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.api.DBRefEntryI;
26 import jalview.datamodel.AlignedCodon;
27 import jalview.datamodel.AlignedCodonFrame;
28 import jalview.datamodel.AlignedCodonFrame.SequenceToSequenceMapping;
29 import jalview.datamodel.Alignment;
30 import jalview.datamodel.AlignmentAnnotation;
31 import jalview.datamodel.AlignmentI;
32 import jalview.datamodel.DBRefEntry;
33 import jalview.datamodel.IncompleteCodonException;
34 import jalview.datamodel.Mapping;
35 import jalview.datamodel.Sequence;
36 import jalview.datamodel.SequenceFeature;
37 import jalview.datamodel.SequenceGroup;
38 import jalview.datamodel.SequenceI;
39 import jalview.io.gff.SequenceOntologyFactory;
40 import jalview.io.gff.SequenceOntologyI;
41 import jalview.schemes.ResidueProperties;
42 import jalview.util.Comparison;
43 import jalview.util.DBRefUtils;
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.Comparator;
55 import java.util.HashMap;
56 import java.util.HashSet;
57 import java.util.Iterator;
58 import java.util.LinkedHashMap;
59 import java.util.List;
61 import java.util.Map.Entry;
62 import java.util.NoSuchElementException;
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;
77 private static final String SEQUENCE_VARIANT = "sequence_variant:";
78 private static final String ID = "ID";
81 * A data model to hold the 'normal' base value at a position, and an optional
82 * sequence variant feature
84 static final class DnaVariant
88 SequenceFeature variant;
90 DnaVariant(String nuc)
96 DnaVariant(String nuc, SequenceFeature var)
102 public String getSource()
104 return variant == null ? null : variant.getFeatureGroup();
109 * given an existing alignment, create a new alignment including all, or up to
110 * flankSize additional symbols from each sequence's dataset sequence
116 public static AlignmentI expandContext(AlignmentI core, int flankSize)
118 List<SequenceI> sq = new ArrayList<SequenceI>();
120 for (SequenceI s : core.getSequences())
122 SequenceI newSeq = s.deriveSequence();
123 final int newSeqStart = newSeq.getStart() - 1;
124 if (newSeqStart > maxoffset
125 && newSeq.getDatasetSequence().getStart() < s.getStart())
127 maxoffset = newSeqStart;
133 maxoffset = Math.min(maxoffset, flankSize);
137 * now add offset left and right to create an expanded alignment
139 for (SequenceI s : sq)
142 while (ds.getDatasetSequence() != null)
144 ds = ds.getDatasetSequence();
146 int s_end = s.findPosition(s.getStart() + s.getLength());
147 // find available flanking residues for sequence
148 int ustream_ds = s.getStart() - ds.getStart();
149 int dstream_ds = ds.getEnd() - s_end;
151 // build new flanked sequence
153 // compute gap padding to start of flanking sequence
154 int offset = maxoffset - ustream_ds;
156 // padding is gapChar x ( maxoffset - min(ustream_ds, flank)
159 if (flankSize < ustream_ds)
161 // take up to flankSize residues
162 offset = maxoffset - flankSize;
163 ustream_ds = flankSize;
165 if (flankSize <= dstream_ds)
167 dstream_ds = flankSize - 1;
170 // TODO use Character.toLowerCase to avoid creating String objects?
171 char[] upstream = new String(ds.getSequence(s.getStart() - 1
172 - ustream_ds, s.getStart() - 1)).toLowerCase().toCharArray();
173 char[] downstream = new String(ds.getSequence(s_end - 1, s_end
174 + dstream_ds)).toLowerCase().toCharArray();
175 char[] coreseq = s.getSequence();
176 char[] nseq = new char[offset + upstream.length + downstream.length
178 char c = core.getGapCharacter();
181 for (; p < offset; p++)
186 System.arraycopy(upstream, 0, nseq, p, upstream.length);
187 System.arraycopy(coreseq, 0, nseq, p + upstream.length,
189 System.arraycopy(downstream, 0, nseq, p + coreseq.length
190 + upstream.length, downstream.length);
191 s.setSequence(new String(nseq));
192 s.setStart(s.getStart() - ustream_ds);
193 s.setEnd(s_end + downstream.length);
195 AlignmentI newAl = new jalview.datamodel.Alignment(
196 sq.toArray(new SequenceI[0]));
197 for (SequenceI s : sq)
199 if (s.getAnnotation() != null)
201 for (AlignmentAnnotation aa : s.getAnnotation())
203 aa.adjustForAlignment(); // JAL-1712 fix
204 newAl.addAnnotation(aa);
208 newAl.setDataset(core.getDataset());
213 * Returns the index (zero-based position) of a sequence in an alignment, or
220 public static int getSequenceIndex(AlignmentI al, SequenceI seq)
224 for (SequenceI alSeq : al.getSequences())
237 * Returns a map of lists of sequences in the alignment, keyed by sequence
238 * name. For use in mapping between different alignment views of the same
241 * @see jalview.datamodel.AlignmentI#getSequencesByName()
243 public static Map<String, List<SequenceI>> getSequencesByName(
246 Map<String, List<SequenceI>> theMap = new LinkedHashMap<String, List<SequenceI>>();
247 for (SequenceI seq : al.getSequences())
249 String name = seq.getName();
252 List<SequenceI> seqs = theMap.get(name);
255 seqs = new ArrayList<SequenceI>();
256 theMap.put(name, seqs);
265 * Build mapping of protein to cDNA alignment. Mappings are made between
266 * sequences where the cDNA translates to the protein sequence. Any new
267 * mappings are added to the protein alignment. Returns true if any mappings
268 * either already exist or were added, else false.
270 * @param proteinAlignment
271 * @param cdnaAlignment
274 public static boolean mapProteinAlignmentToCdna(
275 final AlignmentI proteinAlignment, final AlignmentI cdnaAlignment)
277 if (proteinAlignment == null || cdnaAlignment == null)
282 Set<SequenceI> mappedDna = new HashSet<SequenceI>();
283 Set<SequenceI> mappedProtein = new HashSet<SequenceI>();
286 * First pass - map sequences where cross-references exist. This include
287 * 1-to-many mappings to support, for example, variant cDNA.
289 boolean mappingPerformed = mapProteinToCdna(proteinAlignment,
290 cdnaAlignment, mappedDna, mappedProtein, true);
293 * Second pass - map sequences where no cross-references exist. This only
294 * does 1-to-1 mappings and assumes corresponding sequences are in the same
295 * order in the alignments.
297 mappingPerformed |= mapProteinToCdna(proteinAlignment, cdnaAlignment,
298 mappedDna, mappedProtein, false);
299 return mappingPerformed;
303 * Make mappings between compatible sequences (where the cDNA translation
304 * matches the protein).
306 * @param proteinAlignment
307 * @param cdnaAlignment
309 * a set of mapped DNA sequences (to add to)
310 * @param mappedProtein
311 * a set of mapped Protein sequences (to add to)
313 * if true, only map sequences where xrefs exist
316 protected static boolean mapProteinToCdna(
317 final AlignmentI proteinAlignment,
318 final AlignmentI cdnaAlignment, Set<SequenceI> mappedDna,
319 Set<SequenceI> mappedProtein, boolean xrefsOnly)
321 boolean mappingExistsOrAdded = false;
322 List<SequenceI> thisSeqs = proteinAlignment.getSequences();
323 for (SequenceI aaSeq : thisSeqs)
325 boolean proteinMapped = false;
326 AlignedCodonFrame acf = new AlignedCodonFrame();
328 for (SequenceI cdnaSeq : cdnaAlignment.getSequences())
331 * Always try to map if sequences have xref to each other; this supports
332 * variant cDNA or alternative splicing for a protein sequence.
334 * If no xrefs, try to map progressively, assuming that alignments have
335 * mappable sequences in corresponding order. These are not
336 * many-to-many, as that would risk mixing species with similar cDNA
339 if (xrefsOnly && !AlignmentUtils.haveCrossRef(aaSeq, cdnaSeq))
345 * Don't map non-xrefd sequences more than once each. This heuristic
346 * allows us to pair up similar sequences in ordered alignments.
349 && (mappedProtein.contains(aaSeq) || mappedDna
354 if (mappingExists(proteinAlignment.getCodonFrames(),
355 aaSeq.getDatasetSequence(), cdnaSeq.getDatasetSequence()))
357 mappingExistsOrAdded = true;
361 MapList map = mapCdnaToProtein(aaSeq, cdnaSeq);
364 acf.addMap(cdnaSeq, aaSeq, map);
365 mappingExistsOrAdded = true;
366 proteinMapped = true;
367 mappedDna.add(cdnaSeq);
368 mappedProtein.add(aaSeq);
374 proteinAlignment.addCodonFrame(acf);
377 return mappingExistsOrAdded;
381 * Answers true if the mappings include one between the given (dataset)
384 public static boolean mappingExists(List<AlignedCodonFrame> mappings,
385 SequenceI aaSeq, SequenceI cdnaSeq)
387 if (mappings != null)
389 for (AlignedCodonFrame acf : mappings)
391 if (cdnaSeq == acf.getDnaForAaSeq(aaSeq))
401 * Builds a mapping (if possible) of a cDNA to a protein sequence.
403 * <li>first checks if the cdna translates exactly to the protein sequence</li>
404 * <li>else checks for translation after removing a STOP codon</li>
405 * <li>else checks for translation after removing a START codon</li>
406 * <li>if that fails, inspect CDS features on the cDNA sequence</li>
408 * Returns null if no mapping is determined.
