X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FAlignmentUtils.java;h=0bc81800addb62140ac28047f278c509c4a971cf;hb=cbeb7ad59d51b468c54ca3db2a2a7693060a2509;hp=37f225c02aa70867c3041652349bc6914a2dc797;hpb=2c2992b8822b686f7f12022a66aa41f6969d06d0;p=jalview.git diff --git a/src/jalview/analysis/AlignmentUtils.java b/src/jalview/analysis/AlignmentUtils.java index 37f225c..0bc8180 100644 --- a/src/jalview/analysis/AlignmentUtils.java +++ b/src/jalview/analysis/AlignmentUtils.java @@ -22,7 +22,6 @@ package jalview.analysis; import static jalview.io.gff.GffConstants.CLINICAL_SIGNIFICANCE; -import jalview.api.DBRefEntryI; import jalview.datamodel.AlignedCodon; import jalview.datamodel.AlignedCodonFrame; import jalview.datamodel.AlignedCodonFrame.SequenceToSequenceMapping; @@ -30,17 +29,20 @@ import jalview.datamodel.Alignment; import jalview.datamodel.AlignmentAnnotation; import jalview.datamodel.AlignmentI; import jalview.datamodel.DBRefEntry; +import jalview.datamodel.GeneLociI; import jalview.datamodel.IncompleteCodonException; import jalview.datamodel.Mapping; import jalview.datamodel.Sequence; import jalview.datamodel.SequenceFeature; import jalview.datamodel.SequenceGroup; import jalview.datamodel.SequenceI; -import jalview.io.gff.SequenceOntologyFactory; +import jalview.datamodel.features.SequenceFeatures; +import jalview.io.gff.Gff3Helper; import jalview.io.gff.SequenceOntologyI; import jalview.schemes.ResidueProperties; import jalview.util.Comparison; import jalview.util.DBRefUtils; +import jalview.util.IntRangeComparator; import jalview.util.MapList; import jalview.util.MappingUtils; import jalview.util.StringUtils; @@ -51,7 +53,6 @@ import java.util.ArrayList; import java.util.Arrays; import java.util.Collection; import java.util.Collections; -import java.util.Comparator; import java.util.HashMap; import java.util.HashSet; import java.util.Iterator; @@ -61,6 +62,7 @@ import java.util.Map; import java.util.Map.Entry; import java.util.NoSuchElementException; import java.util.Set; +import java.util.SortedMap; import java.util.TreeMap; /** @@ -72,23 +74,30 @@ import java.util.TreeMap; */ public class AlignmentUtils { + private static final int CODON_LENGTH = 3; private static final String SEQUENCE_VARIANT = "sequence_variant:"; - private static final String ID = "ID"; + + /* + * the 'id' attribute is provided for variant features fetched from + * Ensembl using its REST service with JSON format + */ + public static final String VARIANT_ID = "id"; /** * A data model to hold the 'normal' base value at a position, and an optional * sequence variant feature */ - static class DnaVariant + static final class DnaVariant { - String base; + final String base; SequenceFeature variant; DnaVariant(String nuc) { base = nuc; + variant = null; } DnaVariant(String nuc, SequenceFeature var) @@ -96,6 +105,20 @@ public class AlignmentUtils base = nuc; variant = var; } + + public String getSource() + { + return variant == null ? null : variant.getFeatureGroup(); + } + + /** + * toString for aid in the debugger only + */ + @Override + public String toString() + { + return base + ":" + (variant == null ? "" : variant.getDescription()); + } } /** @@ -108,7 +131,7 @@ public class AlignmentUtils */ public static AlignmentI expandContext(AlignmentI core, int flankSize) { - List sq = new ArrayList(); + List sq = new ArrayList<>(); int maxoffset = 0; for (SequenceI s : core.getSequences()) { @@ -161,10 +184,12 @@ public class AlignmentUtils } } // TODO use Character.toLowerCase to avoid creating String objects? - char[] upstream = new String(ds.getSequence(s.getStart() - 1 - - ustream_ds, s.getStart() - 1)).toLowerCase().toCharArray(); - char[] downstream = new String(ds.getSequence(s_end - 1, s_end - + dstream_ds)).toLowerCase().toCharArray(); + char[] upstream = new String(ds + .getSequence(s.getStart() - 1 - ustream_ds, s.getStart() - 1)) + .toLowerCase().toCharArray(); + char[] downstream = new String( + ds.getSequence(s_end - 1, s_end + dstream_ds)).toLowerCase() + .toCharArray(); char[] coreseq = s.getSequence(); char[] nseq = new char[offset + upstream.length + downstream.length + coreseq.length]; @@ -179,8 +204,8 @@ public class AlignmentUtils System.arraycopy(upstream, 0, nseq, p, upstream.length); System.arraycopy(coreseq, 0, nseq, p + upstream.length, coreseq.length); - System.arraycopy(downstream, 0, nseq, p + coreseq.length - + upstream.length, downstream.length); + System.arraycopy(downstream, 0, nseq, + p + coreseq.length + upstream.length, downstream.length); s.setSequence(new String(nseq)); s.setStart(s.getStart() - ustream_ds); s.setEnd(s_end + downstream.length); @@ -236,7 +261,7 @@ public class AlignmentUtils public static Map> getSequencesByName( AlignmentI al) { - Map> theMap = new LinkedHashMap>(); + Map> theMap = new LinkedHashMap<>(); for (SequenceI seq : al.getSequences()) { String name = seq.getName(); @@ -245,7 +270,7 @@ public class AlignmentUtils List seqs = theMap.get(name); if (seqs == null) { - seqs = new ArrayList(); + seqs = new ArrayList<>(); theMap.put(name, seqs); } seqs.add(seq); @@ -272,8 +297,8 @@ public class AlignmentUtils return false; } - Set mappedDna = new HashSet(); - Set mappedProtein = new HashSet(); + Set mappedDna = new HashSet<>(); + Set mappedProtein = new HashSet<>(); /* * First pass - map sequences where cross-references exist. This include @@ -307,9 +332,9 @@ public class AlignmentUtils * @return */ protected static boolean mapProteinToCdna( - final AlignmentI proteinAlignment, - final AlignmentI cdnaAlignment, Set mappedDna, - Set mappedProtein, boolean xrefsOnly) + final AlignmentI proteinAlignment, final AlignmentI cdnaAlignment, + Set mappedDna, Set mappedProtein, + boolean xrefsOnly) { boolean mappingExistsOrAdded = false; List thisSeqs = proteinAlignment.getSequences(); @@ -338,9 +363,8 @@ public class AlignmentUtils * Don't map non-xrefd sequences more than once each. This heuristic * allows us to pair up similar sequences in ordered alignments. */ - if (!xrefsOnly - && (mappedProtein.contains(aaSeq) || mappedDna - .contains(cdnaSeq))) + if (!xrefsOnly && (mappedProtein.contains(aaSeq) + || mappedDna.contains(cdnaSeq))) { continue; } @@ -374,7 +398,7 @@ public class AlignmentUtils * Answers true if the mappings include one between the given (dataset) * sequences. */ - public static boolean mappingExists(List mappings, + protected static boolean mappingExists(List mappings, SequenceI aaSeq, SequenceI cdnaSeq) { if (mappings != null) @@ -393,7 +417,8 @@ public class AlignmentUtils /** * Builds a mapping (if possible) of a cDNA to a protein sequence. *
    - *
  • first checks if the cdna translates exactly to the protein sequence
    • + *
  • first checks if the cdna translates exactly to the protein + * sequence
    • *
  • else checks for translation after removing a STOP codon
    • *
  • else checks for translation after removing a START codon
    • *
  • if that fails, inspect CDS features on the cDNA sequence
