alleles.split(","))
2426 String codon = base1 + base2 + base;
2427 if (addPeptideVariant(peptide, peptidePos, residue, var, codon))
2440 * Helper method that adds a peptide variant feature, provided the given codon
2441 * translates to a value different to the current residue (is a non-synonymous
2442 * variant). ID and clinical_significance attributes of the dna variant (if
2443 * present) are copied to the new feature.
2450 * @return true if a feature was added, else false
2452 static boolean addPeptideVariant(SequenceI peptide, int peptidePos,
2453 String residue, DnaVariant var, String codon)
2456 * get peptide translation of codon e.g. GAT -> D
2457 * note that variants which are not single alleles,
2458 * e.g. multibase variants or HGMD_MUTATION etc
2459 * are currently ignored here
2461 String trans = codon.contains("-") ? "-"
2462 : (codon.length() > CODON_LENGTH ? null
2463 : ResidueProperties.codonTranslate(codon));
2464 if (trans != null && !trans.equals(residue))
2466 String residue3Char = StringUtils
2467 .toSentenceCase(ResidueProperties.aa2Triplet.get(residue));
2468 String trans3Char = StringUtils
2469 .toSentenceCase(ResidueProperties.aa2Triplet.get(trans));
2470 String desc = "p." + residue3Char + peptidePos + trans3Char;
2471 SequenceFeature sf = new SequenceFeature(
2472 SequenceOntologyI.SEQUENCE_VARIANT, desc, peptidePos,
2473 peptidePos, var.getSource());
2474 StringBuilder attributes = new StringBuilder(32);
2475 String id = (String) var.variant.getValue(ID);
2478 if (id.startsWith(SEQUENCE_VARIANT))
2480 id = id.substring(SEQUENCE_VARIANT.length());
2482 sf.setValue(ID, id);
2483 attributes.append(ID).append("=").append(id);
2484 // TODO handle other species variants JAL-2064
2485 StringBuilder link = new StringBuilder(32);
2488 link.append(desc).append(" ").append(id).append(
2489 "|http://www.ensembl.org/Homo_sapiens/Variation/Summary?v=")
2490 .append(URLEncoder.encode(id, "UTF-8"));
2491 sf.addLink(link.toString());
2492 } catch (UnsupportedEncodingException e)
2497 String clinSig = (String) var.variant.getValue(CLINICAL_SIGNIFICANCE);
2498 if (clinSig != null)
2500 sf.setValue(CLINICAL_SIGNIFICANCE, clinSig);
2501 attributes.append(";").append(CLINICAL_SIGNIFICANCE).append("=")
2504 peptide.addSequenceFeature(sf);
2505 if (attributes.length() > 0)
2507 sf.setAttributes(attributes.toString());
2515 * Builds a map whose key is position in the protein sequence, and value is a
2516 * list of the base and all variants for each corresponding codon position
2519 * @param dnaToProtein
2522 @SuppressWarnings("unchecked")
2523 static LinkedHashMap<Integer, List<DnaVariant>[]> buildDnaVariantsMap(
2524 SequenceI dnaSeq, MapList dnaToProtein)
2527 * map from peptide position to all variants of the codon which codes for it
2528 * LinkedHashMap ensures we keep the peptide features in sequence order
2530 LinkedHashMap<Integer, List<DnaVariant>[]> variants = new LinkedHashMap<Integer, List<DnaVariant>[]>();
2532 List<SequenceFeature> dnaFeatures = dnaSeq.getFeatures()
2533 .getFeaturesByOntology(SequenceOntologyI.SEQUENCE_VARIANT);
2534 if (dnaFeatures.isEmpty())
2539 int dnaStart = dnaSeq.getStart();
2540 int[] lastCodon = null;
2541 int lastPeptidePostion = 0;
2544 * build a map of codon variations for peptides
2546 for (SequenceFeature sf : dnaFeatures)
2548 int dnaCol = sf.getBegin();
2549 if (dnaCol != sf.getEnd())
2551 // not handling multi-locus variant features
2556 * extract dna variants to a string array
2558 String alls = (String) sf.getValue("alleles");
2561 continue; // non-SNP VCF variant perhaps - can't process this
2563 String[] alleles = alls.toUpperCase().split(",");
2565 for (String allele : alleles)
2567 alleles[i++] = allele.trim(); // lose any space characters "A, G"
2570 int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol);
2573 // feature doesn't lie within coding region
2576 int peptidePosition = mapsTo[0];
2577 List<DnaVariant>[] codonVariants = variants.get(peptidePosition);
2578 if (codonVariants == null)
2580 codonVariants = new ArrayList[CODON_LENGTH];
2581 codonVariants[0] = new ArrayList<DnaVariant>();
2582 codonVariants[1] = new ArrayList<DnaVariant>();
2583 codonVariants[2] = new ArrayList<DnaVariant>();
2584 variants.put(peptidePosition, codonVariants);
2588 * get this peptide's codon positions e.g. [3, 4, 5] or [4, 7, 10]
2590 int[] codon = peptidePosition == lastPeptidePostion ? lastCodon
2591 : MappingUtils.flattenRanges(dnaToProtein.locateInFrom(
2592 peptidePosition, peptidePosition));
2593 lastPeptidePostion = peptidePosition;
2597 * save nucleotide (and any variant) for each codon position
2599 for (int codonPos = 0; codonPos < CODON_LENGTH; codonPos++)
2601 String nucleotide = String.valueOf(
2602 dnaSeq.getCharAt(codon[codonPos] - dnaStart)).toUpperCase();
2603 List<DnaVariant> codonVariant = codonVariants[codonPos];
2604 if (codon[codonPos] == dnaCol)
2606 if (!codonVariant.isEmpty()
2607 && codonVariant.get(0).variant == null)