411 * the aligned protein sequence
413 * the aligned cdna sequence
416 public static MapList mapCdnaToProtein(SequenceI proteinSeq,
420 * Here we handle either dataset sequence set (desktop) or absent (applet).
421 * Use only the char[] form of the sequence to avoid creating possibly large
424 final SequenceI proteinDataset = proteinSeq.getDatasetSequence();
425 char[] aaSeqChars = proteinDataset != null ? proteinDataset
426 .getSequence() : proteinSeq.getSequence();
427 final SequenceI cdnaDataset = cdnaSeq.getDatasetSequence();
428 char[] cdnaSeqChars = cdnaDataset != null ? cdnaDataset.getSequence()
429 : cdnaSeq.getSequence();
430 if (aaSeqChars == null || cdnaSeqChars == null)
436 * cdnaStart/End, proteinStartEnd are base 1 (for dataset sequence mapping)
438 final int mappedLength = CODON_LENGTH * aaSeqChars.length;
439 int cdnaLength = cdnaSeqChars.length;
440 int cdnaStart = cdnaSeq.getStart();
441 int cdnaEnd = cdnaSeq.getEnd();
442 final int proteinStart = proteinSeq.getStart();
443 final int proteinEnd = proteinSeq.getEnd();
446 * If lengths don't match, try ignoring stop codon (if present)
448 if (cdnaLength != mappedLength && cdnaLength > 2)
450 String lastCodon = String.valueOf(cdnaSeqChars, cdnaLength - CODON_LENGTH, CODON_LENGTH)
452 for (String stop : ResidueProperties.STOP)
454 if (lastCodon.equals(stop))
456 cdnaEnd -= CODON_LENGTH;
457 cdnaLength -= CODON_LENGTH;
464 * If lengths still don't match, try ignoring start codon.
467 if (cdnaLength != mappedLength
469 && String.valueOf(cdnaSeqChars, 0, CODON_LENGTH).toUpperCase()
470 .equals(ResidueProperties.START))
472 startOffset += CODON_LENGTH;
473 cdnaStart += CODON_LENGTH;
474 cdnaLength -= CODON_LENGTH;
477 if (translatesAs(cdnaSeqChars, startOffset, aaSeqChars))
480 * protein is translation of dna (+/- start/stop codons)
482 MapList map = new MapList(new int[] { cdnaStart, cdnaEnd }, new int[]
483 { proteinStart, proteinEnd }, CODON_LENGTH, 1);
488 * translation failed - try mapping CDS annotated regions of dna
490 return mapCdsToProtein(cdnaSeq, proteinSeq);
494 * Test whether the given cdna sequence, starting at the given offset,
495 * translates to the given amino acid sequence, using the standard translation
496 * table. Designed to fail fast i.e. as soon as a mismatch position is found.
498 * @param cdnaSeqChars
503 protected static boolean translatesAs(char[] cdnaSeqChars, int cdnaStart,
506 if (cdnaSeqChars == null || aaSeqChars == null)
512 int dnaPos = cdnaStart;
513 for (; dnaPos < cdnaSeqChars.length - 2
514 && aaPos < aaSeqChars.length; dnaPos += CODON_LENGTH, aaPos++)
516 String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH);
517 final String translated = ResidueProperties.codonTranslate(codon);
520 * allow * in protein to match untranslatable in dna
522 final char aaRes = aaSeqChars[aaPos];
523 if ((translated == null || "STOP".equals(translated)) && aaRes == '*')
527 if (translated == null || !(aaRes == translated.charAt(0)))
530 // System.out.println(("Mismatch at " + i + "/" + aaResidue + ": "
531 // + codon + "(" + translated + ") != " + aaRes));
537 * check we matched all of the protein sequence
539 if (aaPos != aaSeqChars.length)
545 * check we matched all of the dna except
546 * for optional trailing STOP codon
548 if (dnaPos == cdnaSeqChars.length)
552 if (dnaPos == cdnaSeqChars.length - CODON_LENGTH)
554 String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH);
555 if ("STOP".equals(ResidueProperties.codonTranslate(codon)))
564 * Align sequence 'seq' to match the alignment of a mapped sequence. Note this
565 * currently assumes that we are aligning cDNA to match protein.
568 * the sequence to be realigned
570 * the alignment whose sequence alignment is to be 'copied'
572 * character string represent a gap in the realigned sequence
573 * @param preserveUnmappedGaps
574 * @param preserveMappedGaps
575 * @return true if the sequence was realigned, false if it could not be
577 public static boolean alignSequenceAs(SequenceI seq, AlignmentI al,
578 String gap, boolean preserveMappedGaps,
579 boolean preserveUnmappedGaps)
582 * Get any mappings from the source alignment to the target (dataset)
585 // TODO there may be one AlignedCodonFrame per dataset sequence, or one with
586 // all mappings. Would it help to constrain this?
587 List<AlignedCodonFrame> mappings = al.getCodonFrame(seq);
588 if (mappings == null || mappings.isEmpty())
594 * Locate the aligned source sequence whose dataset sequence is mapped. We
595 * just take the first match here (as we can't align like more than one
598 SequenceI alignFrom = null;
599 AlignedCodonFrame mapping = null;
600 for (AlignedCodonFrame mp : mappings)
602 alignFrom = mp.findAlignedSequence(seq, al);
603 if (alignFrom != null)
610 if (alignFrom == null)
614 alignSequenceAs(seq, alignFrom, mapping, gap, al.getGapCharacter(),
615 preserveMappedGaps, preserveUnmappedGaps);
620 * Align sequence 'alignTo' the same way as 'alignFrom', using the mapping to
621 * match residues and codons. Flags control whether existing gaps in unmapped
622 * (intron) and mapped (exon) regions are preserved or not. Gaps between
623 * intron and exon are only retained if both flags are set.
630 * @param preserveUnmappedGaps
631 * @param preserveMappedGaps
633 public static void alignSequenceAs(SequenceI alignTo,
634 SequenceI alignFrom, AlignedCodonFrame mapping, String myGap,
635 char sourceGap, boolean preserveMappedGaps,
636 boolean preserveUnmappedGaps)
638 // TODO generalise to work for Protein-Protein, dna-dna, dna-protein
640 // aligned and dataset sequence positions, all base zero
644 int basesWritten = 0;
645 char myGapChar = myGap.charAt(0);
646 int ratio = myGap.length();
648 int fromOffset = alignFrom.getStart() - 1;
649 int toOffset = alignTo.getStart() - 1;
650 int sourceGapMappedLength = 0;
651 boolean inExon = false;
652 final char[] thisSeq = alignTo.getSequence();
653 final char[] thatAligned = alignFrom.getSequence();
654 StringBuilder thisAligned = new StringBuilder(2 * thisSeq.length);
657 * Traverse the 'model' aligned sequence
659 for (char sourceChar : thatAligned)
661 if (sourceChar == sourceGap)
663 sourceGapMappedLength += ratio;
668 * Found a non-gap character. Locate its mapped region if any.
671 // Note mapping positions are base 1, our sequence positions base 0
672 int[] mappedPos = mapping.getMappedRegion(alignTo, alignFrom,
673 sourceDsPos + fromOffset);
674 if (mappedPos == null)
677 * unmapped position; treat like a gap
679 sourceGapMappedLength += ratio;
680 // System.err.println("Can't align: no codon mapping to residue "
681 // + sourceDsPos + "(" + sourceChar + ")");
686 int mappedCodonStart = mappedPos[0]; // position (1...) of codon start
687 int mappedCodonEnd = mappedPos[mappedPos.length - 1]; // codon end pos
688 StringBuilder trailingCopiedGap = new StringBuilder();
691 * Copy dna sequence up to and including this codon. Optionally, include
692 * gaps before the codon starts (in introns) and/or after the codon starts
695 * Note this only works for 'linear' splicing, not reverse or interleaved.
696 * But then 'align dna as protein' doesn't make much sense otherwise.
698 int intronLength = 0;
699 while (basesWritten + toOffset < mappedCodonEnd
700 && thisSeqPos < thisSeq.length)
702 final char c = thisSeq[thisSeqPos++];
706 int sourcePosition = basesWritten + toOffset;
707 if (sourcePosition < mappedCodonStart)
710 * Found an unmapped (intron) base. First add in any preceding gaps
713 if (preserveUnmappedGaps && trailingCopiedGap.length() > 0)
715 thisAligned.append(trailingCopiedGap.toString());
716 intronLength += trailingCopiedGap.length();
717 trailingCopiedGap = new StringBuilder();
724 final boolean startOfCodon = sourcePosition == mappedCodonStart;
725 int gapsToAdd = calculateGapsToInsert(preserveMappedGaps,
726 preserveUnmappedGaps, sourceGapMappedLength, inExon,
727 trailingCopiedGap.length(), intronLength, startOfCodon);
728 for (int i = 0; i < gapsToAdd; i++)
730 thisAligned.append(myGapChar);
732 sourceGapMappedLength = 0;
735 thisAligned.append(c);
736 trailingCopiedGap = new StringBuilder();
740 if (inExon && preserveMappedGaps)
742 trailingCopiedGap.append(myGapChar);
744 else if (!inExon && preserveUnmappedGaps)
746 trailingCopiedGap.append(myGapChar);
753 * At end of model aligned sequence. Copy any remaining target sequence, optionally
754 * including (intron) gaps.