    • @@ -415,8 +440,9 @@ public class AlignmentUtils * String objects. */ final SequenceI proteinDataset = proteinSeq.getDatasetSequence(); - char[] aaSeqChars = proteinDataset != null ? proteinDataset - .getSequence() : proteinSeq.getSequence(); + char[] aaSeqChars = proteinDataset != null + ? proteinDataset.getSequence() + : proteinSeq.getSequence(); final SequenceI cdnaDataset = cdnaSeq.getDatasetSequence(); char[] cdnaSeqChars = cdnaDataset != null ? cdnaDataset.getSequence() : cdnaSeq.getSequence(); @@ -428,7 +454,7 @@ public class AlignmentUtils /* * cdnaStart/End, proteinStartEnd are base 1 (for dataset sequence mapping) */ - final int mappedLength = 3 * aaSeqChars.length; + final int mappedLength = CODON_LENGTH * aaSeqChars.length; int cdnaLength = cdnaSeqChars.length; int cdnaStart = cdnaSeq.getStart(); int cdnaEnd = cdnaSeq.getEnd(); @@ -440,14 +466,14 @@ public class AlignmentUtils */ if (cdnaLength != mappedLength && cdnaLength > 2) { - String lastCodon = String.valueOf(cdnaSeqChars, cdnaLength - 3, 3) - .toUpperCase(); - for (String stop : ResidueProperties.STOP) + String lastCodon = String.valueOf(cdnaSeqChars, + cdnaLength - CODON_LENGTH, CODON_LENGTH).toUpperCase(); + for (String stop : ResidueProperties.STOP_CODONS) { if (lastCodon.equals(stop)) { - cdnaEnd -= 3; - cdnaLength -= 3; + cdnaEnd -= CODON_LENGTH; + cdnaLength -= CODON_LENGTH; break; } } @@ -457,14 +483,13 @@ public class AlignmentUtils * If lengths still don't match, try ignoring start codon. */ int startOffset = 0; - if (cdnaLength != mappedLength - && cdnaLength > 2 - && String.valueOf(cdnaSeqChars, 0, 3).toUpperCase() + if (cdnaLength != mappedLength && cdnaLength > 2 + && String.valueOf(cdnaSeqChars, 0, CODON_LENGTH).toUpperCase() .equals(ResidueProperties.START)) { - startOffset += 3; - cdnaStart += 3; - cdnaLength -= 3; + startOffset += CODON_LENGTH; + cdnaStart += CODON_LENGTH; + cdnaLength -= CODON_LENGTH; } if (translatesAs(cdnaSeqChars, startOffset, aaSeqChars)) @@ -472,8 +497,9 @@ public class AlignmentUtils /* * protein is translation of dna (+/- start/stop codons) */ - MapList map = new MapList(new int[] { cdnaStart, cdnaEnd }, new int[] - { proteinStart, proteinEnd }, 3, 1); + MapList map = new MapList(new int[] { cdnaStart, cdnaEnd }, + new int[] + { proteinStart, proteinEnd }, CODON_LENGTH, 1); return map; } @@ -504,16 +530,17 @@ public class AlignmentUtils int aaPos = 0; int dnaPos = cdnaStart; for (; dnaPos < cdnaSeqChars.length - 2 - && aaPos < aaSeqChars.length; dnaPos += 3, aaPos++) + && aaPos < aaSeqChars.length; dnaPos += CODON_LENGTH, aaPos++) { - String codon = String.valueOf(cdnaSeqChars, dnaPos, 3); + String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH); final String translated = ResidueProperties.codonTranslate(codon); /* * allow * in protein to match untranslatable in dna */ final char aaRes = aaSeqChars[aaPos]; - if ((translated == null || "STOP".equals(translated)) && aaRes == '*') + if ((translated == null || ResidueProperties.STOP.equals(translated)) + && aaRes == '*') { continue; } @@ -542,10 +569,11 @@ public class AlignmentUtils { return true; } - if (dnaPos == cdnaSeqChars.length - 3) + if (dnaPos == cdnaSeqChars.length - CODON_LENGTH) { - String codon = String.valueOf(cdnaSeqChars, dnaPos, 3); - if ("STOP".equals(ResidueProperties.codonTranslate(codon))) + String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH); + if (ResidueProperties.STOP + .equals(ResidueProperties.codonTranslate(codon))) { return true; } @@ -623,10 +651,9 @@ public class AlignmentUtils * @param preserveUnmappedGaps * @param preserveMappedGaps */ - public static void alignSequenceAs(SequenceI alignTo, - SequenceI alignFrom, AlignedCodonFrame mapping, String myGap, - char sourceGap, boolean preserveMappedGaps, - boolean preserveUnmappedGaps) + public static void alignSequenceAs(SequenceI alignTo, SequenceI alignFrom, + AlignedCodonFrame mapping, String myGap, char sourceGap, + boolean preserveMappedGaps, boolean preserveUnmappedGaps) { // TODO generalise to work for Protein-Protein, dna-dna, dna-protein @@ -642,15 +669,16 @@ public class AlignmentUtils int toOffset = alignTo.getStart() - 1; int sourceGapMappedLength = 0; boolean inExon = false; - final char[] thisSeq = alignTo.getSequence(); - final char[] thatAligned = alignFrom.getSequence(); - StringBuilder thisAligned = new StringBuilder(2 * thisSeq.length); + final int toLength = alignTo.getLength(); + final int fromLength = alignFrom.getLength(); + StringBuilder thisAligned = new StringBuilder(2 * toLength); /* * Traverse the 'model' aligned sequence */ - for (char sourceChar : thatAligned) + for (int i = 0; i < fromLength; i++) { + char sourceChar = alignFrom.getCharAt(i); if (sourceChar == sourceGap) { sourceGapMappedLength += ratio; @@ -690,9 +718,9 @@ public class AlignmentUtils */ int intronLength = 0; while (basesWritten + toOffset < mappedCodonEnd - && thisSeqPos < thisSeq.length) + && thisSeqPos < toLength) { - final char c = thisSeq[thisSeqPos++]; + final char c = alignTo.getCharAt(thisSeqPos++); if (c != myGapChar) { basesWritten++; @@ -718,7 +746,7 @@ public class AlignmentUtils int gapsToAdd = calculateGapsToInsert(preserveMappedGaps, preserveUnmappedGaps, sourceGapMappedLength, inExon, trailingCopiedGap.length(), intronLength, startOfCodon); - for (int i = 0; i < gapsToAdd; i++) + for (int k = 0; k < gapsToAdd; k++) { thisAligned.append(myGapChar); } @@ -746,9 +774,9 @@ public class AlignmentUtils * At end of model aligned sequence. Copy any remaining target sequence, optionally * including (intron) gaps. */ - while (thisSeqPos < thisSeq.length) + while (thisSeqPos < toLength) { - final char c = thisSeq[thisSeqPos++]; + final char c = alignTo.getCharAt(thisSeqPos++); if (c != myGapChar || preserveUnmappedGaps) { thisAligned.append(c); @@ -821,8 +849,9 @@ public class AlignmentUtils } else { - gapsToAdd = Math.min(intronLength + trailingGapLength - - sourceGapMappedLength, trailingGapLength); + gapsToAdd = Math.min( + intronLength + trailingGapLength - sourceGapMappedLength, + trailingGapLength); } } } @@ -857,7 +886,7 @@ public class AlignmentUtils System.err.println("Wrong alignment type in alignProteinAsDna"); return 0; } - List unmappedProtein = new ArrayList(); + List unmappedProtein = new ArrayList<>(); Map> alignedCodons = buildCodonColumnsMap( protein, dna, unmappedProtein); return alignProteinAs(protein, alignedCodons, unmappedProtein); @@ -867,6 +896,8 @@ public class AlignmentUtils * Realigns the given dna to match the alignment of the protein, using codon * mappings to translate aligned peptide positions to codons. * + * Always produces a padded CDS alignment. + * * @param dna * the alignment whose sequences are realigned by this method * @param protein @@ -883,6 +914,7 @@ public class AlignmentUtils // todo: implement this List mappings = protein.getCodonFrames(); int alignedCount = 0; + int width = 0; // alignment width for padding CDS for (SequenceI dnaSeq : dna.getSequences()) { if (alignCdsSequenceAsProtein(dnaSeq, protein, mappings, @@ -890,6 +922,18 @@ public class AlignmentUtils { alignedCount++; } + width = Math.max(dnaSeq.getLength(), width); + } + int oldwidth; + int diff; + for (SequenceI dnaSeq : dna.getSequences()) + { + oldwidth = dnaSeq.getLength(); + diff = width - oldwidth; + if (diff > 0) + { + dnaSeq.insertCharAt(oldwidth, diff, dna.getGapCharacter()); + } } return alignedCount; } @@ -906,7 +950,8 @@ public class AlignmentUtils * @return */ static boolean alignCdsSequenceAsProtein(SequenceI cdsSeq, - AlignmentI protein, List mappings, char gapChar) + AlignmentI protein, List mappings, + char gapChar) { SequenceI cdsDss = cdsSeq.getDatasetSequence(); if (cdsDss == null) @@ -915,15 +960,15 @@ public class AlignmentUtils .println("alignCdsSequenceAsProtein needs aligned sequence!"); return false; } - + List dnaMappings = MappingUtils .findMappingsForSequence(cdsSeq, mappings); for (AlignedCodonFrame mapping : dnaMappings) { SequenceI peptide = mapping.findAlignedSequence(cdsSeq, protein); - int peptideLength = peptide.getLength(); if (peptide != null) { + final int peptideLength = peptide.getLength(); Mapping map = mapping.getMappingBetween(cdsSeq, peptide); if (map != null) { @@ -932,44 +977,45 @@ public class AlignmentUtils { mapList = mapList.getInverse(); } - int cdsLength = cdsDss.getLength(); + final int cdsLength = cdsDss.getLength(); int mappedFromLength = MappingUtils.getLength(mapList .getFromRanges()); int mappedToLength = MappingUtils .getLength(mapList.getToRanges()); - boolean addStopCodon = (cdsLength == mappedFromLength * 3 + 3) - || (peptide.getDatasetSequence().getLength() == mappedFromLength - 1); + boolean addStopCodon = (cdsLength == mappedFromLength + * CODON_LENGTH + CODON_LENGTH) + || (peptide.getDatasetSequence() + .getLength() == mappedFromLength - 1); if (cdsLength != mappedToLength && !addStopCodon) { - System.err - .println(String - .format("Can't align cds as protein (length mismatch %d/%d): %s", - cdsLength, mappedToLength, - cdsSeq.getName())); + System.err.println(String.format( + "Can't align cds as protein (length mismatch %d/%d): %s", + cdsLength, mappedToLength, cdsSeq.getName())); } /* * pre-fill the aligned cds sequence with gaps */ - char[] alignedCds = new char[peptideLength * 3 - + (addStopCodon ? 