X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FAlignmentUtils.java;h=8e0335fee4a161b0414377a99fe60992228f6f12;hb=5c6564f903f75960af960720a8635ab8709afc37;hp=c88a4628302e52f2afb5cb97f2c2462ef532e2d9;hpb=6fe36904fddf9ecb85e67974f48081bba373e8ab;p=jalview.git diff --git a/src/jalview/analysis/AlignmentUtils.java b/src/jalview/analysis/AlignmentUtils.java index c88a462..8e0335f 100644 --- a/src/jalview/analysis/AlignmentUtils.java +++ b/src/jalview/analysis/AlignmentUtils.java @@ -22,6 +22,7 @@ package jalview.analysis; import static jalview.io.gff.GffConstants.CLINICAL_SIGNIFICANCE; +import jalview.commands.RemoveGapColCommand; import jalview.datamodel.AlignedCodon; import jalview.datamodel.AlignedCodonFrame; import jalview.datamodel.AlignedCodonFrame.SequenceToSequenceMapping; @@ -29,6 +30,7 @@ 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; @@ -73,12 +75,15 @@ 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 @@ -127,7 +132,7 @@ 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()) { @@ -257,7 +262,7 @@ 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(); @@ -266,7 +271,7 @@ public class AlignmentUtils List seqs = theMap.get(name); if (seqs == null) { - seqs = new ArrayList(); + seqs = new ArrayList<>(); theMap.put(name, seqs); } seqs.add(seq); @@ -293,8 +298,8 @@ 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 @@ -394,7 +399,7 @@ 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) @@ -464,7 +469,7 @@ public class AlignmentUtils { String lastCodon = String.valueOf(cdnaSeqChars, cdnaLength - CODON_LENGTH, CODON_LENGTH).toUpperCase(); - for (String stop : ResidueProperties.STOP) + for (String stop : ResidueProperties.STOP_CODONS) { if (lastCodon.equals(stop)) { @@ -535,7 +540,8 @@ public class AlignmentUtils * 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; } @@ -567,7 +573,8 @@ public class AlignmentUtils if (dnaPos == cdnaSeqChars.length - CODON_LENGTH) { String codon = String.valueOf(cdnaSeqChars, dnaPos, CODON_LENGTH); - if ("STOP".equals(ResidueProperties.codonTranslate(codon))) + if (ResidueProperties.STOP + .equals(ResidueProperties.codonTranslate(codon))) { return true; } @@ -880,7 +887,7 @@ 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); @@ -1091,7 +1098,7 @@ 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()) @@ -1137,9 +1144,9 @@ 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()) @@ -1318,7 +1325,7 @@ 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); @@ -1455,7 +1462,7 @@ public class AlignmentUtils { continue; } - final List result = new ArrayList(); + final List result = new ArrayList<>(); for (AlignmentAnnotation dsann : datasetAnnotations) { /* @@ -1633,22 +1640,22 @@ public class AlignmentUtils AlignmentI dataset, SequenceI[] products) { if (dataset == null || dataset.getDataset() != null) - { + { 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 .getDatasetSequence()); - } + } } /* @@ -1670,15 +1677,15 @@ public class AlignmentUtils List seqMappings = MappingUtils .findMappingsForSequence(dnaSeq, mappings); for (AlignedCodonFrame mapping : seqMappings) - { + { List mappingsFromSequence = mapping .getMappingsFromSequence(dnaSeq); for (Mapping aMapping : mappingsFromSequence) - { + { MapList mapList = aMapping.getMap(); if (mapList.getFromRatio() == 1) - { + { /* * not a dna-to-protein mapping (likely dna-to-cds) */ @@ -1704,15 +1711,15 @@ public class AlignmentUtils if (cdsSeq != null) { if (!foundSeqs.contains(cdsSeq)) - { + { foundSeqs.add(cdsSeq); SequenceI derivedSequence = cdsSeq.deriveSequence(); cdsSeqs.add(derivedSequence); if (!dataset.getSequences().contains(cdsSeq)) - { + { dataset.addSequence(cdsSeq); + } } - } continue; } @@ -1740,8 +1747,10 @@ public class AlignmentUtils /* * add a mapping from CDS to the (unchanged) mapped to range */ - List cdsRange = Collections.singletonList(new int[] { 1, - cdsSeq.getLength() }); + List cdsRange = Collections + .singletonList(new int[] + { cdsSeq.getStart(), + cdsSeq.getLength() + cdsSeq.getStart() - 1 }); MapList cdsToProteinMap = new MapList(cdsRange, mapList.getToRanges(), mapList.getFromRatio(), mapList.getToRatio()); @@ -1763,7 +1772,7 @@ public class AlignmentUtils * add another mapping from original 'from' range to CDS */ AlignedCodonFrame dnaToCdsMapping = new AlignedCodonFrame(); - MapList dnaToCdsMap = new MapList(mapList.getFromRanges(), + final MapList dnaToCdsMap = new MapList(mapList.getFromRanges(), cdsRange, 1, 1); dnaToCdsMapping.addMap(dnaSeq.getDatasetSequence(), cdsSeqDss, dnaToCdsMap); @@ -1773,6 +1782,13 @@ 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 @@ -1795,13 +1811,11 @@ public class AlignmentUtils * 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(dnaToCdsMap - .getInverse()))); + cdsCrossRef.setMap(new Mapping(dnaDss, new MapList(cdsToDnaMap))); cdsSeqDss.addDBRef(cdsCrossRef); dnaSeq.addDBRef(new DBRefEntry(source, version, cdsSeq @@ -1818,16 +1832,16 @@ public class AlignmentUtils proteinToCdsRef.setMap(new Mapping(cdsSeqDss, cdsToProteinMap .getInverse())); proteinProduct.addDBRef(proteinToCdsRef); - } + } /* * transfer any features on dna that overlap the CDS */ transferFeatures(dnaSeq, cdsSeq, dnaToCdsMap, null, SequenceOntologyI.CDS); + } } } - } AlignmentI cds = new Alignment(cdsSeqs.toArray(new SequenceI[cdsSeqs .size()])); @@ -1837,6 +1851,38 @@ 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. @@ -1848,7 +1894,7 @@ 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, @@ -1873,7 +1919,15 @@ public class AlignmentUtils 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; + } } /* @@ -1898,10 +1952,12 @@ public class AlignmentUtils { /* * 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()) @@ -1931,39 +1987,61 @@ public class AlignmentUtils static SequenceI makeCdsSequence(SequenceI seq, Mapping mapping, AlignmentI dataset) { - char[] seqChars = seq.getSequence(); - List fromRanges = mapping.getMap().getFromRanges(); - int cdsWidth = MappingUtils.getLength(fromRanges); - char[] newSeqChars = new char[cdsWidth]; + /* + * construct CDS sequence name as "CDS|" with 'from id' held in the mapping + * if set (e.g. EMBL protein_id), else sequence name appended + */ + String mapFromId = mapping.getMappedFromId(); + final String seqId = "CDS|" + + (mapFromId != null ? mapFromId : seq.getName()); - int newPos = 0; - for (int[] range : fromRanges) + SequenceI newSeq = null; + + final MapList maplist = mapping.getMap(); + if (maplist.isContiguous() && maplist.isFromForwardStrand()) { - if (range[0] <= range[1]) - { - // forward strand mapping - just copy the range - int length = range[1] - range[0] + 1; - System.arraycopy(seqChars, range[0] - 1, newSeqChars, newPos, - length); - newPos += length; - } - else + /* + * just a subsequence, keep same dataset sequence + */ + int start = maplist.getFromLowest(); + int end = maplist.getFromHighest(); + newSeq = seq.getSubSequence(start - 1, end); + newSeq.setName(seqId); + } + else + { + /* + * construct by splicing mapped from ranges + */ + char[] seqChars = seq.getSequence(); + List fromRanges = maplist.getFromRanges(); + int cdsWidth = MappingUtils.getLength(fromRanges); + char[] newSeqChars = new char[cdsWidth]; + + int newPos = 0; + for (int[] range : fromRanges) { - // reverse strand mapping - copy and complement one by one - for (int i = range[0]; i >= range[1]; i--) + if (range[0] <= range[1]) { - newSeqChars[newPos++] = Dna.getComplement(seqChars[i - 1]); + // forward strand mapping - just copy the range + int length = range[1] - range[0] + 1; + System.arraycopy(seqChars, range[0] - 1, newSeqChars, newPos, + length); + newPos += length; + } + else + { + // reverse strand mapping - copy and complement one by one + for (int i = range[0]; i >= range[1]; i--) + { + newSeqChars[newPos++] = Dna.getComplement(seqChars[i - 1]); + } } } + + newSeq = new Sequence(seqId, newSeqChars, 1, newPos); } - /* - * assign 'from id' held in the mapping if set (e.g. EMBL protein_id), - * else generate a sequence name - */ - 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()); @@ -2004,21 +2082,23 @@ public class AlignmentUtils } /** - * add any DBRefEntrys to cdsSeq from contig that have a Mapping congruent to + * 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. + * @return list of DBRefEntrys added */ - public static List propagateDBRefsToCDS(SequenceI cdsSeq, + protected static List propagateDBRefsToCDS(SequenceI cdsSeq, SequenceI contig, SequenceI proteinProduct, Mapping mapping) { - // gather direct refs from contig congrent with mapping - List direct = new ArrayList(); - HashSet directSources = new HashSet(); + // 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()) @@ -2038,7 +2118,7 @@ public class AlignmentUtils DBRefEntry[] onSource = DBRefUtils.selectRefs( proteinProduct.getDBRefs(), directSources.toArray(new String[0])); - List propagated = new ArrayList(); + List propagated = new ArrayList<>(); // and generate appropriate mappings for (DBRefEntry cdsref : direct) @@ -2096,7 +2176,7 @@ public class AlignmentUtils * subtypes in the Sequence Ontology) * @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; @@ -2104,6 +2184,10 @@ public class AlignmentUtils { copyTo = copyTo.getDatasetSequence(); } + if (fromSeq == copyTo || fromSeq.getDatasetSequence() == copyTo) + { + return 0; // shared dataset sequence + } /* * get features, optionally restricted by an ontology term @@ -2179,7 +2263,10 @@ public class AlignmentUtils /** * 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 @@ -2191,6 +2278,16 @@ 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(); @@ -2205,7 +2302,7 @@ public class AlignmentUtils proteinStart++; proteinLength--; } - List proteinRange = new ArrayList(); + List proteinRange = new ArrayList<>(); /* * dna length should map to protein (or protein plus stop codon) @@ -2214,8 +2311,12 @@ public class AlignmentUtils 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 }); @@ -2226,7 +2327,7 @@ public class AlignmentUtils /** * 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. @@ -2234,9 +2335,9 @@ public class AlignmentUtils * @param dnaSeq * @return */ - public static List findCdsPositions(SequenceI dnaSeq) + protected static List findCdsPositions(SequenceI dnaSeq) { - List result = new ArrayList(); + List result = new ArrayList<>(); List sfs = dnaSeq.getFeatures().getFeaturesByOntology( SequenceOntologyI.CDS); @@ -2245,7 +2346,6 @@ public class AlignmentUtils return result; } SequenceFeatures.sortFeatures(sfs, true); - int startPhase = 0; for (SequenceFeature sf : sfs) { @@ -2263,7 +2363,7 @@ public class AlignmentUtils */ int begin = sf.getBegin(); int end = sf.getEnd(); - if (result.isEmpty()) + if (result.isEmpty() && phase > 0) { begin += phase; if (begin > end) @@ -2278,16 +2378,6 @@ public class AlignmentUtils } /* - * 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; - } - - /* * Finally sort ranges by start position. This avoids a dependency on * keeping features in order on the sequence (if they are in order anyway, * the sort will have almost no work to do). The implicit assumption is CDS @@ -2351,23 +2441,24 @@ public class AlignmentUtils } /** - * Computes non-synonymous peptide variants from codon variants and adds them - * as sequence_variant features on the protein sequence (one feature per - * allele variant). Selected attributes (variant id, clinical significance) - * are copied over to the new features. + * Computes non-synonymous peptide variants from codon variants and adds them as + * sequence_variant features on the protein sequence (one feature per allele + * variant). Selected attributes (variant id, clinical significance) are copied + * over to the new features. * * @param peptide - * the protein sequence + * the protein dataset (ungapped) sequence * @param peptidePos - * the position to compute peptide variants for + * the position to compute peptide variants for * @param codonVariants - * a list of dna variants per codon position + * a list of dna variants per codon position * @return the number of features added */ static int computePeptideVariants(SequenceI peptide, int peptidePos, List[] codonVariants) { - String residue = String.valueOf(peptide.getCharAt(peptidePos - 1)); + String residue = String + .valueOf(peptide.getCharAt(peptidePos - peptide.getStart())); int count = 0; String base1 = codonVariants[0].get(0).base; String base2 = codonVariants[1].get(0).base; @@ -2385,10 +2476,17 @@ public class AlignmentUtils { 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++; + } } } } @@ -2407,10 +2505,17 @@ public class AlignmentUtils { 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++; + } } } } @@ -2429,10 +2534,17 @@ public class AlignmentUtils { 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++; + } } } } @@ -2443,20 +2555,22 @@ 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 @@ -2464,62 +2578,79 @@ public class AlignmentUtils * e.g. multibase variants or HGMD_MUTATION etc * are currently ignored here */ - String trans = codon.contains("-") ? "-" + String trans = codon.contains("-") ? null : (codon.length() > CODON_LENGTH ? null : ResidueProperties.codonTranslate(codon)); - if (trans != null && !trans.equals(residue)) + 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; - SequenceFeature sf = new SequenceFeature( - SequenceOntologyI.SEQUENCE_VARIANT, desc, peptidePos, - peptidePos, var.getSource()); - 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 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) - { - // 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 { - sf.setAttributes(attributes.toString()); + 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 } - 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 @@ -2533,7 +2664,7 @@ 