756 while (thisSeqPos < thisSeq.length)
758 final char c = thisSeq[thisSeqPos++];
759 if (c != myGapChar || preserveUnmappedGaps)
761 thisAligned.append(c);
763 sourceGapMappedLength--;
767 * finally add gaps to pad for any trailing source gaps or
768 * unmapped characters
770 if (preserveUnmappedGaps)
772 while (sourceGapMappedLength > 0)
774 thisAligned.append(myGapChar);
775 sourceGapMappedLength--;
780 * All done aligning, set the aligned sequence.
782 alignTo.setSequence(new String(thisAligned));
786 * Helper method to work out how many gaps to insert when realigning.
788 * @param preserveMappedGaps
789 * @param preserveUnmappedGaps
790 * @param sourceGapMappedLength
792 * @param trailingCopiedGap
793 * @param intronLength
794 * @param startOfCodon
797 protected static int calculateGapsToInsert(boolean preserveMappedGaps,
798 boolean preserveUnmappedGaps, int sourceGapMappedLength,
799 boolean inExon, int trailingGapLength, int intronLength,
800 final boolean startOfCodon)
806 * Reached start of codon. Ignore trailing gaps in intron unless we are
807 * preserving gaps in both exon and intron. Ignore them anyway if the
808 * protein alignment introduces a gap at least as large as the intronic
811 if (inExon && !preserveMappedGaps)
813 trailingGapLength = 0;
815 if (!inExon && !(preserveMappedGaps && preserveUnmappedGaps))
817 trailingGapLength = 0;
821 gapsToAdd = Math.max(sourceGapMappedLength, trailingGapLength);
825 if (intronLength + trailingGapLength <= sourceGapMappedLength)
827 gapsToAdd = sourceGapMappedLength - intronLength;
831 gapsToAdd = Math.min(intronLength + trailingGapLength
832 - sourceGapMappedLength, trailingGapLength);
839 * second or third base of codon; check for any gaps in dna
841 if (!preserveMappedGaps)
843 trailingGapLength = 0;
845 gapsToAdd = Math.max(sourceGapMappedLength, trailingGapLength);
851 * Realigns the given protein to match the alignment of the dna, using codon
852 * mappings to translate aligned codon positions to protein residues.
855 * the alignment whose sequences are realigned by this method
857 * the dna alignment whose alignment we are 'copying'
858 * @return the number of sequences that were realigned
860 public static int alignProteinAsDna(AlignmentI protein, AlignmentI dna)
862 if (protein.isNucleotide() || !dna.isNucleotide())
864 System.err.println("Wrong alignment type in alignProteinAsDna");
867 List<SequenceI> unmappedProtein = new ArrayList<SequenceI>();
868 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons = buildCodonColumnsMap(
869 protein, dna, unmappedProtein);
870 return alignProteinAs(protein, alignedCodons, unmappedProtein);
874 * Realigns the given dna to match the alignment of the protein, using codon
875 * mappings to translate aligned peptide positions to codons.
877 * Always produces a padded CDS alignment.
880 * the alignment whose sequences are realigned by this method
882 * the protein alignment whose alignment we are 'copying'
883 * @return the number of sequences that were realigned
885 public static int alignCdsAsProtein(AlignmentI dna, AlignmentI protein)
887 if (protein.isNucleotide() || !dna.isNucleotide())
889 System.err.println("Wrong alignment type in alignProteinAsDna");
892 // todo: implement this
893 List<AlignedCodonFrame> mappings = protein.getCodonFrames();
894 int alignedCount = 0;
895 int width = 0; // alignment width for padding CDS
896 for (SequenceI dnaSeq : dna.getSequences())
898 if (alignCdsSequenceAsProtein(dnaSeq, protein, mappings,
899 dna.getGapCharacter()))
903 width = Math.max(dnaSeq.getLength(), width);
907 for (SequenceI dnaSeq : dna.getSequences())
909 oldwidth = dnaSeq.getLength();
910 diff = width - oldwidth;
913 dnaSeq.insertCharAt(oldwidth, diff, dna.getGapCharacter());
920 * Helper method to align (if possible) the dna sequence to match the
921 * alignment of a mapped protein sequence. This is currently limited to
922 * handling coding sequence only.
930 static boolean alignCdsSequenceAsProtein(SequenceI cdsSeq,
931 AlignmentI protein, List<AlignedCodonFrame> mappings, char gapChar)
933 SequenceI cdsDss = cdsSeq.getDatasetSequence();
937 .println("alignCdsSequenceAsProtein needs aligned sequence!");
941 List<AlignedCodonFrame> dnaMappings = MappingUtils
942 .findMappingsForSequence(cdsSeq, mappings);
943 for (AlignedCodonFrame mapping : dnaMappings)
945 SequenceI peptide = mapping.findAlignedSequence(cdsSeq, protein);
948 int peptideLength = peptide.getLength();
949 Mapping map = mapping.getMappingBetween(cdsSeq, peptide);
952 MapList mapList = map.getMap();
953 if (map.getTo() == peptide.getDatasetSequence())
955 mapList = mapList.getInverse();
957 int cdsLength = cdsDss.getLength();
958 int mappedFromLength = MappingUtils.getLength(mapList
960 int mappedToLength = MappingUtils
961 .getLength(mapList.getToRanges());
962 boolean addStopCodon = (cdsLength == mappedFromLength * CODON_LENGTH + CODON_LENGTH)
963 || (peptide.getDatasetSequence().getLength() == mappedFromLength - 1);
964 if (cdsLength != mappedToLength && !addStopCodon)
968 .format("Can't align cds as protein (length mismatch %d/%d): %s",
969 cdsLength, mappedToLength,
974 * pre-fill the aligned cds sequence with gaps
976 char[] alignedCds = new char[peptideLength * CODON_LENGTH
977 + (addStopCodon ? CODON_LENGTH : 0)];
978 Arrays.fill(alignedCds, gapChar);
981 * walk over the aligned peptide sequence and insert mapped
982 * codons for residues in the aligned cds sequence
984 char[] alignedPeptide = peptide.getSequence();
985 char[] nucleotides = cdsDss.getSequence();
987 int cdsStart = cdsDss.getStart();
988 int proteinPos = peptide.getStart() - 1;
990 for (char residue : alignedPeptide)
992 if (Comparison.isGap(residue))
994 cdsCol += CODON_LENGTH;
999 int[] codon = mapList.locateInTo(proteinPos, proteinPos);
1002 // e.g. incomplete start codon, X in peptide
1003 cdsCol += CODON_LENGTH;
1007 for (int j = codon[0]; j <= codon[1]; j++)
1009 char mappedBase = nucleotides[j - cdsStart];
1010 alignedCds[cdsCol++] = mappedBase;
1018 * append stop codon if not mapped from protein,
1019 * closing it up to the end of the mapped sequence
1021 if (copiedBases == nucleotides.length - CODON_LENGTH)
1023 for (int i = alignedCds.length - 1; i >= 0; i--)
1025 if (!Comparison.isGap(alignedCds[i]))
1027 cdsCol = i + 1; // gap just after end of sequence
1031 for (int i = nucleotides.length - CODON_LENGTH; i < nucleotides.length; i++)
1033 alignedCds[cdsCol++] = nucleotides[i];
1036 cdsSeq.setSequence(new String(alignedCds));
1045 * Builds a map whose key is an aligned codon position (3 alignment column
1046 * numbers base 0), and whose value is a map from protein sequence to each
1047 * protein's peptide residue for that codon. The map generates an ordering of
1048 * the codons, and allows us to read off the peptides at each position in
1049 * order to assemble 'aligned' protein sequences.
1052 * the protein alignment
1054 * the coding dna alignment
1055 * @param unmappedProtein
1056 * any unmapped proteins are added to this list
1059 protected static Map<AlignedCodon, Map<SequenceI, AlignedCodon>> buildCodonColumnsMap(
1060 AlignmentI protein, AlignmentI dna,
1061 List<SequenceI> unmappedProtein)
1064 * maintain a list of any proteins with no mappings - these will be
1065 * rendered 'as is' in the protein alignment as we can't align them
1067 unmappedProtein.addAll(protein.getSequences());
1069 List<AlignedCodonFrame> mappings = protein.getCodonFrames();
1072 * Map will hold, for each aligned codon position e.g. [3, 5, 6], a map of
1073 * {dnaSequence, {proteinSequence, codonProduct}} at that position. The
1074 * comparator keeps the codon positions ordered.