3 : 0)]; + char[] alignedCds = new char[peptideLength * CODON_LENGTH + + (addStopCodon ? CODON_LENGTH : 0)]; Arrays.fill(alignedCds, gapChar); /* * walk over the aligned peptide sequence and insert mapped * codons for residues in the aligned cds sequence */ - char[] alignedPeptide = peptide.getSequence(); - char[] nucleotides = cdsDss.getSequence(); int copiedBases = 0; int cdsStart = cdsDss.getStart(); int proteinPos = peptide.getStart() - 1; int cdsCol = 0; - for (char residue : alignedPeptide) + + for (int col = 0; col < peptideLength; col++) { + char residue = peptide.getCharAt(col); + if (Comparison.isGap(residue)) { - cdsCol += 3; + cdsCol += CODON_LENGTH; } else { @@ -978,13 +1024,13 @@ public class AlignmentUtils if (codon == null) { // e.g. incomplete start codon, X in peptide - cdsCol += 3; + cdsCol += CODON_LENGTH; } else { for (int j = codon[0]; j <= codon[1]; j++) { - char mappedBase = nucleotides[j - cdsStart]; + char mappedBase = cdsDss.getCharAt(j - cdsStart); alignedCds[cdsCol++] = mappedBase; copiedBases++; } @@ -996,7 +1042,7 @@ public class AlignmentUtils * append stop codon if not mapped from protein, * closing it up to the end of the mapped sequence */ - if (copiedBases == nucleotides.length - 3) + if (copiedBases == cdsLength - CODON_LENGTH) { for (int i = alignedCds.length - 1; i >= 0; i--) { @@ -1006,9 +1052,9 @@ public class AlignmentUtils break; } } - for (int i = nucleotides.length - 3; i < nucleotides.length; i++) + for (int i = cdsLength - CODON_LENGTH; i < cdsLength; i++) { - alignedCds[cdsCol++] = nucleotides[i]; + alignedCds[cdsCol++] = cdsDss.getCharAt(i); } } cdsSeq.setSequence(new String(alignedCds)); @@ -1051,7 +1097,7 @@ public class AlignmentUtils * {dnaSequence, {proteinSequence, codonProduct}} at that position. The * comparator keeps the codon positions ordered. */ - Map> alignedCodons = new TreeMap>( + Map> alignedCodons = new TreeMap<>( new CodonComparator()); for (SequenceI dnaSeq : dna.getSequences()) @@ -1062,8 +1108,8 @@ public class AlignmentUtils if (prot != null) { Mapping seqMap = mapping.getMappingForSequence(dnaSeq); - addCodonPositions(dnaSeq, prot, protein.getGapCharacter(), - seqMap, alignedCodons); + addCodonPositions(dnaSeq, prot, protein.getGapCharacter(), seqMap, + alignedCodons); unmappedProtein.remove(prot); } } @@ -1076,7 +1122,7 @@ public class AlignmentUtils // TODO resolve JAL-2022 so this fudge can be removed int mappedSequenceCount = protein.getHeight() - unmappedProtein.size(); addUnmappedPeptideStarts(alignedCodons, mappedSequenceCount); - + return alignedCodons; } @@ -1097,9 +1143,9 @@ public class AlignmentUtils // TODO delete this ugly hack once JAL-2022 is resolved // i.e. we can model startPhase > 0 (incomplete start codon) - List sequencesChecked = new ArrayList(); + List sequencesChecked = new ArrayList<>(); AlignedCodon lastCodon = null; - Map toAdd = new HashMap(); + Map toAdd = new HashMap<>(); for (Entry> entry : alignedCodons .entrySet()) @@ -1116,8 +1162,8 @@ public class AlignmentUtils AlignedCodon codon = sequenceCodon.getValue(); if (codon.peptideCol > 1) { - System.err - .println("Problem mapping protein with >1 unmapped start positions: " + System.err.println( + "Problem mapping protein with >1 unmapped start positions: " + seq.getName()); } else if (codon.peptideCol == 1) @@ -1128,8 +1174,8 @@ public class AlignmentUtils if (lastCodon != null) { AlignedCodon firstPeptide = new AlignedCodon(lastCodon.pos1, - lastCodon.pos2, lastCodon.pos3, String.valueOf(seq - .getCharAt(0)), 0); + lastCodon.pos2, lastCodon.pos3, + String.valueOf(seq.getCharAt(0)), 0); toAdd.put(seq, firstPeptide); } else @@ -1178,21 +1224,26 @@ public class AlignmentUtils List unmappedProtein) { /* - * Prefill aligned sequences with gaps before inserting aligned protein - * residues. + * prefill peptide sequences with gaps */ int alignedWidth = alignedCodons.size(); char[] gaps = new char[alignedWidth]; Arrays.fill(gaps, protein.getGapCharacter()); - String allGaps = String.valueOf(gaps); + Map peptides = new HashMap<>(); for (SequenceI seq : protein.getSequences()) { if (!unmappedProtein.contains(seq)) { - seq.setSequence(allGaps); + peptides.put(seq, Arrays.copyOf(gaps, gaps.length)); } } + /* + * Traverse the codons left to right (as defined by CodonComparator) + * and insert peptides in each column where the sequence is mapped. + * This gives a peptide 'alignment' where residues are aligned if their + * corresponding codons occupy the same columns in the cdna alignment. + */ int column = 0; for (AlignedCodon codon : alignedCodons.keySet()) { @@ -1200,12 +1251,20 @@ public class AlignmentUtils .get(codon); for (Entry entry : columnResidues.entrySet()) { - // place translated codon at its column position in sequence - entry.getKey().getSequence()[column] = entry.getValue().product - .charAt(0); + char residue = entry.getValue().product.charAt(0); + peptides.get(entry.getKey())[column] = residue; } column++; } + + /* + * and finally set the constructed sequences + */ + for (Entry entry : peptides.entrySet()) + { + entry.getKey().setSequence(new String(entry.getValue())); + } + return 0; } @@ -1265,7 +1324,7 @@ public class AlignmentUtils Map seqProduct = alignedCodons.get(codon); if (seqProduct == null) { - seqProduct = new HashMap(); + seqProduct = new HashMap<>(); alignedCodons.put(codon, seqProduct); } seqProduct.put(protein, codon); @@ -1278,7 +1337,8 @@ public class AlignmentUtils *
      *
    • One alignment must be nucleotide, and the other protein
      • *
    • At least one pair of sequences must be already mapped, or mappable
      • - *
    • Mappable means the nucleotide translation matches the protein sequence
      • + *
    • Mappable means the nucleotide translation matches the protein + * sequence
      • *
    • The translation may ignore start and stop codons if present in the * nucleotide
      • *