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[]>(); + LinkedHashMap[]> variants = new LinkedHashMap<>(); List dnaFeatures = dnaSeq.getFeatures() .getFeaturesByOntology(SequenceOntologyI.SEQUENCE_VARIANT); @@ -2559,18 +2690,26 @@ public class AlignmentUtils } /* - * extract dna variants to a string array + * ignore variant if not a SNP */ String alls = (String) sf.getValue(Gff3Helper.ALLELES); if (alls == null) { continue; // non-SNP VCF variant perhaps - can't process this } + String[] alleles = alls.toUpperCase().split(","); - int i = 0; + boolean isSnp = true; for (String allele : alleles) { - alleles[i++] = allele.trim(); // lose any space characters "A, G" + if (allele.trim().length() > 1) + { + isSnp = false; + } + } + if (!isSnp) + { + continue; } int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol); @@ -2584,9 +2723,9 @@ public class AlignmentUtils if (codonVariants == null) { codonVariants = new ArrayList[CODON_LENGTH]; - codonVariants[0] = new ArrayList(); - codonVariants[1] = new ArrayList(); - codonVariants[2] = new ArrayList(); + codonVariants[0] = new ArrayList<>(); + codonVariants[1] = new ArrayList<>(); + codonVariants[2] = new ArrayList<>(); variants.put(peptidePosition, codonVariants); } @@ -2710,7 +2849,7 @@ 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(); @@ -2770,10 +2909,10 @@ public class AlignmentUtils * true; else returns false * * @param unaligned - * - sequences to be aligned based on aligned + * - sequences to be aligned based on aligned * @param aligned - * - 'guide' alignment containing sequences derived from same dataset - * as unaligned + * - 'guide' alignment containing sequences derived from same + * dataset as unaligned * @return */ static boolean alignAsSameSequences(AlignmentI unaligned, @@ -2785,7 +2924,7 @@ public class AlignmentUtils } // map from dataset sequence to alignment sequence(s) - Map> alignedDatasets = new HashMap>(); + Map> alignedDatasets = new HashMap<>(); for (SequenceI seq : aligned.getSequences()) { SequenceI ds = seq.getDatasetSequence(); @@ -2797,15 +2936,22 @@ public class AlignmentUtils } /* - * first pass - check whether all sequences to be aligned share a dataset - * sequence with an aligned sequence + * first pass - check whether all sequences to be aligned share a + * dataset sequence with an aligned sequence; also note the leftmost + * ungapped column from which to copy */ + int leftmost = Integer.MAX_VALUE; for (SequenceI seq : unaligned.getSequences()) { - if (!alignedDatasets.containsKey(seq.getDatasetSequence())) + final SequenceI ds = seq.getDatasetSequence(); + if (!alignedDatasets.containsKey(ds)) { return false; } + SequenceI alignedSeq = alignedDatasets.get(ds) + .get(0); + int startCol = alignedSeq.findIndex(seq.getStart()); // 1.. + leftmost = Math.min(leftmost, startCol); } /* @@ -2813,13 +2959,25 @@ public class AlignmentUtils * heuristic rule: pair off sequences in order for the case where * more than one shares the same dataset sequence */ + final char gapCharacter = aligned.getGapCharacter(); for (SequenceI seq : unaligned.getSequences()) { 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(alignedSequences.get(0).getSequenceAsString()); + SequenceI alignedSeq = alignedSequences.get(0); + + /* + * gap fill for leading (5') UTR if any + */ + // TODO this copies intron columns - wrong! + int startCol = alignedSeq.findIndex(seq.getStart()); // 1.. + int endCol = alignedSeq.findIndex(seq.getEnd()); + char[] seqchars = new char[endCol - leftmost + 1]; + Arrays.fill(seqchars, gapCharacter); + char[] toCopy = alignedSeq.getSequence(startCol - 1, endCol); + System.arraycopy(toCopy, 0, seqchars, startCol - leftmost, + toCopy.length); + seq.setSequence(String.valueOf(seqchars)); if (alignedSequences.size() > 0) { // pop off aligned sequences (except the last one) @@ -2827,6 +2985,12 @@ public class AlignmentUtils } } + /* + * finally remove gapped columns (e.g. introns) + */ + new RemoveGapColCommand("", unaligned.getSequencesArray(), 0, + unaligned.getWidth() - 1, unaligned); + return true; } @@ -2848,7 +3012,7 @@ public class AlignmentUtils * {unalignedSequence, characterPerSequence} at that position. * TreeMap keeps the entries in ascending column order. */ - SortedMap> map = new TreeMap>(); + SortedMap> map = new TreeMap<>(); /* * record any sequences that have no mapping so can't be realigned @@ -2953,7 +3117,7 @@ public class AlignmentUtils Map seqsMap = map.get(fromCol); if (seqsMap == null) { - seqsMap = new HashMap(); + seqsMap = new HashMap<>(); map.put(fromCol, seqsMap); } seqsMap.put(seq, seq.getCharAt(mappedCharPos - toStart));