1076 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons = new TreeMap<AlignedCodon, Map<SequenceI, AlignedCodon>>(
1077 new CodonComparator());
1079 for (SequenceI dnaSeq : dna.getSequences())
1081 for (AlignedCodonFrame mapping : mappings)
1083 SequenceI prot = mapping.findAlignedSequence(dnaSeq, protein);
1086 Mapping seqMap = mapping.getMappingForSequence(dnaSeq);
1087 addCodonPositions(dnaSeq, prot, protein.getGapCharacter(),
1088 seqMap, alignedCodons);
1089 unmappedProtein.remove(prot);
1095 * Finally add any unmapped peptide start residues (e.g. for incomplete
1096 * codons) as if at the codon position before the second residue
1098 // TODO resolve JAL-2022 so this fudge can be removed
1099 int mappedSequenceCount = protein.getHeight() - unmappedProtein.size();
1100 addUnmappedPeptideStarts(alignedCodons, mappedSequenceCount);
1102 return alignedCodons;
1106 * Scans for any protein mapped from position 2 (meaning unmapped start
1107 * position e.g. an incomplete codon), and synthesizes a 'codon' for it at the
1108 * preceding position in the alignment
1110 * @param alignedCodons
1111 * the codon-to-peptide map
1112 * @param mappedSequenceCount
1113 * the number of distinct sequences in the map
1115 protected static void addUnmappedPeptideStarts(
1116 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons,
1117 int mappedSequenceCount)
1119 // TODO delete this ugly hack once JAL-2022 is resolved
1120 // i.e. we can model startPhase > 0 (incomplete start codon)
1122 List<SequenceI> sequencesChecked = new ArrayList<SequenceI>();
1123 AlignedCodon lastCodon = null;
1124 Map<SequenceI, AlignedCodon> toAdd = new HashMap<SequenceI, AlignedCodon>();
1126 for (Entry<AlignedCodon, Map<SequenceI, AlignedCodon>> entry : alignedCodons
1129 for (Entry<SequenceI, AlignedCodon> sequenceCodon : entry.getValue()
1132 SequenceI seq = sequenceCodon.getKey();
1133 if (sequencesChecked.contains(seq))
1137 sequencesChecked.add(seq);
1138 AlignedCodon codon = sequenceCodon.getValue();
1139 if (codon.peptideCol > 1)
1142 .println("Problem mapping protein with >1 unmapped start positions: "
1145 else if (codon.peptideCol == 1)
1148 * first position (peptideCol == 0) was unmapped - add it
1150 if (lastCodon != null)
1152 AlignedCodon firstPeptide = new AlignedCodon(lastCodon.pos1,
1153 lastCodon.pos2, lastCodon.pos3, String.valueOf(seq
1155 toAdd.put(seq, firstPeptide);
1160 * unmapped residue at start of alignment (no prior column) -
1161 * 'insert' at nominal codon [0, 0, 0]
1163 AlignedCodon firstPeptide = new AlignedCodon(0, 0, 0,
1164 String.valueOf(seq.getCharAt(0)), 0);
1165 toAdd.put(seq, firstPeptide);
1168 if (sequencesChecked.size() == mappedSequenceCount)
1170 // no need to check past first mapped position in all sequences
1174 lastCodon = entry.getKey();
1178 * add any new codons safely after iterating over the map
1180 for (Entry<SequenceI, AlignedCodon> startCodon : toAdd.entrySet())
1182 addCodonToMap(alignedCodons, startCodon.getValue(),
1183 startCodon.getKey());
1188 * Update the aligned protein sequences to match the codon alignments given in
1192 * @param alignedCodons
1193 * an ordered map of codon positions (columns), with sequence/peptide
1194 * values present in each column
1195 * @param unmappedProtein
1198 protected static int alignProteinAs(AlignmentI protein,
1199 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons,
1200 List<SequenceI> unmappedProtein)
1203 * Prefill aligned sequences with gaps before inserting aligned protein
1206 int alignedWidth = alignedCodons.size();
1207 char[] gaps = new char[alignedWidth];
1208 Arrays.fill(gaps, protein.getGapCharacter());
1209 String allGaps = String.valueOf(gaps);
1210 for (SequenceI seq : protein.getSequences())
1212 if (!unmappedProtein.contains(seq))
1214 seq.setSequence(allGaps);
1219 for (AlignedCodon codon : alignedCodons.keySet())
1221 final Map<SequenceI, AlignedCodon> columnResidues = alignedCodons
1223 for (Entry<SequenceI, AlignedCodon> entry : columnResidues.entrySet())
1225 // place translated codon at its column position in sequence
1226 entry.getKey().getSequence()[column] = entry.getValue().product
1235 * Populate the map of aligned codons by traversing the given sequence
1236 * mapping, locating the aligned positions of mapped codons, and adding those
1237 * positions and their translation products to the map.
1240 * the aligned sequence we are mapping from
1242 * the sequence to be aligned to the codons
1244 * the gap character in the dna sequence
1246 * a mapping to a sequence translation
1247 * @param alignedCodons
1248 * the map we are building up
1250 static void addCodonPositions(SequenceI dna, SequenceI protein,
1251 char gapChar, Mapping seqMap,
1252 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons)
1254 Iterator<AlignedCodon> codons = seqMap.getCodonIterator(dna, gapChar);
1257 * add codon positions, and their peptide translations, to the alignment
1258 * map, while remembering the first codon mapped
1260 while (codons.hasNext())
1264 AlignedCodon codon = codons.next();
1265 addCodonToMap(alignedCodons, codon, protein);
1266 } catch (IncompleteCodonException e)
1268 // possible incomplete trailing codon - ignore
1269 } catch (NoSuchElementException e)
1271 // possibly peptide lacking STOP
1277 * Helper method to add a codon-to-peptide entry to the aligned codons map
1279 * @param alignedCodons
1283 protected static void addCodonToMap(
1284 Map<AlignedCodon, Map<SequenceI, AlignedCodon>> alignedCodons,
1285 AlignedCodon codon, SequenceI protein)
1287 Map<SequenceI, AlignedCodon> seqProduct = alignedCodons.get(codon);
1288 if (seqProduct == null)
1290 seqProduct = new HashMap<SequenceI, AlignedCodon>();
1291 alignedCodons.put(codon, seqProduct);
1293 seqProduct.put(protein, codon);
1297 * Returns true if a cDNA/Protein mapping either exists, or could be made,
1298 * between at least one pair of sequences in the two alignments. Currently,
1301 * <li>One alignment must be nucleotide, and the other protein</li>
1302 * <li>At least one pair of sequences must be already mapped, or mappable</li>
1303 * <li>Mappable means the nucleotide translation matches the protein sequence</li>
1304 * <li>The translation may ignore start and stop codons if present in the
1312 public static boolean isMappable(AlignmentI al1, AlignmentI al2)
1314 if (al1 == null || al2 == null)
1320 * Require one nucleotide and one protein
1322 if (al1.isNucleotide() == al2.isNucleotide())
1326 AlignmentI dna = al1.isNucleotide() ? al1 : al2;
1327 AlignmentI protein = dna == al1 ? al2 : al1;
1328 List<AlignedCodonFrame> mappings = protein.getCodonFrames();
1329 for (SequenceI dnaSeq : dna.getSequences())
1331 for (SequenceI proteinSeq : protein.getSequences())
1333 if (isMappable(dnaSeq, proteinSeq, mappings))
1343 * Returns true if the dna sequence is mapped, or could be mapped, to the
1351 protected static boolean isMappable(SequenceI dnaSeq,
1352 SequenceI proteinSeq, List<AlignedCodonFrame> mappings)
1354 if (dnaSeq == null || proteinSeq == null)
1359 SequenceI dnaDs = dnaSeq.getDatasetSequence() == null ? dnaSeq : dnaSeq
1360 .getDatasetSequence();
1361 SequenceI proteinDs = proteinSeq.getDatasetSequence() == null ? proteinSeq
1362 : proteinSeq.getDatasetSequence();
1364 for (AlignedCodonFrame mapping : mappings)
1366 if (proteinDs == mapping.getAaForDnaSeq(dnaDs))
1376 * Just try to make a mapping (it is not yet stored), test whether
1379 return mapCdnaToProtein(proteinDs, dnaDs) != null;
1383 * Finds any reference annotations associated with the sequences in
1384 * sequenceScope, that are not already added to the alignment, and adds them
1385 * to the 'candidates' map. Also populates a lookup table of annotation
1386 * labels, keyed by calcId, for use in constructing tooltips or the like.
1388 * @param sequenceScope
1389 * the sequences to scan for reference annotations
1390 * @param labelForCalcId
1391 * (optional) map to populate with label for calcId
1393 * map to populate with annotations for sequence
1395 * the alignment to check for presence of annotations
1397 public static void findAddableReferenceAnnotations(
1398 List<SequenceI> sequenceScope,
1399 Map<String, String> labelForCalcId,
1400 final Map<SequenceI, List<AlignmentAnnotation>> candidates,
1403 if (sequenceScope == null)
1409 * For each sequence in scope, make a list of any annotations on the
1410 * underlying dataset sequence which are not already on the alignment.
1412 * Add to a map of { alignmentSequence, <List of annotations to add> }
1414 for (SequenceI seq : sequenceScope)
1416 SequenceI dataset = seq.getDatasetSequence();
1417 if (dataset == null)
1421 AlignmentAnnotation[] datasetAnnotations = dataset.getAnnotation();
1422 if (datasetAnnotations == null)
1426 final List<AlignmentAnnotation> result = new ArrayList<AlignmentAnnotation>();
1427 for (AlignmentAnnotation dsann : datasetAnnotations)
1430 * Find matching annotations on the alignment. If none is found, then
1431 * add this annotation to the list of 'addable' annotations for this
1434 final Iterable<AlignmentAnnotation> matchedAlignmentAnnotations = al
1435 .findAnnotations(seq, dsann.getCalcId(), dsann.label);
1436 if (!matchedAlignmentAnnotations.iterator().hasNext())
1439 if (labelForCalcId != null)
1441 labelForCalcId.put(dsann.getCalcId(), dsann.label);
1446 * Save any addable annotations for this sequence
1448 if (!result.isEmpty())
1450 candidates.put(seq, result);
1456 * Adds annotations to the top of the alignment annotations, in the same order
1457 * as their related sequences.