    @@ -1334,9 +1394,10 @@ public class AlignmentUtils return false; } - SequenceI dnaDs = dnaSeq.getDatasetSequence() == null ? dnaSeq : dnaSeq - .getDatasetSequence(); - SequenceI proteinDs = proteinSeq.getDatasetSequence() == null ? proteinSeq + SequenceI dnaDs = dnaSeq.getDatasetSequence() == null ? dnaSeq + : dnaSeq.getDatasetSequence(); + SequenceI proteinDs = proteinSeq.getDatasetSequence() == null + ? proteinSeq : proteinSeq.getDatasetSequence(); for (AlignedCodonFrame mapping : mappings) @@ -1373,8 +1434,7 @@ public class AlignmentUtils * the alignment to check for presence of annotations */ public static void findAddableReferenceAnnotations( - List sequenceScope, - Map labelForCalcId, + List sequenceScope, Map labelForCalcId, final Map> candidates, AlignmentI al) { @@ -1401,7 +1461,7 @@ public class AlignmentUtils { continue; } - final List result = new ArrayList(); + final List result = new ArrayList<>(); for (AlignmentAnnotation dsann : datasetAnnotations) { /* @@ -1478,8 +1538,8 @@ public class AlignmentUtils /** * Set visibility of alignment annotations of specified types (labels), for - * specified sequences. This supports controls like - * "Show all secondary structure", "Hide all Temp factor", etc. + * specified sequences. This supports controls like "Show all secondary + * structure", "Hide all Temp factor", etc. * * @al the alignment to scan for annotations * @param types @@ -1503,9 +1563,8 @@ public class AlignmentUtils { if (anyType || types.contains(aa.label)) { - if ((aa.sequenceRef != null) - && (forSequences == null || forSequences - .contains(aa.sequenceRef))) + if ((aa.sequenceRef != null) && (forSequences == null + || forSequences.contains(aa.sequenceRef))) { aa.visible = doShow; } @@ -1584,13 +1643,13 @@ public class AlignmentUtils throw new IllegalArgumentException( "IMPLEMENTATION ERROR: dataset.getDataset() must be null!"); } - List foundSeqs = new ArrayList(); - List cdsSeqs = new ArrayList(); + List foundSeqs = new ArrayList<>(); + List cdsSeqs = new ArrayList<>(); List mappings = dataset.getCodonFrames(); HashSet productSeqs = null; if (products != null) { - productSeqs = new HashSet(); + productSeqs = new HashSet<>(); for (SequenceI seq : products) { productSeqs.add(seq.getDatasetSequence() == null ? seq : seq @@ -1667,11 +1726,20 @@ public class AlignmentUtils * didn't find mapped CDS sequence - construct it and add * its dataset sequence to the dataset */ - cdsSeq = makeCdsSequence(dnaSeq.getDatasetSequence(), aMapping); - SequenceI cdsSeqDss = cdsSeq.createDatasetSequence(); + cdsSeq = makeCdsSequence(dnaSeq.getDatasetSequence(), aMapping, + dataset).deriveSequence(); + // cdsSeq has a name constructed as CDS| + // will be either the accession for the coding sequence, + // marked in the /via/ dbref to the protein product accession + // or it will be the original nucleotide accession. + SequenceI cdsSeqDss = cdsSeq.getDatasetSequence(); + cdsSeqs.add(cdsSeq); + if (!dataset.getSequences().contains(cdsSeqDss)) { + // check if this sequence is a newly created one + // so needs adding to the dataset dataset.addSequence(cdsSeqDss); } @@ -1680,10 +1748,12 @@ public class AlignmentUtils */ List cdsRange = Collections.singletonList(new int[] { 1, cdsSeq.getLength() }); - MapList cdsToProteinMap = new MapList(cdsRange, mapList.getToRanges(), - mapList.getFromRatio(), mapList.getToRatio()); + MapList cdsToProteinMap = new MapList(cdsRange, + mapList.getToRanges(), mapList.getFromRatio(), + mapList.getToRatio()); AlignedCodonFrame cdsToProteinMapping = new AlignedCodonFrame(); - cdsToProteinMapping.addMap(cdsSeq, proteinProduct, cdsToProteinMap); + cdsToProteinMapping.addMap(cdsSeqDss, proteinProduct, + cdsToProteinMap); /* * guard against duplicating the mapping if repeating this action @@ -1693,31 +1763,15 @@ public class AlignmentUtils mappings.add(cdsToProteinMapping); } - /* - * copy protein's dbrefs to CDS sequence - * this enables Get Cross-References from CDS alignment - */ - DBRefEntry[] proteinRefs = DBRefUtils.selectDbRefs(false, - proteinProduct.getDBRefs()); - if (proteinRefs != null) - { - for (DBRefEntry ref : proteinRefs) - { - DBRefEntry cdsToProteinRef = new DBRefEntry(ref); - cdsToProteinRef.setMap(new Mapping(proteinProduct, - cdsToProteinMap)); - cdsSeqDss.addDBRef(cdsToProteinRef); - } - } - + propagateDBRefsToCDS(cdsSeqDss, dnaSeq.getDatasetSequence(), + proteinProduct, aMapping); /* * add another mapping from original 'from' range to CDS */ AlignedCodonFrame dnaToCdsMapping = new AlignedCodonFrame(); - MapList dnaToCdsMap = new MapList(mapList.getFromRanges(), - cdsRange, 1, - 1); - dnaToCdsMapping.addMap(dnaSeq.getDatasetSequence(), cdsSeq, + final MapList dnaToCdsMap = new MapList(mapList.getFromRanges(), + cdsRange, 1, 1); + dnaToCdsMapping.addMap(dnaSeq.getDatasetSequence(), cdsSeqDss, dnaToCdsMap); if (!mappings.contains(dnaToCdsMapping)) { @@ -1725,18 +1779,53 @@ public class AlignmentUtils } /* + * transfer dna chromosomal loci (if known) to the CDS + * sequence (via the mapping) + */ + final MapList cdsToDnaMap = dnaToCdsMap.getInverse(); + transferGeneLoci(dnaSeq, cdsToDnaMap, cdsSeq); + + /* * add DBRef with mapping from protein to CDS * (this enables Get Cross-References from protein alignment) * This is tricky because we can't have two DBRefs with the * same source and accession, so need a different accession for * the CDS from the dna sequence */ - DBRefEntryI dnaRef = dnaDss.getSourceDBRef(); - if (dnaRef != null) + + // specific use case: + // Genomic contig ENSCHR:1, contains coding regions for ENSG01, + // ENSG02, ENSG03, with transcripts and products similarly named. + // cannot add distinct dbrefs mapping location on ENSCHR:1 to ENSG01 + + // JBPNote: ?? can't actually create an example that demonstrates we + // need to + // synthesize an xref. + + for (DBRefEntry primRef : dnaDss.getPrimaryDBRefs()) { + /* + * create a cross-reference from CDS to the source sequence's + * primary reference and vice versa + */ + String source = primRef.getSource(); + String version = primRef.getVersion(); + DBRefEntry cdsCrossRef = new DBRefEntry(source, source + ":" + + version, primRef.getAccessionId()); + cdsCrossRef.setMap(new Mapping(dnaDss, new MapList(cdsToDnaMap))); + cdsSeqDss.addDBRef(cdsCrossRef); + + dnaSeq.addDBRef(new DBRefEntry(source, version, cdsSeq + .getName(), new Mapping(cdsSeqDss, dnaToCdsMap))); + + // problem here is that the cross-reference is synthesized - + // cdsSeq.getName() may be like 'CDS|dnaaccession' or + // 'CDS|emblcdsacc' // assuming cds version same as dna ?!? - DBRefEntry proteinToCdsRef = new DBRefEntry(dnaRef.getSource(), - dnaRef.getVersion(), cdsSeq.getName()); + + DBRefEntry proteinToCdsRef = new DBRefEntry(source, version, + cdsSeq.getName()); + // proteinToCdsRef.setMap(new Mapping(cdsSeqDss, cdsToProteinMap .getInverse())); proteinProduct.addDBRef(proteinToCdsRef); @@ -1745,7 +1834,7 @@ public class AlignmentUtils /* * transfer any features on dna that overlap the CDS */ - transferFeatures(dnaSeq, cdsSeq, cdsToProteinMap, null, + transferFeatures(dnaSeq, cdsSeq, dnaToCdsMap, null, SequenceOntologyI.CDS); } } @@ -1759,6 +1848,38 @@ public class AlignmentUtils } /** + * Tries to transfer gene loci (dbref to chromosome positions) from fromSeq to + * toSeq, mediated by the given mapping between the sequences + * + * @param fromSeq + * @param targetToFrom + * Map + * @param targetSeq + */ + protected static void transferGeneLoci(SequenceI fromSeq, + MapList targetToFrom, SequenceI targetSeq) + { + if (targetSeq.getGeneLoci() != null) + { + // already have - don't override + return; + } + GeneLociI fromLoci = fromSeq.getGeneLoci(); + if (fromLoci == null) + { + return; + } + + MapList newMap = targetToFrom.traverse(fromLoci.getMapping()); + + if (newMap != null) + { + targetSeq.setGeneLoci(fromLoci.getSpeciesId(), + fromLoci.getAssemblyId(), fromLoci.getChromosomeId(), newMap); + } + } + + /** * A helper method that finds a CDS sequence in the alignment dataset that is * mapped to the given protein sequence, and either is, or has a mapping from, * the given dna sequence. @@ -1770,7 +1891,7 @@ public class AlignmentUtils * @param seqMappings * the set of mappings involving dnaSeq * @param aMapping - * an initial candidate from seqMappings + * a transcript-to-peptide mapping * @return */ static SequenceI findCdsForProtein(List mappings, @@ -1789,12 +1910,21 @@ public class AlignmentUtils * is this mapping from the whole dna sequence (i.e. CDS)? * allowing for possible stop codon on dna but not peptide */ - int mappedFromLength = MappingUtils.getLength(aMapping.getMap() - .getFromRanges()); + int mappedFromLength = MappingUtils + .getLength(aMapping.getMap().getFromRanges()); int dnaLength = seqDss.getLength(); - if (mappedFromLength == dnaLength || mappedFromLength == dnaLength - 3) + if (mappedFromLength == dnaLength + || mappedFromLength == dnaLength - CODON_LENGTH) { - return seqDss; + /* + * if sequence has CDS features, this is a transcript with no UTR + * - do not take this as the CDS sequence! (JAL-2789) + */ + if (seqDss.getFeatures().getFeaturesByOntology(SequenceOntologyI.CDS) + .isEmpty()) + { + return seqDss; + } } /* @@ -1808,20 +1938,23 @@ public class AlignmentUtils for (SequenceToSequenceMapping map : acf.getMappings()) { Mapping mapping = map.getMapping(); - if (mapping != aMapping && mapping.getMap().getFromRatio() == 3 + if (mapping != aMapping + && mapping.getMap().getFromRatio() == CODON_LENGTH && proteinProduct == mapping.getTo() && seqDss != map.getFromSeq()) { - mappedFromLength = MappingUtils.getLength(mapping.getMap() - .getFromRanges()); + mappedFromLength = MappingUtils + .getLength(mapping.getMap().getFromRanges()); if (mappedFromLength == map.getFromSeq().getLength()) { /* * found a 3:1 mapping to the protein product which covers - * the whole dna sequence i.e. is from CDS; finally check it - * is from the dna start sequence + * the whole dna sequence i.e. is from CDS; finally check the CDS + * is mapped from the given dna start sequence */ SequenceI cdsSeq = map.getFromSeq(); + // todo this test is weak if seqMappings contains multiple mappings; + // we get away with it if transcript:cds relationship is 1:1 List dnaToCdsMaps = MappingUtils .findMappingsForSequence(cdsSeq, seqMappings); if (!dnaToCdsMaps.isEmpty()) @@ -1842,9 +1975,14 @@ public class AlignmentUtils * * @param seq * @param mapping + * @param dataset + * - existing dataset. We check for sequences that look like the CDS + * we are about to construct, if one exists already, then we will + * just return that one. * @return CDS sequence (as a dataset sequence) */ - static SequenceI makeCdsSequence(SequenceI seq, Mapping mapping) + static SequenceI makeCdsSequence(SequenceI seq, Mapping mapping, + AlignmentI dataset) { char[] seqChars = seq.getSequence(); List fromRanges = mapping.getMap().getFromRanges(); @@ -1879,26 +2017,141 @@ public class AlignmentUtils String mapFromId = mapping.getMappedFromId(); String seqId = "CDS|" + (mapFromId != null ? mapFromId : seq.getName()); SequenceI newSeq = new Sequence(seqId, newSeqChars, 1, newPos); + if (dataset != null) + { + SequenceI[] matches = dataset.findSequenceMatch(newSeq.getName()); + if (matches != null) + { + boolean matched = false; + for (SequenceI mtch : matches) + { + if (mtch.getStart() != newSeq.getStart()) + { + continue; + } + if (mtch.getEnd() != newSeq.getEnd()) + { + continue; + } + if (!Arrays.equals(mtch.getSequence(), newSeq.getSequence())) + { + continue; + } + if (!matched) + { + matched = true; + newSeq = mtch; + } + else + { + System.err.println( + "JAL-2154 regression: warning - found (and ignnored a duplicate CDS sequence):" + + mtch.toString()); + } + } + } + } // newSeq.setDescription(mapFromId); return newSeq; } /** + * Adds any DBRefEntrys to cdsSeq from contig that have a Mapping congruent to + * the given mapping. + * + * @param cdsSeq + * @param contig + * @param proteinProduct + * @param mapping + * @return list of DBRefEntrys added + */ + protected static List propagateDBRefsToCDS(SequenceI cdsSeq, + SequenceI contig, SequenceI proteinProduct, Mapping mapping) + { + + // gather direct refs from contig congruent with mapping + List direct = new ArrayList<>(); + HashSet directSources = new HashSet<>(); + + if (contig.getDBRefs() != null) + { + for (DBRefEntry dbr : contig.getDBRefs()) + { + if (dbr.hasMap() && dbr.getMap().getMap().isTripletMap()) + { + MapList map = dbr.getMap().getMap(); + // check if map is the CDS mapping + if (mapping.getMap().equals(map)) + { + direct.add(dbr); + directSources.add(dbr.getSource()); + } + } + } + } + DBRefEntry[] onSource = DBRefUtils.selectRefs( + proteinProduct.getDBRefs(), + directSources.toArray(new String[0])); + List propagated = new ArrayList<>(); + + // and generate appropriate mappings + for (DBRefEntry cdsref : direct) + { + // clone maplist and mapping + MapList cdsposmap = new MapList( + Arrays.asList(new int[][] + { new int[] { cdsSeq.getStart(), cdsSeq.getEnd() } }), + cdsref.getMap().getMap().getToRanges(), 3, 1); + Mapping cdsmap = new Mapping(cdsref.getMap().getTo(), + cdsref.getMap().getMap()); + + // create dbref + DBRefEntry newref = new DBRefEntry(cdsref.getSource(), + cdsref.getVersion(), cdsref.getAccessionId(), + new Mapping(cdsmap.getTo(), cdsposmap)); + + // and see if we can map to the protein product for this mapping. + // onSource is the filtered set of accessions on protein that we are + // tranferring, so we assume accession is the same. + if (cdsmap.getTo() == null && onSource != null) + { + List sourceRefs = DBRefUtils.searchRefs(onSource, + cdsref.getAccessionId()); + if (sourceRefs != null) + { + for (DBRefEntry srcref : sourceRefs) + { + if (srcref.getSource().equalsIgnoreCase(cdsref.getSource())) + { + // we have found a complementary dbref on the protein product, so + // update mapping's getTo + newref.getMap().setTo(proteinProduct); + } + } + } + } + cdsSeq.addDBRef(newref); + propagated.add(newref); + } + return propagated; + } + + /** * Transfers co-located features on 'fromSeq' to 'toSeq', adjusting the * feature start/end ranges, optionally omitting specified feature types. * Returns the number of features copied. * * @param fromSeq * @param toSeq + * @param mapping + * the mapping from 'fromSeq' to 'toSeq' * @param select * if not null, only features of this type are copied (including * subtypes in the Sequence Ontology) - * @param mapping - * the mapping from 'fromSeq' to 'toSeq' * @param omitting */ - public static int transferFeatures(SequenceI fromSeq, SequenceI toSeq, + protected static int transferFeatures(SequenceI fromSeq, SequenceI toSeq, MapList mapping, String select, String... omitting) { SequenceI copyTo = toSeq; @@ -1907,82 +2160,84 @@ public class AlignmentUtils copyTo = copyTo.getDatasetSequence(); } - SequenceOntologyI so = SequenceOntologyFactory.getInstance(); + /* + * get features, optionally restricted by an ontology term + */ + List sfs = select == null ? fromSeq.getFeatures() + .getPositionalFeatures() : fromSeq.getFeatures() + .getFeaturesByOntology(select); + int count = 0; - SequenceFeature[] sfs = fromSeq.getSequenceFeatures(); - if (sfs != null) + for (SequenceFeature sf : sfs) { - for (SequenceFeature sf : sfs) + String type = sf.getType(); + boolean omit = false; + for (String toOmit : omitting) { - String type = sf.getType(); - if (select != null && !so.isA(type, select)) + if (type.equals(toOmit)) { - continue; - } - boolean omit = false; - for (String toOmit : omitting) - { - if (type.equals(toOmit)) - { - omit = true; - } - } - if (omit) - { - continue; + omit = true; } + } + if (omit) + { + continue; + } - /* - * locate the mapped range - null if either start or end is - * not mapped (no partial overlaps are calculated) - */ - int start = sf.getBegin(); - int end = sf.getEnd(); - int[] mappedTo = mapping.locateInTo(start, end); - /* - * if whole exon range doesn't map, try interpreting it - * as 5' or 3' exon overlapping the CDS range - */ - if (mappedTo == null) - { - mappedTo = mapping.locateInTo(end, end); - if (mappedTo != null) - { - /* - * end of exon is in CDS range - 5' overlap - * to a range from the start of the peptide - */ - mappedTo[0] = 1; - } - } - if (mappedTo == null) + /* + * locate the mapped range - null if either start or end is + * not mapped (no partial overlaps are calculated) + */ + int start = sf.getBegin(); + int end = sf.getEnd(); + int[] mappedTo = mapping.locateInTo(start, end); + /* + * if whole exon range doesn't map, try interpreting it + * as 5' or 