2610 * already recorded base value, add this variant
2612 codonVariant.get(0).variant = sf;
2617 * add variant with base value
2619 codonVariant.add(new DnaVariant(nucleotide, sf));
2622 else if (codonVariant.isEmpty())
2625 * record (possibly non-varying) base value
2627 codonVariant.add(new DnaVariant(nucleotide));
2635 * Makes an alignment with a copy of the given sequences, adding in any
2636 * non-redundant sequences which are mapped to by the cross-referenced
2642 * the alignment dataset shared by the new copy
2645 public static AlignmentI makeCopyAlignment(SequenceI[] seqs,
2646 SequenceI[] xrefs, AlignmentI dataset)
2648 AlignmentI copy = new Alignment(new Alignment(seqs));
2649 copy.setDataset(dataset);
2650 boolean isProtein = !copy.isNucleotide();
2651 SequenceIdMatcher matcher = new SequenceIdMatcher(seqs);
2654 for (SequenceI xref : xrefs)
2656 DBRefEntry[] dbrefs = xref.getDBRefs();
2659 for (DBRefEntry dbref : dbrefs)
2661 if (dbref.getMap() == null || dbref.getMap().getTo() == null
2662 || dbref.getMap().getTo().isProtein() != isProtein)
2666 SequenceI mappedTo = dbref.getMap().getTo();
2667 SequenceI match = matcher.findIdMatch(mappedTo);
2670 matcher.add(mappedTo);
2671 copy.addSequence(mappedTo);
2681 * Try to align sequences in 'unaligned' to match the alignment of their
2682 * mapped regions in 'aligned'. For example, could use this to align CDS
2683 * sequences which are mapped to their parent cDNA sequences.
2685 * This method handles 1:1 mappings (dna-to-dna or protein-to-protein). For
2686 * dna-to-protein or protein-to-dna use alternative methods.
2689 * sequences to be aligned
2691 * holds aligned sequences and their mappings
2694 public static int alignAs(AlignmentI unaligned, AlignmentI aligned)
2697 * easy case - aligning a copy of aligned sequences
2699 if (alignAsSameSequences(unaligned, aligned))
2701 return unaligned.getHeight();
2705 * fancy case - aligning via mappings between sequences
2707 List<SequenceI> unmapped = new ArrayList<SequenceI>();
2708 Map<Integer, Map<SequenceI, Character>> columnMap = buildMappedColumnsMap(
2709 unaligned, aligned, unmapped);
2710 int width = columnMap.size();
2711 char gap = unaligned.getGapCharacter();
2712 int realignedCount = 0;
2713 // TODO: verify this loop scales sensibly for very wide/high alignments
2715 for (SequenceI seq : unaligned.getSequences())
2717 if (!unmapped.contains(seq))
2719 char[] newSeq = new char[width];
2720 Arrays.fill(newSeq, gap); // JBPComment - doubt this is faster than the
2721 // Integer iteration below
2726 * traverse the map to find columns populated
2729 for (Integer column : columnMap.keySet())
2731 Character c = columnMap.get(column).get(seq);
2735 * sequence has a character at this position
2745 * trim trailing gaps
2747 if (lastCol < width)
2749 char[] tmp = new char[lastCol + 1];
2750 System.arraycopy(newSeq, 0, tmp, 0, lastCol + 1);
2753 // TODO: optimise SequenceI to avoid char[]->String->char[]
2754 seq.setSequence(String.valueOf(newSeq));
2758 return realignedCount;
2762 * If unaligned and aligned sequences share the same dataset sequences, then
2763 * simply copies the aligned sequences to the unaligned sequences and returns
2764 * true; else returns false
2767 * - sequences to be aligned based on aligned
2769 * - 'guide' alignment containing sequences derived from same dataset
2773 static boolean alignAsSameSequences(AlignmentI unaligned,
2776 if (aligned.getDataset() == null || unaligned.getDataset() == null)
2778 return false; // should only pass alignments with datasets here
2781 // map from dataset sequence to alignment sequence(s)
2782 Map<SequenceI, List<SequenceI>> alignedDatasets = new HashMap<SequenceI, List<SequenceI>>();
2783 for (SequenceI seq : aligned.getSequences())
2785 SequenceI ds = seq.getDatasetSequence();
2786 if (alignedDatasets.get(ds) == null)
2788 alignedDatasets.put(ds, new ArrayList<SequenceI>());
2790 alignedDatasets.get(ds).add(seq);
2794 * first pass - check whether all sequences to be aligned share a dataset
2795 * sequence with an aligned sequence
2797 for (SequenceI seq : unaligned.getSequences())
2799 if (!alignedDatasets.containsKey(seq.getDatasetSequence()))
2806 * second pass - copy aligned sequences;
2807 * heuristic rule: pair off sequences in order for the case where
2808 * more than one shares the same dataset sequence
2810 for (SequenceI seq : unaligned.getSequences())
2812 List<SequenceI> alignedSequences = alignedDatasets
2813 .get(seq.getDatasetSequence());
2814 // TODO: getSequenceAsString() will be deprecated in the future
2815 // TODO: need to leave to SequenceI implementor to update gaps
2816 seq.setSequence(alignedSequences.get(0).getSequenceAsString());
2817 if (alignedSequences.size() > 0)
2819 // pop off aligned sequences (except the last one)
2820 alignedSequences.remove(0);
2828 * Returns a map whose key is alignment column number (base 1), and whose
2829 * values are a map of sequence characters in that column.