1459 * @param annotations
1460 * the annotations to add
1462 * the alignment to add them to
1463 * @param selectionGroup
1464 * current selection group (or null if none)
1466 public static void addReferenceAnnotations(
1467 Map<SequenceI, List<AlignmentAnnotation>> annotations,
1468 final AlignmentI alignment, final SequenceGroup selectionGroup)
1470 for (SequenceI seq : annotations.keySet())
1472 for (AlignmentAnnotation ann : annotations.get(seq))
1474 AlignmentAnnotation copyAnn = new AlignmentAnnotation(ann);
1476 int endRes = ann.annotations.length;
1477 if (selectionGroup != null)
1479 startRes = selectionGroup.getStartRes();
1480 endRes = selectionGroup.getEndRes();
1482 copyAnn.restrict(startRes, endRes);
1485 * Add to the sequence (sets copyAnn.datasetSequence), unless the
1486 * original annotation is already on the sequence.
1488 if (!seq.hasAnnotation(ann))
1490 seq.addAlignmentAnnotation(copyAnn);
1493 copyAnn.adjustForAlignment();
1494 // add to the alignment and set visible
1495 alignment.addAnnotation(copyAnn);
1496 copyAnn.visible = true;
1502 * Set visibility of alignment annotations of specified types (labels), for
1503 * specified sequences. This supports controls like
1504 * "Show all secondary structure", "Hide all Temp factor", etc.
1506 * @al the alignment to scan for annotations
1508 * the types (labels) of annotations to be updated
1509 * @param forSequences
1510 * if not null, only annotations linked to one of these sequences are
1511 * in scope for update; if null, acts on all sequence annotations
1513 * if this flag is true, 'types' is ignored (label not checked)
1515 * if true, set visibility on, else set off
1517 public static void showOrHideSequenceAnnotations(AlignmentI al,
1518 Collection<String> types, List<SequenceI> forSequences,
1519 boolean anyType, boolean doShow)
1521 AlignmentAnnotation[] anns = al.getAlignmentAnnotation();
1524 for (AlignmentAnnotation aa : anns)
1526 if (anyType || types.contains(aa.label))
1528 if ((aa.sequenceRef != null)
1529 && (forSequences == null || forSequences
1530 .contains(aa.sequenceRef)))
1532 aa.visible = doShow;
1540 * Returns true if either sequence has a cross-reference to the other
1546 public static boolean haveCrossRef(SequenceI seq1, SequenceI seq2)
1548 // Note: moved here from class CrossRef as the latter class has dependencies
1549 // not availability to the applet's classpath
1550 return hasCrossRef(seq1, seq2) || hasCrossRef(seq2, seq1);
1554 * Returns true if seq1 has a cross-reference to seq2. Currently this assumes
1555 * that sequence name is structured as Source|AccessionId.
1561 public static boolean hasCrossRef(SequenceI seq1, SequenceI seq2)
1563 if (seq1 == null || seq2 == null)
1567 String name = seq2.getName();
1568 final DBRefEntry[] xrefs = seq1.getDBRefs();
1571 for (DBRefEntry xref : xrefs)
1573 String xrefName = xref.getSource() + "|" + xref.getAccessionId();
1574 // case-insensitive test, consistent with DBRefEntry.equalRef()
1575 if (xrefName.equalsIgnoreCase(name))
1585 * Constructs an alignment consisting of the mapped (CDS) regions in the given
1586 * nucleotide sequences, and updates mappings to match. The CDS sequences are
1587 * added to the original alignment's dataset, which is shared by the new
1588 * alignment. Mappings from nucleotide to CDS, and from CDS to protein, are
1589 * added to the alignment dataset.
1592 * aligned nucleotide (dna or cds) sequences
1594 * the alignment dataset the sequences belong to
1596 * (optional) to restrict results to CDS that map to specified
1598 * @return an alignment whose sequences are the cds-only parts of the dna
1599 * sequences (or null if no mappings are found)
1601 public static AlignmentI makeCdsAlignment(SequenceI[] dna,
1602 AlignmentI dataset, SequenceI[] products)
1604 if (dataset == null || dataset.getDataset() != null)
1606 throw new IllegalArgumentException(
1607 "IMPLEMENTATION ERROR: dataset.getDataset() must be null!");
1609 List<SequenceI> foundSeqs = new ArrayList<SequenceI>();
1610 List<SequenceI> cdsSeqs = new ArrayList<SequenceI>();
1611 List<AlignedCodonFrame> mappings = dataset.getCodonFrames();
1612 HashSet<SequenceI> productSeqs = null;
1613 if (products != null)
1615 productSeqs = new HashSet<SequenceI>();
1616 for (SequenceI seq : products)
1618 productSeqs.add(seq.getDatasetSequence() == null ? seq : seq
1619 .getDatasetSequence());
1624 * Construct CDS sequences from mappings on the alignment dataset.
1626 * - find the protein product(s) mapped to from each dna sequence
1627 * - if the mapping covers the whole dna sequence (give or take start/stop
1628 * codon), take the dna as the CDS sequence
1629 * - else search dataset mappings for a suitable dna sequence, i.e. one
1630 * whose whole sequence is mapped to the protein
1631 * - if no sequence found, construct one from the dna sequence and mapping
1632 * (and add it to dataset so it is found if this is repeated)
1634 for (SequenceI dnaSeq : dna)
1636 SequenceI dnaDss = dnaSeq.getDatasetSequence() == null ? dnaSeq
1637 : dnaSeq.getDatasetSequence();
1639 List<AlignedCodonFrame> seqMappings = MappingUtils
1640 .findMappingsForSequence(dnaSeq, mappings);
1641 for (AlignedCodonFrame mapping : seqMappings)
1643 List<Mapping> mappingsFromSequence = mapping
1644 .getMappingsFromSequence(dnaSeq);
1646 for (Mapping aMapping : mappingsFromSequence)
1648 MapList mapList = aMapping.getMap();
1649 if (mapList.getFromRatio() == 1)
1652 * not a dna-to-protein mapping (likely dna-to-cds)
1658 * skip if mapping is not to one of the target set of proteins
1660 SequenceI proteinProduct = aMapping.getTo();
1661 if (productSeqs != null && !productSeqs.contains(proteinProduct))
1667 * try to locate the CDS from the dataset mappings;
1668 * guard against duplicate results (for the case that protein has
1669 * dbrefs to both dna and cds sequences)
1671 SequenceI cdsSeq = findCdsForProtein(mappings, dnaSeq,
1672 seqMappings, aMapping);
1675 if (!foundSeqs.contains(cdsSeq))
1677 foundSeqs.add(cdsSeq);
1678 SequenceI derivedSequence = cdsSeq.deriveSequence();
1679 cdsSeqs.add(derivedSequence);
1680 if (!dataset.getSequences().contains(cdsSeq))
1682 dataset.addSequence(cdsSeq);
1689 * didn't find mapped CDS sequence - construct it and add
1690 * its dataset sequence to the dataset
1692 cdsSeq = makeCdsSequence(dnaSeq.getDatasetSequence(), aMapping);
1693 SequenceI cdsSeqDss = cdsSeq.createDatasetSequence();
1694 cdsSeqs.add(cdsSeq);
1695 if (!dataset.getSequences().contains(cdsSeqDss))
1697 dataset.addSequence(cdsSeqDss);
1701 * add a mapping from CDS to the (unchanged) mapped to range
1703 List<int[]> cdsRange = Collections.singletonList(new int[] { 1,
1704 cdsSeq.getLength() });
1705 MapList cdsToProteinMap = new MapList(cdsRange, mapList.getToRanges(),
1706 mapList.getFromRatio(), mapList.getToRatio());
1707 AlignedCodonFrame cdsToProteinMapping = new AlignedCodonFrame();
1708 cdsToProteinMapping.addMap(cdsSeq, proteinProduct, cdsToProteinMap);
1711 * guard against duplicating the mapping if repeating this action
1713 if (!mappings.contains(cdsToProteinMapping))
1715 mappings.add(cdsToProteinMapping);
1719 * copy protein's dbrefs to CDS sequence
1720 * this enables Get Cross-References from CDS alignment
1722 DBRefEntry[] proteinRefs = DBRefUtils.selectDbRefs(false,
1723 proteinProduct.getDBRefs());
1724 if (proteinRefs != null)
1726 for (DBRefEntry ref : proteinRefs)
1728 DBRefEntry cdsToProteinRef = new DBRefEntry(ref);
1729 cdsToProteinRef.setMap(new Mapping(proteinProduct,
1731 cdsSeqDss.addDBRef(cdsToProteinRef);
1736 * add another mapping from original 'from' range to CDS
1738 AlignedCodonFrame dnaToCdsMapping = new AlignedCodonFrame();
1739 MapList dnaToCdsMap = new MapList(mapList.getFromRanges(),
1742 dnaToCdsMapping.addMap(dnaSeq.getDatasetSequence(), cdsSeq,
1744 if (!mappings.contains(dnaToCdsMapping))
1746 mappings.add(dnaToCdsMapping);
1750 * add DBRef with mapping from protein to CDS
1751 * (this enables Get Cross-References from protein alignment)
1752 * This is tricky because we can't have two DBRefs with the
1753 * same source and accession, so need a different accession for
1754 * the CDS from the dna sequence
1756 DBRefEntryI dnaRef = dnaDss.getSourceDBRef();
1759 // assuming cds version same as dna ?!?
1760 DBRefEntry proteinToCdsRef = new DBRefEntry(dnaRef.getSource(),
1761 dnaRef.getVersion(), cdsSeq.getName());
1762 proteinToCdsRef.setMap(new Mapping(cdsSeqDss, cdsToProteinMap
1764 proteinProduct.addDBRef(proteinToCdsRef);
1768 * transfer any features on dna that overlap the CDS
1770 transferFeatures(dnaSeq, cdsSeq, cdsToProteinMap, null,
1771 SequenceOntologyI.CDS);
1776 AlignmentI cds = new Alignment(cdsSeqs.toArray(new SequenceI[cdsSeqs
1778 cds.setDataset(dataset);
1784 * A helper method that finds a CDS sequence in the alignment dataset that is
1785 * mapped to the given protein sequence, and either is, or has a mapping from,
1786 * the given dna sequence.