3' exon overlapping the CDS range + */ + if (mappedTo == null) + { + mappedTo = mapping.locateInTo(end, end); + if (mappedTo != null) { - mappedTo = mapping.locateInTo(start, start); - if (mappedTo != null) - { - /* - * start of exon is in CDS range - 3' overlap - * to a range up to the end of the peptide - */ - mappedTo[1] = toSeq.getLength(); - } + /* + * end of exon is in CDS range - 5' overlap + * to a range from the start of the peptide + */ + mappedTo[0] = 1; } + } + if (mappedTo == null) + { + mappedTo = mapping.locateInTo(start, start); if (mappedTo != null) { - SequenceFeature copy = new SequenceFeature(sf); - copy.setBegin(Math.min(mappedTo[0], mappedTo[1])); - copy.setEnd(Math.max(mappedTo[0], mappedTo[1])); - copyTo.addSequenceFeature(copy); - count++; + /* + * start of exon is in CDS range - 3' overlap + * to a range up to the end of the peptide + */ + mappedTo[1] = toSeq.getLength(); } } + if (mappedTo != null) + { + int newBegin = Math.min(mappedTo[0], mappedTo[1]); + int newEnd = Math.max(mappedTo[0], mappedTo[1]); + SequenceFeature copy = new SequenceFeature(sf, newBegin, newEnd, + sf.getFeatureGroup(), sf.getScore()); + copyTo.addSequenceFeature(copy); + count++; + } } return count; } /** * Returns a mapping from dna to protein by inspecting sequence features of - * type "CDS" on the dna. + * type "CDS" on the dna. A mapping is constructed if the total CDS feature + * length is 3 times the peptide length (optionally after dropping a trailing + * stop codon). This method does not check whether the CDS nucleotide sequence + * translates to the peptide sequence. * * @param dnaSeq * @param proteinSeq @@ -1994,6 +2249,16 @@ public class AlignmentUtils List ranges = findCdsPositions(dnaSeq); int mappedDnaLength = MappingUtils.getLength(ranges); + /* + * if not a whole number of codons, truncate mapping + */ + int codonRemainder = mappedDnaLength % CODON_LENGTH; + if (codonRemainder > 0) + { + mappedDnaLength -= codonRemainder; + MappingUtils.removeEndPositions(codonRemainder, ranges); + } + int proteinLength = proteinSeq.getLength(); int proteinStart = proteinSeq.getStart(); int proteinEnd = proteinSeq.getEnd(); @@ -2008,28 +2273,32 @@ public class AlignmentUtils proteinStart++; proteinLength--; } - List proteinRange = new ArrayList(); + List proteinRange = new ArrayList<>(); /* * dna length should map to protein (or protein plus stop codon) */ - int codesForResidues = mappedDnaLength / 3; + int codesForResidues = mappedDnaLength / CODON_LENGTH; if (codesForResidues == (proteinLength + 1)) { // assuming extra codon is for STOP and not in peptide + // todo: check trailing codon is indeed a STOP codon codesForResidues--; + mappedDnaLength -= CODON_LENGTH; + MappingUtils.removeEndPositions(CODON_LENGTH, ranges); } + if (codesForResidues == proteinLength) { proteinRange.add(new int[] { proteinStart, proteinEnd }); - return new MapList(ranges, proteinRange, 3, 1); + return new MapList(ranges, proteinRange, CODON_LENGTH, 1); } return null; } /** * Returns a list of CDS ranges found (as sequence positions base 1), i.e. of - * start/end positions of sequence features of type "CDS" (or a sub-type of + * [start, end] positions of sequence features of type "CDS" (or a sub-type of * CDS in the Sequence Ontology). The ranges are sorted into ascending start * position order, so this method is only valid for linear CDS in the same * sense as the protein product. @@ -2037,62 +2306,46 @@ public class AlignmentUtils * @param dnaSeq * @return */ - public static List findCdsPositions(SequenceI dnaSeq) + protected static List findCdsPositions(SequenceI dnaSeq) { - List result = new ArrayList(); - SequenceFeature[] sfs = dnaSeq.getSequenceFeatures(); - if (sfs == null) + List result = new ArrayList<>(); + + List sfs = dnaSeq.getFeatures().getFeaturesByOntology( + SequenceOntologyI.CDS); + if (sfs.isEmpty()) { return result; } - - SequenceOntologyI so = SequenceOntologyFactory.getInstance(); - int startPhase = 0; + SequenceFeatures.sortFeatures(sfs, true); for (SequenceFeature sf : sfs) { + int phase = 0; + try + { + phase = Integer.parseInt(sf.getPhase()); + } catch (NumberFormatException e) + { + // ignore + } /* - * process a CDS feature (or a sub-type of CDS) + * phase > 0 on first codon means 5' incomplete - skip to the start + * of the next codon; example ENST00000496384 */ - if (so.isA(sf.getType(), SequenceOntologyI.CDS)) + int begin = sf.getBegin(); + int end = sf.getEnd(); + if (result.isEmpty() && phase > 0) { - int phase = 0; - try - { - phase = Integer.parseInt(sf.getPhase()); - } catch (NumberFormatException e) + begin += phase; + if (begin > end) { - // ignore - } - /* - * phase > 0 on first codon means 5' incomplete - skip to the start - * of the next codon; example ENST00000496384 - */ - int begin = sf.getBegin(); - int end = sf.getEnd(); - if (result.isEmpty()) - { - begin += phase; - if (begin > end) - { - // shouldn't happen! - System.err - .println("Error: start phase extends beyond start CDS in " - + dnaSeq.getName()); - } + // shouldn't happen! + System.err + .println("Error: start phase extends beyond start CDS in " + + dnaSeq.getName()); } - result.add(new int[] { begin, end }); } - } - - /* - * remove 'startPhase' positions (usually 0) from the first range - * so we begin at the start of a complete codon - */ - if (!result.isEmpty()) - { - // TODO JAL-2022 correctly model start phase > 0 - result.get(0)[0] += startPhase; + result.add(new int[] { begin, end }); } /* @@ -2102,14 +2355,7 @@ public class AlignmentUtils * ranges are assembled in order. Other cases should not use this method, * but instead construct an explicit mapping for CDS (e.g. EMBL parsing). */ - Collections.sort(result, new Comparator() - { - @Override - public int compare(int[] o1, int[] o2) - { - return Integer.compare(o1[0], o2[0]); - } - }); + Collections.sort(result, IntRangeComparator.ASCENDING); return result; } @@ -2162,24 +2408,6 @@ public class AlignmentUtils count += computePeptideVariants(peptide, peptidePos, codonVariants); } - /* - * sort to get sequence features in start position order - * - would be better to store in Sequence as a TreeSet or NCList? - */ - if (peptide.getSequenceFeatures() != null) - { - Arrays.sort(peptide.getSequenceFeatures(), - new Comparator() - { - @Override - public int compare(SequenceFeature o1, SequenceFeature o2) - { - int c = Integer.compare(o1.getBegin(), o2.getBegin()); - return c == 0 ? Integer.compare(o1.getEnd(), o2.getEnd()) - : c; - } - }); - } return count; } @@ -2213,15 +2441,22 @@ public class AlignmentUtils { if (var.variant != null) { - String alleles = (String) var.variant.getValue("alleles"); + String alleles = (String) var.variant.getValue(Gff3Helper.ALLELES); if (alleles != null) { for (String base : alleles.split(",")) { - String codon = base + base2 + base3; - if (addPeptideVariant(peptide, peptidePos, residue, var, codon)) + if (!base1.equalsIgnoreCase(base)) { - count++; + String codon = base.toUpperCase() + base2.toLowerCase() + + base3.toLowerCase(); + String canonical = base1.toUpperCase() + base2.toLowerCase() + + base3.toLowerCase(); + if (addPeptideVariant(peptide, peptidePos, residue, var, + codon, canonical)) + { + count++; + } } } } @@ -2235,15 +2470,22 @@ public class AlignmentUtils { if (var.variant != null) { - String alleles = (String) var.variant.getValue("alleles"); + String alleles = (String) var.variant.getValue(Gff3Helper.ALLELES); if (alleles != null) { for (String base : alleles.split(",")) { - String codon = base1 + base + base3; - if (addPeptideVariant(peptide, peptidePos, residue, var, codon)) + if (!base2.equalsIgnoreCase(base)) { - count++; + String codon = base1.toLowerCase() + base.toUpperCase() + + base3.toLowerCase(); + String