X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FAlignmentUtils.java;h=f1084a7d74e40b05120a189540077ee1fddc29d0;hb=947109cebef3e98693ffc0a70b5bffcbd32217e2;hp=228446bc7cdb8162d58d4e40b349141ebaf7c6ed;hpb=e464da0ea20b0322846ce623598a16f2a1cba7ae;p=jalview.git diff --git a/src/jalview/analysis/AlignmentUtils.java b/src/jalview/analysis/AlignmentUtils.java index 228446b..f1084a7 100644 --- a/src/jalview/analysis/AlignmentUtils.java +++ b/src/jalview/analysis/AlignmentUtils.java @@ -24,7 +24,6 @@ import static jalview.io.gff.GffConstants.CLINICAL_SIGNIFICANCE; import jalview.datamodel.AlignedCodon; import jalview.datamodel.AlignedCodonFrame; -import jalview.datamodel.AlignedCodonFrame.SequenceToSequenceMapping; import jalview.datamodel.Alignment; import jalview.datamodel.AlignmentAnnotation; import jalview.datamodel.AlignmentI; @@ -36,6 +35,7 @@ import jalview.datamodel.Sequence; import jalview.datamodel.SequenceFeature; import jalview.datamodel.SequenceGroup; import jalview.datamodel.SequenceI; +import jalview.datamodel.SequenceMapping; import jalview.datamodel.features.SequenceFeatures; import jalview.io.gff.Gff3Helper; import jalview.io.gff.SequenceOntologyI; @@ -74,12 +74,15 @@ 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 @@ -128,7 +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()) { @@ -258,7 +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(); @@ -267,7 +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); @@ -294,8 +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 @@ -465,7 +468,7 @@ 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)) { @@ -536,7 +539,8 @@ 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; } @@ -568,7 +572,8 @@ 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; } @@ -881,7 +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); @@ -1092,7 +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()) @@ -1102,7 +1107,7 @@ public class AlignmentUtils SequenceI prot = mapping.findAlignedSequence(dnaSeq, protein); if (prot != null) { - Mapping seqMap = mapping.getMappingForSequence(dnaSeq); + Mapping seqMap = mapping.getMappingForSequence(dnaSeq, false); addCodonPositions(dnaSeq, prot, protein.getGapCharacter(), seqMap, alignedCodons); unmappedProtein.remove(prot); @@ -1138,9 +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()) @@ -1319,7 +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); @@ -1456,7 +1461,7 @@ public class AlignmentUtils { continue; } - final List result = new ArrayList(); + final List result = new ArrayList<>(); for (AlignmentAnnotation dsann : datasetAnnotations) { /* @@ -1638,13 +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 @@ -1886,7 +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, @@ -1911,7 +1916,15 @@ 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; + } } /* @@ -1922,7 +1935,7 @@ public class AlignmentUtils .findMappingsForSequence(proteinProduct, mappings); for (AlignedCodonFrame acf : mappingsToPeptide) { - for (SequenceToSequenceMapping map : acf.getMappings()) + for (SequenceMapping map : acf.getMappings()) { Mapping mapping = map.getMapping(); if (mapping != aMapping @@ -1936,10 +1949,12 @@ 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()) @@ -2054,9 +2069,11 @@ public class AlignmentUtils 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(); + List direct = new ArrayList<>(); + HashSet directSources = new HashSet<>(); + if (contig.getDBRefs() != null) { for (DBRefEntry dbr : contig.getDBRefs()) @@ -2076,7 +2093,7 @@ 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) @@ -2217,7 +2234,10 @@ 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 @@ -2229,6 +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(); @@ -2243,7 +2273,7 @@ public class AlignmentUtils proteinStart++; proteinLength--; } - List proteinRange = new ArrayList(); + List proteinRange = new ArrayList<>(); /* * dna length should map to protein (or protein plus stop codon) @@ -2252,8 +2282,12 @@ 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 }); @@ -2264,7 +2298,7 @@ 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. @@ -2274,7 +2308,7 @@ public class AlignmentUtils */ protected static List findCdsPositions(SequenceI dnaSeq) { - List result = new ArrayList(); + List result = new ArrayList<>(); List sfs = dnaSeq.getFeatures().getFeaturesByOntology( SequenceOntologyI.CDS); @@ -2283,7 +2317,6 @@ public class AlignmentUtils return result; } SequenceFeatures.sortFeatures(sfs, true); - int startPhase = 0; for (SequenceFeature sf : sfs) { @@ -2301,7 +2334,7 @@ public class AlignmentUtils */ int begin = sf.getBegin(); int end = sf.getEnd(); - if (result.isEmpty()) + if (result.isEmpty() && phase > 0) { begin += phase; if (begin > end) @@ -2316,16 +2349,6 @@ 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 @@ -2405,7 +2428,8 @@ public class AlignmentUtils 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; @@ -2423,10 +2447,17 @@ 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++; + } } } } @@ -2445,10 +2476,17 @@ 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++; + } } } } @@ -2467,10 +2505,17 @@ 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++; + } } } } @@ -2481,20 +2526,22 @@ 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 @@ -2502,57 +2549,71 @@ 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 + { + 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; } /** @@ -2574,7 +2635,7 @@ 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); @@ -2633,9 +2694,9 @@ 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); } @@ -2759,7 +2820,7 @@ 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(); @@ -2834,7 +2895,7 @@ 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(); @@ -2897,7 +2958,7 @@ 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 @@ -3002,7 +3063,7 @@ 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));