1789 * set of all mappings on the dataset
1791 * a dna (or cds) sequence we are searching from
1792 * @param seqMappings
1793 * the set of mappings involving dnaSeq
1795 * an initial candidate from seqMappings
1798 static SequenceI findCdsForProtein(List<AlignedCodonFrame> mappings,
1799 SequenceI dnaSeq, List<AlignedCodonFrame> seqMappings,
1803 * TODO a better dna-cds-protein mapping data representation to allow easy
1804 * navigation; until then this clunky looping around lists of mappings
1806 SequenceI seqDss = dnaSeq.getDatasetSequence() == null ? dnaSeq
1807 : dnaSeq.getDatasetSequence();
1808 SequenceI proteinProduct = aMapping.getTo();
1811 * is this mapping from the whole dna sequence (i.e. CDS)?
1812 * allowing for possible stop codon on dna but not peptide
1814 int mappedFromLength = MappingUtils.getLength(aMapping.getMap()
1816 int dnaLength = seqDss.getLength();
1817 if (mappedFromLength == dnaLength || mappedFromLength == dnaLength - CODON_LENGTH)
1823 * looks like we found the dna-to-protein mapping; search for the
1824 * corresponding cds-to-protein mapping
1826 List<AlignedCodonFrame> mappingsToPeptide = MappingUtils
1827 .findMappingsForSequence(proteinProduct, mappings);
1828 for (AlignedCodonFrame acf : mappingsToPeptide)
1830 for (SequenceToSequenceMapping map : acf.getMappings())
1832 Mapping mapping = map.getMapping();
1833 if (mapping != aMapping && mapping.getMap().getFromRatio() == CODON_LENGTH
1834 && proteinProduct == mapping.getTo()
1835 && seqDss != map.getFromSeq())
1837 mappedFromLength = MappingUtils.getLength(mapping.getMap()
1839 if (mappedFromLength == map.getFromSeq().getLength())
1842 * found a 3:1 mapping to the protein product which covers
1843 * the whole dna sequence i.e. is from CDS; finally check it
1844 * is from the dna start sequence
1846 SequenceI cdsSeq = map.getFromSeq();
1847 List<AlignedCodonFrame> dnaToCdsMaps = MappingUtils
1848 .findMappingsForSequence(cdsSeq, seqMappings);
1849 if (!dnaToCdsMaps.isEmpty())
1861 * Helper method that makes a CDS sequence as defined by the mappings from the
1862 * given sequence i.e. extracts the 'mapped from' ranges (which may be on
1863 * forward or reverse strand).
1867 * @return CDS sequence (as a dataset sequence)
1869 static SequenceI makeCdsSequence(SequenceI seq, Mapping mapping)
1871 char[] seqChars = seq.getSequence();
1872 List<int[]> fromRanges = mapping.getMap().getFromRanges();
1873 int cdsWidth = MappingUtils.getLength(fromRanges);
1874 char[] newSeqChars = new char[cdsWidth];
1877 for (int[] range : fromRanges)
1879 if (range[0] <= range[1])
1881 // forward strand mapping - just copy the range
1882 int length = range[1] - range[0] + 1;
1883 System.arraycopy(seqChars, range[0] - 1, newSeqChars, newPos,
1889 // reverse strand mapping - copy and complement one by one
1890 for (int i = range[0]; i >= range[1]; i--)
1892 newSeqChars[newPos++] = Dna.getComplement(seqChars[i - 1]);
1898 * assign 'from id' held in the mapping if set (e.g. EMBL protein_id),
1899 * else generate a sequence name
1901 String mapFromId = mapping.getMappedFromId();
1902 String seqId = "CDS|" + (mapFromId != null ? mapFromId : seq.getName());
1903 SequenceI newSeq = new Sequence(seqId, newSeqChars, 1, newPos);
1904 // newSeq.setDescription(mapFromId);
1910 * Transfers co-located features on 'fromSeq' to 'toSeq', adjusting the
1911 * feature start/end ranges, optionally omitting specified feature types.
1912 * Returns the number of features copied.
1917 * if not null, only features of this type are copied (including
1918 * subtypes in the Sequence Ontology)
1920 * the mapping from 'fromSeq' to 'toSeq'
1923 public static int transferFeatures(SequenceI fromSeq, SequenceI toSeq,
1924 MapList mapping, String select, String... omitting)
1926 SequenceI copyTo = toSeq;
1927 while (copyTo.getDatasetSequence() != null)
1929 copyTo = copyTo.getDatasetSequence();
1932 SequenceOntologyI so = SequenceOntologyFactory.getInstance();
1934 SequenceFeature[] sfs = fromSeq.getSequenceFeatures();
1937 for (SequenceFeature sf : sfs)
1939 String type = sf.getType();
1940 if (select != null && !so.isA(type, select))
1944 boolean omit = false;
1945 for (String toOmit : omitting)
1947 if (type.equals(toOmit))
1958 * locate the mapped range - null if either start or end is
1959 * not mapped (no partial overlaps are calculated)
1961 int start = sf.getBegin();
1962 int end = sf.getEnd();
1963 int[] mappedTo = mapping.locateInTo(start, end);
1965 * if whole exon range doesn't map, try interpreting it
1966 * as 5' or 3' exon overlapping the CDS range
1968 if (mappedTo == null)
1970 mappedTo = mapping.locateInTo(end, end);
1971 if (mappedTo != null)
1974 * end of exon is in CDS range - 5' overlap
1975 * to a range from the start of the peptide
1980 if (mappedTo == null)
1982 mappedTo = mapping.locateInTo(start, start);
1983 if (mappedTo != null)
1986 * start of exon is in CDS range - 3' overlap
1987 * to a range up to the end of the peptide
1989 mappedTo[1] = toSeq.getLength();
1992 if (mappedTo != null)
1994 SequenceFeature copy = new SequenceFeature(sf);
1995 copy.setBegin(Math.min(mappedTo[0], mappedTo[1]));
1996 copy.setEnd(Math.max(mappedTo[0], mappedTo[1]));
1997 copyTo.addSequenceFeature(copy);
2006 * Returns a mapping from dna to protein by inspecting sequence features of
2007 * type "CDS" on the dna.
2013 public static MapList mapCdsToProtein(SequenceI dnaSeq,
2014 SequenceI proteinSeq)
2016 List<int[]> ranges = findCdsPositions(dnaSeq);
2017 int mappedDnaLength = MappingUtils.getLength(ranges);
2019 int proteinLength = proteinSeq.getLength();
2020 int proteinStart = proteinSeq.getStart();
2021 int proteinEnd = proteinSeq.getEnd();
2024 * incomplete start codon may mean X at start of peptide
2025 * we ignore both for mapping purposes
2027 if (proteinSeq.getCharAt(0) == 'X')
2029 // todo JAL-2022 support startPhase > 0
2033 List<int[]> proteinRange = new ArrayList<int[]>();
2036 * dna length should map to protein (or protein plus stop codon)
2038 int codesForResidues = mappedDnaLength / CODON_LENGTH;
2039 if (codesForResidues == (proteinLength + 1))
2041 // assuming extra codon is for STOP and not in peptide
2044 if (codesForResidues == proteinLength)
2046 proteinRange.add(new int[] { proteinStart, proteinEnd });
2047 return new MapList(ranges, proteinRange, CODON_LENGTH, 1);
2053 * Returns a list of CDS ranges found (as sequence positions base 1), i.e. of
2054 * start/end positions of sequence features of type "CDS" (or a sub-type of
2055 * CDS in the Sequence Ontology). The ranges are sorted into ascending start
2056 * position order, so this method is only valid for linear CDS in the same
2057 * sense as the protein product.
2062 public static List<int[]> findCdsPositions(SequenceI dnaSeq)
2064 List<int[]> result = new ArrayList<int[]>();
2065 SequenceFeature[] sfs = dnaSeq.getSequenceFeatures();
2071 SequenceOntologyI so = SequenceOntologyFactory.getInstance();
2074 for (SequenceFeature sf : sfs)
2077 * process a CDS feature (or a sub-type of CDS)
2079 if (so.isA(sf.getType(), SequenceOntologyI.CDS))
2084 phase = Integer.parseInt(sf.getPhase());
2085 } catch (NumberFormatException e)
2090 * phase > 0 on first codon means 5' incomplete - skip to the start
2091 * of the next codon; example ENST00000496384
2093 int begin = sf.getBegin();
2094 int end = sf.getEnd();
2095 if (result.isEmpty())
2100 // shouldn't happen!
2102 .println("Error: start phase extends beyond start CDS in "
2103 + dnaSeq.getName());
2106 result.add(new int[] { begin, end });
2111 * remove 'startPhase' positions (usually 0) from the first range
2112 * so we begin at the start of a complete codon
2114 if (!result.isEmpty())
2116 // TODO JAL-2022 correctly model start phase > 0
2117 result.get(0)[0] += startPhase;
2121 * Finally sort ranges by start position. This avoids a dependency on
2122 * keeping features in order on the sequence (if they are in order anyway,
2123 * the sort will have almost no work to do). The implicit assumption is CDS
2124 * ranges are assembled in order. Other cases should not use this method,
2125 * but instead construct an explicit mapping for CDS (e.g. EMBL parsing).