canonical = base1.toLowerCase() + base2.toUpperCase() + + base3.toLowerCase(); + if (addPeptideVariant(peptide, peptidePos, residue, var, + codon, canonical)) + { + count++; + } } } } @@ -2257,15 +2499,22 @@ public class AlignmentUtils { if (var.variant != null) { - String alleles = (String) var.variant.getValue("alleles"); + String alleles = (String) var.variant.getValue(Gff3Helper.ALLELES); if (alleles != null) { for (String base : alleles.split(",")) { - String codon = base1 + base2 + base; - if (addPeptideVariant(peptide, peptidePos, residue, var, codon)) + if (!base3.equalsIgnoreCase(base)) { - count++; + String codon = base1.toLowerCase() + base2.toLowerCase() + + base.toUpperCase(); + String canonical = base1.toLowerCase() + base2.toLowerCase() + + base3.toUpperCase(); + if (addPeptideVariant(peptide, peptidePos, residue, var, + codon, canonical)) + { + count++; + } } } } @@ -2276,20 +2525,22 @@ public class AlignmentUtils } /** - * Helper method that adds a peptide variant feature, provided the given codon - * translates to a value different to the current residue (is a non-synonymous - * variant). ID and clinical_significance attributes of the dna variant (if - * present) are copied to the new feature. + * Helper method that adds a peptide variant feature. ID and + * clinical_significance attributes of the dna variant (if present) are copied + * to the new feature. * * @param peptide * @param peptidePos * @param residue * @param var * @param codon + * the variant codon e.g. aCg + * @param canonical + * the 'normal' codon e.g. aTg * @return true if a feature was added, else false */ static boolean addPeptideVariant(SequenceI peptide, int peptidePos, - String residue, DnaVariant var, String codon) + String residue, DnaVariant var, String codon, String canonical) { /* * get peptide translation of codon e.g. GAT -> D @@ -2297,69 +2548,85 @@ public class AlignmentUtils * e.g. multibase variants or HGMD_MUTATION etc * are currently ignored here */ - String trans = codon.contains("-") ? "-" - : (codon.length() > 3 ? null : ResidueProperties - .codonTranslate(codon)); - if (trans != null && !trans.equals(residue)) + String trans = codon.contains("-") ? null + : (codon.length() > CODON_LENGTH ? null + : ResidueProperties.codonTranslate(codon)); + if (trans == null) + { + return false; + } + String desc = canonical + "/" + codon; + String featureType = ""; + if (trans.equals(residue)) + { + featureType = SequenceOntologyI.SYNONYMOUS_VARIANT; + } + else if (ResidueProperties.STOP.equals(trans)) + { + featureType = SequenceOntologyI.STOP_GAINED; + } + else { String residue3Char = StringUtils .toSentenceCase(ResidueProperties.aa2Triplet.get(residue)); String trans3Char = StringUtils .toSentenceCase(ResidueProperties.aa2Triplet.get(trans)); - String desc = "p." + residue3Char + peptidePos + trans3Char; - // set score to 0f so 'graduated colour' option is offered! JAL-2060 - SequenceFeature sf = new SequenceFeature( - SequenceOntologyI.SEQUENCE_VARIANT, desc, peptidePos, - peptidePos, 0f, "Jalview"); - StringBuilder attributes = new StringBuilder(32); - String id = (String) var.variant.getValue(ID); - if (id != null) - { - if (id.startsWith(SEQUENCE_VARIANT)) - { - id = id.substring(SEQUENCE_VARIANT.length()); - } - sf.setValue(ID, id); - attributes.append(ID).append("=").append(id); - // TODO handle other species variants - StringBuilder link = new StringBuilder(32); - try - { - link.append(desc).append(" ").append(id) - .append("|http://www.ensembl.org/Homo_sapiens/Variation/Summary?v=") - .append(URLEncoder.encode(id, "UTF-8")); - sf.addLink(link.toString()); - } catch (UnsupportedEncodingException e) - { - // as if - } - } - String clinSig = (String) var.variant - .getValue(CLINICAL_SIGNIFICANCE); - if (clinSig != null) + desc = "p." + residue3Char + peptidePos + trans3Char; + featureType = SequenceOntologyI.NONSYNONYMOUS_VARIANT; + } + SequenceFeature sf = new SequenceFeature(featureType, desc, peptidePos, + peptidePos, var.getSource()); + + StringBuilder attributes = new StringBuilder(32); + String id = (String) var.variant.getValue(VARIANT_ID); + if (id != null) + { + if (id.startsWith(SEQUENCE_VARIANT)) { - sf.setValue(CLINICAL_SIGNIFICANCE, clinSig); - attributes.append(";").append(CLINICAL_SIGNIFICANCE).append("=") - .append(clinSig); + id = id.substring(SEQUENCE_VARIANT.length()); } - peptide.addSequenceFeature(sf); - if (attributes.length() > 0) + sf.setValue(VARIANT_ID, id); + attributes.append(VARIANT_ID).append("=").append(id); + // TODO handle other species variants JAL-2064 + StringBuilder link = new StringBuilder(32); + try + { + link.append(desc).append(" ").append(id).append( + "|http://www.ensembl.org/Homo_sapiens/Variation/Summary?v=") + .append(URLEncoder.encode(id, "UTF-8")); + sf.addLink(link.toString()); + } catch (UnsupportedEncodingException e) { - sf.setAttributes(attributes.toString()); + // as if } - return true; } - return false; + String clinSig = (String) var.variant.getValue(CLINICAL_SIGNIFICANCE); + if (clinSig != null) + { + sf.setValue(CLINICAL_SIGNIFICANCE, clinSig); + attributes.append(";").append(CLINICAL_SIGNIFICANCE).append("=") + .append(clinSig); + } + peptide.addSequenceFeature(sf); + if (attributes.length() > 0) + { + sf.setAttributes(attributes.toString()); + } + return true; } /** * Builds a map whose key is position in the protein sequence, and value is a - * list of the base and all variants for each corresponding codon position + * list of the base and all variants for each corresponding codon position. + *

    + * This depends on dna variants being held as a comma-separated list as + * property "alleles" on variant features. * * @param dnaSeq * @param dnaToProtein * @return */ + @SuppressWarnings("unchecked") static LinkedHashMap[]> buildDnaVariantsMap( SequenceI dnaSeq, MapList dnaToProtein) { @@ -2367,11 +2634,11 @@ public class AlignmentUtils * map from peptide position to all variants of the codon which codes for it * LinkedHashMap ensures we keep the peptide features in sequence order */ - LinkedHashMap[]> variants = new LinkedHashMap[]>(); - SequenceOntologyI so = SequenceOntologyFactory.getInstance(); + LinkedHashMap[]> variants = new LinkedHashMap<>(); - SequenceFeature[] dnaFeatures = dnaSeq.getSequenceFeatures(); - if (dnaFeatures == null) + List dnaFeatures = dnaSeq.getFeatures() + .getFeaturesByOntology(SequenceOntologyI.SEQUENCE_VARIANT); + if (dnaFeatures.isEmpty()) { return variants; } @@ -2391,84 +2658,89 @@ public class AlignmentUtils // not handling multi-locus variant features continue; } - if (so.isA(sf.getType(), SequenceOntologyI.SEQUENCE_VARIANT)) + + /* + * ignore variant if not a SNP + */ + String alls = (String) sf.getValue(Gff3Helper.ALLELES); + if (alls == null) { - int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol); - if (mapsTo == null) - { - // feature doesn't lie within coding region - continue; - } - int peptidePosition = mapsTo[0]; - List[] codonVariants = variants.get(peptidePosition); - if (codonVariants == null) - { - codonVariants = new ArrayList[3]; - codonVariants[0] = new ArrayList(); - codonVariants[1] = new ArrayList(); - codonVariants[2] = new ArrayList(); - variants.put(peptidePosition, codonVariants); - } + continue; // non-SNP VCF variant perhaps - can't process this + } - /* - * extract dna variants to a string array - */ - String alls = (String) sf.getValue("alleles"); - if (alls == null) - { - continue; - } - String[] alleles = alls.toUpperCase().split(","); - int i = 0; - for (String allele : alleles) + String[] alleles = alls.toUpperCase().split(","); + boolean isSnp = true; + for (String allele : alleles) + { + if (allele.trim().length() > 1) { - alleles[i++] = allele.trim(); // lose any space characters "A, G" + isSnp = false; } + } + if (!isSnp) + { + continue; + } - /* - * get this peptide's codon positions e.g. [3, 4, 5] or [4, 7, 10] - */ - int[] codon = peptidePosition == lastPeptidePostion ? lastCodon - : MappingUtils.flattenRanges(dnaToProtein.locateInFrom( - peptidePosition, peptidePosition)); - lastPeptidePostion = peptidePosition; - lastCodon = codon; + int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol); + if (mapsTo == null) + { + // feature doesn't lie within coding region + continue; + } + int peptidePosition = mapsTo[0]; + List[] codonVariants = variants.get(peptidePosition); + if (codonVariants == null) + { + codonVariants = new ArrayList[CODON_LENGTH]; + codonVariants[0] = new ArrayList<>(); + codonVariants[1] = new ArrayList<>(); + codonVariants[2] = new ArrayList<>(); + variants.put(peptidePosition, codonVariants); + } - /* - * save nucleotide (and any variant) for each codon position - */ - for (int codonPos = 0; codonPos < 3; codonPos++) + /* + * get this peptide's codon positions e.g. [3, 4, 5] or [4, 7, 10] + */ + int[] codon = peptidePosition == lastPeptidePostion ? lastCodon + : MappingUtils.flattenRanges(dnaToProtein.locateInFrom( + peptidePosition, peptidePosition)); + lastPeptidePostion = peptidePosition; + lastCodon = codon; + + /* + * save nucleotide (and any variant) for each codon position + */ + for (int codonPos = 0; codonPos < CODON_LENGTH; codonPos++) + { + String nucleotide = String.valueOf( + dnaSeq.getCharAt(codon[codonPos] - dnaStart)).toUpperCase(); + List codonVariant = codonVariants[codonPos]; + if (codon[codonPos] == dnaCol) { - String nucleotide = String.valueOf( - dnaSeq.getCharAt(codon[codonPos] - dnaStart)) - .toUpperCase(); - List codonVariant = codonVariants[codonPos]; - if (codon[codonPos] == dnaCol) + if (!codonVariant.isEmpty() + && codonVariant.get(0).variant == null) { - if (!codonVariant.isEmpty() - && codonVariant.get(0).variant == null) - { - /* - * already recorded base value, add this variant - */ - codonVariant.get(0).variant = sf; - } - else - { - /* - * add variant with base value - */ - codonVariant.add(new DnaVariant(nucleotide, sf)); - } + /* + * already recorded base value, add this variant + */ + codonVariant.get(0).variant = sf; } - else if (codonVariant.isEmpty()) + else { /* - * record (possibly non-varying) base value + * add variant with base value */ - codonVariant.add(new DnaVariant(nucleotide)); + codonVariant.add(new DnaVariant(nucleotide, sf)); } } + else if (codonVariant.isEmpty()) + { + /* + * record (possibly non-varying) base value + */ + codonVariant.add(new DnaVariant(nucleotide)); + } } } return variants; @@ -2490,7 +2762,7 @@ public class AlignmentUtils { AlignmentI copy = new Alignment(new Alignment(seqs)); copy.setDataset(dataset); - + boolean isProtein = !copy.isNucleotide(); SequenceIdMatcher matcher = new SequenceIdMatcher(seqs); if (xrefs != null) { @@ -2501,7 +2773,8 @@ public class AlignmentUtils { for (DBRefEntry dbref : dbrefs) { - if (dbref.getMap() == null || dbref.getMap().getTo() == null) + if (dbref.getMap() == null || dbref.getMap().getTo() == null + || dbref.getMap().getTo().isProtein() != isProtein) { continue; } @@ -2546,12 +2819,13 @@ public class AlignmentUtils /* * fancy case - aligning via mappings between sequences */ - List unmapped = new ArrayList(); + List unmapped = new ArrayList<>(); Map> columnMap = buildMappedColumnsMap( unaligned, aligned, unmapped); int width = columnMap.size(); char gap = unaligned.getGapCharacter(); int realignedCount = 0; + // TODO: verify this loop scales sensibly for very wide/high alignments for (SequenceI seq : unaligned.getSequences()) { @@ -2581,7 +2855,7 @@ public class AlignmentUtils } newCol++; } - + /* * trim trailing gaps */ @@ -2605,7 +2879,10 @@ public class AlignmentUtils * true; else returns false * * @param unaligned + * - sequences to be aligned based on aligned * @param aligned + * - 'guide' alignment containing sequences derived from same dataset + * as unaligned * @return */ static boolean alignAsSameSequences(AlignmentI unaligned, @@ -2616,10 +2893,16 @@ public class AlignmentUtils return false; // should only pass alignments with datasets here } - Map alignedDatasets = new HashMap(); + // map from dataset sequence to alignment sequence(s) + Map> alignedDatasets = new HashMap<>(); for (SequenceI seq : aligned.getSequences()) { - alignedDatasets.put(seq.getDatasetSequence(), seq); + SequenceI ds = seq.getDatasetSequence(); + if (alignedDatasets.get(ds) == null) + { + alignedDatasets.put(ds, new ArrayList()); + } + alignedDatasets.get(ds).add(seq); } /* @@ -2635,15 +2918,22 @@ public class AlignmentUtils } /* - * second pass - copy aligned sequences + * second pass - copy aligned sequences; + * heuristic rule: pair off sequences in order for the case where + * more than one shares the same dataset sequence */ for (SequenceI seq : unaligned.getSequences()) { - SequenceI alignedSequence = alignedDatasets.get(seq - .getDatasetSequence()); + List alignedSequences = alignedDatasets + .get(seq.getDatasetSequence()); // TODO: getSequenceAsString() will be deprecated in the future // TODO: need to leave to SequenceI implementor to update gaps - seq.setSequence(alignedSequence.getSequenceAsString()); + seq.setSequence(alignedSequences.get(0).getSequenceAsString()); + if (alignedSequences.size() > 0) + { + // pop off aligned sequences (except the last one) + alignedSequences.remove(0); + } } return true; @@ -2658,15 +2948,16 @@ public class AlignmentUtils * @param unmapped * @return */ - static Map> buildMappedColumnsMap( - AlignmentI unaligned, AlignmentI aligned, List unmapped) + static SortedMap> buildMappedColumnsMap( + AlignmentI unaligned, AlignmentI aligned, + List unmapped) { /* * Map will hold, for each aligned column position, a map of * {unalignedSequence, characterPerSequence} at that position. * TreeMap keeps the entries in ascending column order. */ - Map> map = new TreeMap>(); + SortedMap> map = new TreeMap<>(); /* * record any sequences that have no mapping so can't be realigned @@ -2694,7 +2985,8 @@ public class AlignmentUtils } /** - * Helper method that adds to a map the mapped column positions of a sequence.
    + * Helper method that adds to a map the mapped column positions of a sequence. + *
    * For example if aaTT-Tg-gAAA is mapped to TTTAAA then the map should record * that columns 3,4,6,10,11,12 map to characters T,T,T,A,A,A of the mapped to * sequence. @@ -2723,13 +3015,11 @@ public class AlignmentUtils */ if (seqMap.getTo() == fromSeq.getDatasetSequence()) { - seqMap = new Mapping(seq.getDatasetSequence(), seqMap.getMap() - .getInverse()); + seqMap = new Mapping(seq.getDatasetSequence(), + seqMap.getMap().getInverse()); } - char[] fromChars = fromSeq.getSequence(); int toStart = seq.getStart(); - char[] toChars = seq.getSequence(); /* * traverse [start, end, start, end...] ranges in fromSeq @@ -2760,10 +3050,10 @@ public class AlignmentUtils * of the next character of the mapped-to sequence; stop when all * the characters of the range have been counted */ - while (mappedCharPos <= range[1] && fromCol <= fromChars.length + while (mappedCharPos <= range[1] && fromCol <= fromSeq.getLength() && fromCol >= 0) { - if (!Comparison.isGap(fromChars[fromCol - 1])) + if (!Comparison.isGap(fromSeq.getCharAt(fromCol - 1))) { /* * mapped from sequence has a character in this column @@ -2772,10 +3062,10 @@ public class AlignmentUtils Map seqsMap = map.get(fromCol); if (seqsMap == null) { - seqsMap = new HashMap(); + seqsMap = new HashMap<>(); map.put(fromCol, seqsMap); } - seqsMap.put(seq, toChars[mappedCharPos - toStart]); + seqsMap.put(seq, seq.getCharAt(mappedCharPos - toStart)); mappedCharPos++; } fromCol += (forward ? 1 : -1);