2836 static SortedMap<Integer, Map<SequenceI, Character>> buildMappedColumnsMap(
2837 AlignmentI unaligned, AlignmentI aligned,
2838 List<SequenceI> unmapped)
2841 * Map will hold, for each aligned column position, a map of
2842 * {unalignedSequence, characterPerSequence} at that position.
2843 * TreeMap keeps the entries in ascending column order.
2845 SortedMap<Integer, Map<SequenceI, Character>> map = new TreeMap<Integer, Map<SequenceI, Character>>();
2848 * record any sequences that have no mapping so can't be realigned
2850 unmapped.addAll(unaligned.getSequences());
2852 List<AlignedCodonFrame> mappings = aligned.getCodonFrames();
2854 for (SequenceI seq : unaligned.getSequences())
2856 for (AlignedCodonFrame mapping : mappings)
2858 SequenceI fromSeq = mapping.findAlignedSequence(seq, aligned);
2859 if (fromSeq != null)
2861 Mapping seqMap = mapping.getMappingBetween(fromSeq, seq);
2862 if (addMappedPositions(seq, fromSeq, seqMap, map))
2864 unmapped.remove(seq);
2873 * Helper method that adds to a map the mapped column positions of a sequence.
2875 * For example if aaTT-Tg-gAAA is mapped to TTTAAA then the map should record
2876 * that columns 3,4,6,10,11,12 map to characters T,T,T,A,A,A of the mapped to
2880 * the sequence whose column positions we are recording
2882 * a sequence that is mapped to the first sequence
2884 * the mapping from 'fromSeq' to 'seq'
2886 * a map to add the column positions (in fromSeq) of the mapped
2890 static boolean addMappedPositions(SequenceI seq, SequenceI fromSeq,
2891 Mapping seqMap, Map<Integer, Map<SequenceI, Character>> map)
2899 * invert mapping if it is from unaligned to aligned sequence
2901 if (seqMap.getTo() == fromSeq.getDatasetSequence())
2903 seqMap = new Mapping(seq.getDatasetSequence(),
2904 seqMap.getMap().getInverse());
2907 int toStart = seq.getStart();
2910 * traverse [start, end, start, end...] ranges in fromSeq
2912 for (int[] fromRange : seqMap.getMap().getFromRanges())
2914 for (int i = 0; i < fromRange.length - 1; i += 2)
2916 boolean forward = fromRange[i + 1] >= fromRange[i];
2919 * find the range mapped to (sequence positions base 1)
2921 int[] range = seqMap.locateMappedRange(fromRange[i],
2925 System.err.println("Error in mapping " + seqMap + " from "
2926 + fromSeq.getName());
2929 int fromCol = fromSeq.findIndex(fromRange[i]);
2930 int mappedCharPos = range[0];
2933 * walk over the 'from' aligned sequence in forward or reverse
2934 * direction; when a non-gap is found, record the column position
2935 * of the next character of the mapped-to sequence; stop when all
2936 * the characters of the range have been counted
2938 while (mappedCharPos <= range[1] && fromCol <= fromSeq.getLength()
2941 if (!Comparison.isGap(fromSeq.getCharAt(fromCol - 1)))
2944 * mapped from sequence has a character in this column
2945 * record the column position for the mapped to character
2947 Map<SequenceI, Character> seqsMap = map.get(fromCol);
2948 if (seqsMap == null)
2950 seqsMap = new HashMap<SequenceI, Character>();
2951 map.put(fromCol, seqsMap);
2953 seqsMap.put(seq, seq.getCharAt(mappedCharPos - toStart));
2956 fromCol += (forward ? 1 : -1);
2963 // strictly temporary hack until proper criteria for aligning protein to cds
2964 // are in place; this is so Ensembl -> fetch xrefs Uniprot aligns the Uniprot
2965 public static boolean looksLikeEnsembl(AlignmentI alignment)
2967 for (SequenceI seq : alignment.getSequences())
2969 String name = seq.getName();
2970 if (!name.startsWith("ENSG") && !name.startsWith("ENST"))