2127 Collections.sort(result, new Comparator<int[]>()
2130 public int compare(int[] o1, int[] o2)
2132 return Integer.compare(o1[0], o2[0]);
2139 * Maps exon features from dna to protein, and computes variants in peptide
2140 * product generated by variants in dna, and adds them as sequence_variant
2141 * features on the protein sequence. Returns the number of variant features
2146 * @param dnaToProtein
2148 public static int computeProteinFeatures(SequenceI dnaSeq,
2149 SequenceI peptide, MapList dnaToProtein)
2151 while (dnaSeq.getDatasetSequence() != null)
2153 dnaSeq = dnaSeq.getDatasetSequence();
2155 while (peptide.getDatasetSequence() != null)
2157 peptide = peptide.getDatasetSequence();
2160 transferFeatures(dnaSeq, peptide, dnaToProtein, SequenceOntologyI.EXON);
2163 * compute protein variants from dna variants and codon mappings;
2164 * NB - alternatively we could retrieve this using the REST service e.g.
2165 * http://rest.ensembl.org/overlap/translation
2166 * /ENSP00000288602?feature=transcript_variation;content-type=text/xml
2167 * which would be a bit slower but possibly more reliable
2171 * build a map with codon variations for each potentially varying peptide
2173 LinkedHashMap<Integer, List<DnaVariant>[]> variants = buildDnaVariantsMap(
2174 dnaSeq, dnaToProtein);
2177 * scan codon variations, compute peptide variants and add to peptide sequence
2180 for (Entry<Integer, List<DnaVariant>[]> variant : variants.entrySet())
2182 int peptidePos = variant.getKey();
2183 List<DnaVariant>[] codonVariants = variant.getValue();
2184 count += computePeptideVariants(peptide, peptidePos, codonVariants);
2188 * sort to get sequence features in start position order
2189 * - would be better to store in Sequence as a TreeSet or NCList?
2191 if (peptide.getSequenceFeatures() != null)
2193 Arrays.sort(peptide.getSequenceFeatures(),
2194 new Comparator<SequenceFeature>()
2197 public int compare(SequenceFeature o1, SequenceFeature o2)
2199 int c = Integer.compare(o1.getBegin(), o2.getBegin());
2200 return c == 0 ? Integer.compare(o1.getEnd(), o2.getEnd())
2209 * Computes non-synonymous peptide variants from codon variants and adds them
2210 * as sequence_variant features on the protein sequence (one feature per
2211 * allele variant). Selected attributes (variant id, clinical significance)
2212 * are copied over to the new features.
2215 * the protein sequence
2217 * the position to compute peptide variants for
2218 * @param codonVariants
2219 * a list of dna variants per codon position
2220 * @return the number of features added
2222 static int computePeptideVariants(SequenceI peptide, int peptidePos,
2223 List<DnaVariant>[] codonVariants)
2225 String residue = String.valueOf(peptide.getCharAt(peptidePos - 1));
2227 String base1 = codonVariants[0].get(0).base;
2228 String base2 = codonVariants[1].get(0).base;
2229 String base3 = codonVariants[2].get(0).base;
2232 * variants in first codon base
2234 for (DnaVariant var : codonVariants[0])
2236 if (var.variant != null)
2238 String alleles = (String) var.variant.getValue("alleles");
2239 if (alleles != null)
2241 for (String base : alleles.split(","))
2243 String codon = base + base2 + base3;
2244 if (addPeptideVariant(peptide, peptidePos, residue, var, codon))
2254 * variants in second codon base
2256 for (DnaVariant var : codonVariants[1])
2258 if (var.variant != null)
2260 String alleles = (String) var.variant.getValue("alleles");
2261 if (alleles != null)
2263 for (String base : alleles.split(","))
2265 String codon = base1 + base + base3;
2266 if (addPeptideVariant(peptide, peptidePos, residue, var, codon))
2276 * variants in third codon base
2278 for (DnaVariant var : codonVariants[2])
2280 if (var.variant != null)
2282 String alleles = (String) var.variant.getValue("alleles");
2283 if (alleles != null)
2285 for (String base : alleles.split(","))
2287 String codon = base1 + base2 + base;
2288 if (addPeptideVariant(peptide, peptidePos, residue, var, codon))
2301 * Helper method that adds a peptide variant feature, provided the given codon
2302 * translates to a value different to the current residue (is a non-synonymous
2303 * variant). ID and clinical_significance attributes of the dna variant (if
2304 * present) are copied to the new feature.
2311 * @return true if a feature was added, else false
2313 static boolean addPeptideVariant(SequenceI peptide, int peptidePos,
2314 String residue, DnaVariant var, String codon)
2317 * get peptide translation of codon e.g. GAT -> D
2318 * note that variants which are not single alleles,
2319 * e.g. multibase variants or HGMD_MUTATION etc
2320 * are currently ignored here
2322 String trans = codon.contains("-") ? "-"
2323 : (codon.length() > CODON_LENGTH ? null : ResidueProperties
2324 .codonTranslate(codon));
2325 if (trans != null && !trans.equals(residue))
2327 String residue3Char = StringUtils
2328 .toSentenceCase(ResidueProperties.aa2Triplet.get(residue));
2329 String trans3Char = StringUtils
2330 .toSentenceCase(ResidueProperties.aa2Triplet.get(trans));
2331 String desc = "p." + residue3Char + peptidePos + trans3Char;
2332 // set score to 0f so 'graduated colour' option is offered! JAL-2060
2333 SequenceFeature sf = new SequenceFeature(
2334 SequenceOntologyI.SEQUENCE_VARIANT, desc, peptidePos,
2335 peptidePos, 0f, var.getSource());
2336 StringBuilder attributes = new StringBuilder(32);
2337 String id = (String) var.variant.getValue(ID);
2340 if (id.startsWith(SEQUENCE_VARIANT))
2342 id = id.substring(SEQUENCE_VARIANT.length());
2344 sf.setValue(ID, id);
2345 attributes.append(ID).append("=").append(id);
2346 // TODO handle other species variants JAL-2064
2347 StringBuilder link = new StringBuilder(32);
2350 link.append(desc).append(" ").append(id)
2351 .append("|http://www.ensembl.org/Homo_sapiens/Variation/Summary?v=")
2352 .append(URLEncoder.encode(id, "UTF-8"));
2353 sf.addLink(link.toString());
2354 } catch (UnsupportedEncodingException e)
2359 String clinSig = (String) var.variant
2360 .getValue(CLINICAL_SIGNIFICANCE);
2361 if (clinSig != null)
2363 sf.setValue(CLINICAL_SIGNIFICANCE, clinSig);
2364 attributes.append(";").append(CLINICAL_SIGNIFICANCE).append("=")
2367 peptide.addSequenceFeature(sf);
2368 if (attributes.length() > 0)
2370 sf.setAttributes(attributes.toString());
2378 * Builds a map whose key is position in the protein sequence, and value is a
2379 * list of the base and all variants for each corresponding codon position
2382 * @param dnaToProtein
2385 @SuppressWarnings("unchecked")
2386 static LinkedHashMap<Integer, List<DnaVariant>[]> buildDnaVariantsMap(
2387 SequenceI dnaSeq, MapList dnaToProtein)
2390 * map from peptide position to all variants of the codon which codes for it
2391 * LinkedHashMap ensures we keep the peptide features in sequence order
2393 LinkedHashMap<Integer, List<DnaVariant>[]> variants = new LinkedHashMap<Integer, List<DnaVariant>[]>();
2394 SequenceOntologyI so = SequenceOntologyFactory.getInstance();
2396 SequenceFeature[] dnaFeatures = dnaSeq.getSequenceFeatures();
2397 if (dnaFeatures == null)
2402 int dnaStart = dnaSeq.getStart();
2403 int[] lastCodon = null;
2404 int lastPeptidePostion = 0;
2407 * build a map of codon variations for peptides
2409 for (SequenceFeature sf : dnaFeatures)
2411 int dnaCol = sf.getBegin();
2412 if (dnaCol != sf.getEnd())
2414 // not handling multi-locus variant features
2417 if (so.isA(sf.getType(), SequenceOntologyI.SEQUENCE_VARIANT))
2419 int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol);
2422 // feature doesn't lie within coding region
2425 int peptidePosition = mapsTo[0];
2426 List<DnaVariant>[] codonVariants = variants.get(peptidePosition);
2427 if (codonVariants == null)
2429 codonVariants = new ArrayList[CODON_LENGTH];
2430 codonVariants[0] = new ArrayList<DnaVariant>();
2431 codonVariants[1] = new ArrayList<DnaVariant>();
2432 codonVariants[2] = new ArrayList<DnaVariant>();
2433 variants.put(peptidePosition, codonVariants);
2437 * extract dna variants to a string array
2439 String alls = (String) sf.getValue("alleles");
2444 String[] alleles = alls.toUpperCase().split(",");
2446 for (String allele : alleles)
2448 alleles[i++] = allele.trim(); // lose any space characters "A, G"
2452 * get this peptide's codon positions e.g. [3, 4, 5] or [4, 7, 10]
2454 int[] codon = peptidePosition == lastPeptidePostion ? lastCodon
2455 : MappingUtils.flattenRanges(dnaToProtein.locateInFrom(
2456 peptidePosition, peptidePosition));
2457 lastPeptidePostion = peptidePosition;
2461 * save nucleotide (and any variant) for each codon position
2463 for (int codonPos = 0; codonPos < CODON_LENGTH; codonPos++)
2465 String nucleotide = String.valueOf(
2466 dnaSeq.getCharAt(codon[codonPos] - dnaStart))
2468 List<DnaVariant> codonVariant = codonVariants[codonPos];
2469 if (codon[codonPos] == dnaCol)
2471 if (!codonVariant.isEmpty()
2472 && codonVariant.get(0).variant == null)
2475 * already recorded base value, add this variant
2477 codonVariant.get(0).variant = sf;
2482 * add variant with base value
2484 codonVariant.add(new DnaVariant(nucleotide, sf));
2487 else if (codonVariant.isEmpty())
2490 * record (possibly non-varying) base value
2492 codonVariant.add(new DnaVariant(nucleotide));
2501 * Makes an alignment with a copy of the given sequences, adding in any
2502 * non-redundant sequences which are mapped to by the cross-referenced
2508 * the alignment dataset shared by the new copy
2511 public static AlignmentI makeCopyAlignment(SequenceI[] seqs,
2512 SequenceI[] xrefs, AlignmentI dataset)
2514 AlignmentI copy = new Alignment(new Alignment(seqs));
2515 copy.setDataset(dataset);
2517 SequenceIdMatcher matcher = new SequenceIdMatcher(seqs);
2520 for (SequenceI xref : xrefs)
2522 DBRefEntry[] dbrefs = xref.getDBRefs();
2525 for (DBRefEntry dbref : dbrefs)
2527 if (dbref.getMap() == null || dbref.getMap().getTo() == null)
2531 SequenceI mappedTo = dbref.getMap().getTo();
2532 SequenceI match = matcher.findIdMatch(mappedTo);
2535 matcher.add(mappedTo);
2536 copy.addSequence(mappedTo);
2546 * Try to align sequences in 'unaligned' to match the alignment of their
2547 * mapped regions in 'aligned'. For example, could use this to align CDS
2548 * sequences which are mapped to their parent cDNA sequences.
2550 * This method handles 1:1 mappings (dna-to-dna or protein-to-protein). For
2551 * dna-to-protein or protein-to-dna use alternative methods.
2554 * sequences to be aligned
2556 * holds aligned sequences and their mappings
2559 public static int alignAs(AlignmentI unaligned, AlignmentI aligned)
2562 * easy case - aligning a copy of aligned sequences
2564 if (alignAsSameSequences(unaligned, aligned))
2566 return unaligned.getHeight();
2570 * fancy case - aligning via mappings between sequences
2572 List<SequenceI> unmapped = new ArrayList<SequenceI>();
2573 Map<Integer, Map<SequenceI, Character>> columnMap = buildMappedColumnsMap(
2574 unaligned, aligned, unmapped);
2575 int width = columnMap.size();
2576 char gap = unaligned.getGapCharacter();
2577 int realignedCount = 0;
2578 // TODO: verify this loop scales sensibly for very wide/high alignments
2580 for (SequenceI seq : unaligned.getSequences())
2582 if (!unmapped.contains(seq))
2584 char[] newSeq = new char[width];
2585 Arrays.fill(newSeq, gap); // JBPComment - doubt this is faster than the
2586 // Integer iteration below
2591 * traverse the map to find columns populated
2594 for (Integer column : columnMap.keySet())
2596 Character c = columnMap.get(column).get(seq);
2600 * sequence has a character at this position
2610 * trim trailing gaps
2612 if (lastCol < width)
2614 char[] tmp = new char[lastCol + 1];
2615 System.arraycopy(newSeq, 0, tmp, 0, lastCol + 1);
2618 // TODO: optimise SequenceI to avoid char[]->String->char[]
2619 seq.setSequence(String.valueOf(newSeq));
2623 return realignedCount;
2627 * If unaligned and aligned sequences share the same dataset sequences, then
2628 * simply copies the aligned sequences to the unaligned sequences and returns
2629 * true; else returns false
2632 * - sequences to be aligned based on aligned
2634 * - 'guide' alignment containing sequences derived from same dataset
2638 static boolean alignAsSameSequences(AlignmentI unaligned,
2641 if (aligned.getDataset() == null || unaligned.getDataset() == null)
2643 return false; // should only pass alignments with datasets here
2646 // map from dataset sequence to alignment sequence(s)
2647 Map<SequenceI, List<SequenceI>> alignedDatasets = new HashMap<SequenceI, List<SequenceI>>();
2648 for (SequenceI seq : aligned.getSequences())
2650 SequenceI ds = seq.getDatasetSequence();
2651 if (alignedDatasets.get(ds) == null)
2653 alignedDatasets.put(ds, new ArrayList<SequenceI>());
2655 alignedDatasets.get(ds).add(seq);
2659 * first pass - check whether all sequences to be aligned share a dataset
2660 * sequence with an aligned sequence
2662 for (SequenceI seq : unaligned.getSequences())
2664 if (!alignedDatasets.containsKey(seq.getDatasetSequence()))
2671 * second pass - copy aligned sequences;
2672 * heuristic rule: pair off sequences in order for the case where
2673 * more than one shares the same dataset sequence
2675 for (SequenceI seq : unaligned.getSequences())
2677 List<SequenceI> alignedSequences = alignedDatasets.get(seq
2678 .getDatasetSequence());
2679 // TODO: getSequenceAsString() will be deprecated in the future
2680 // TODO: need to leave to SequenceI implementor to update gaps
2681 seq.setSequence(alignedSequences.get(0).getSequenceAsString());
2682 if (alignedSequences.size() > 0)
2684 // pop off aligned sequences (except the last one)
2685 alignedSequences.remove(0);
2693 * Returns a map whose key is alignment column number (base 1), and whose
2694 * values are a map of sequence characters in that column.
2701 static Map<Integer, Map<SequenceI, Character>> buildMappedColumnsMap(
2702 AlignmentI unaligned, AlignmentI aligned, List<SequenceI> unmapped)
2705 * Map will hold, for each aligned column position, a map of
2706 * {unalignedSequence, characterPerSequence} at that position.
2707 * TreeMap keeps the entries in ascending column order.
2709 Map<Integer, Map<SequenceI, Character>> map = new TreeMap<Integer, Map<SequenceI, Character>>();
2712 * record any sequences that have no mapping so can't be realigned
2714 unmapped.addAll(unaligned.getSequences());
2716 List<AlignedCodonFrame> mappings = aligned.getCodonFrames();
2718 for (SequenceI seq : unaligned.getSequences())
2720 for (AlignedCodonFrame mapping : mappings)
2722 SequenceI fromSeq = mapping.findAlignedSequence(seq, aligned);
2723 if (fromSeq != null)
2725 Mapping seqMap = mapping.getMappingBetween(fromSeq, seq);
2726 if (addMappedPositions(seq, fromSeq, seqMap, map))
2728 unmapped.remove(seq);
2737 * Helper method that adds to a map the mapped column positions of a sequence. <br>
2738 * For example if aaTT-Tg-gAAA is mapped to TTTAAA then the map should record
2739 * that columns 3,4,6,10,11,12 map to characters T,T,T,A,A,A of the mapped to
2743 * the sequence whose column positions we are recording
2745 * a sequence that is mapped to the first sequence
2747 * the mapping from 'fromSeq' to 'seq'
2749 * a map to add the column positions (in fromSeq) of the mapped
2753 static boolean addMappedPositions(SequenceI seq, SequenceI fromSeq,
2754 Mapping seqMap, Map<Integer, Map<SequenceI, Character>> map)
2762 * invert mapping if it is from unaligned to aligned sequence
2764 if (seqMap.getTo() == fromSeq.getDatasetSequence())
2766 seqMap = new Mapping(seq.getDatasetSequence(), seqMap.getMap()
2770 char[] fromChars = fromSeq.getSequence();
2771 int toStart = seq.getStart();
2772 char[] toChars = seq.getSequence();
2775 * traverse [start, end, start, end...] ranges in fromSeq
2777 for (int[] fromRange : seqMap.getMap().getFromRanges())
2779 for (int i = 0; i < fromRange.length - 1; i += 2)
2781 boolean forward = fromRange[i + 1] >= fromRange[i];
2784 * find the range mapped to (sequence positions base 1)
2786 int[] range = seqMap.locateMappedRange(fromRange[i],
2790 System.err.println("Error in mapping " + seqMap + " from "
2791 + fromSeq.getName());
2794 int fromCol = fromSeq.findIndex(fromRange[i]);
2795 int mappedCharPos = range[0];
2798 * walk over the 'from' aligned sequence in forward or reverse
2799 * direction; when a non-gap is found, record the column position
2800 * of the next character of the mapped-to sequence; stop when all
2801 * the characters of the range have been counted
2803 while (mappedCharPos <= range[1] && fromCol <= fromChars.length
2806 if (!Comparison.isGap(fromChars[fromCol - 1]))
2809 * mapped from sequence has a character in this column
2810 * record the column position for the mapped to character
2812 Map<SequenceI, Character> seqsMap = map.get(fromCol);
2813 if (seqsMap == null)
2815 seqsMap = new HashMap<SequenceI, Character>();
2816 map.put(fromCol, seqsMap);
2818 seqsMap.put(seq, toChars[mappedCharPos - toStart]);
2821 fromCol += (forward ? 1 : -1);
2828 // strictly temporary hack until proper criteria for aligning protein to cds
2829 // are in place; this is so Ensembl -> fetch xrefs Uniprot aligns the Uniprot
2830 public static boolean looksLikeEnsembl(AlignmentI alignment)
2832 for (SequenceI seq : alignment.getSequences())
2834 String name = seq.getName();
2835 if (!name.startsWith("ENSG") && !name.startsWith("ENST"))