X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fdatamodel%2FMappedFeatures.java;h=0fa03cf0d6c632601da074b2f41e81a51967b9d1;hb=bf68bd00205b1216a5744c9db11b5a8ebc4c7af9;hp=4c63916b54f3d1de6785ccfd1faa1505f6b084f3;hpb=108953c611409ce47ef8b22c21a7854d10ca0fbd;p=jalview.git diff --git a/src/jalview/datamodel/MappedFeatures.java b/src/jalview/datamodel/MappedFeatures.java index 4c63916..0fa03cf 100644 --- a/src/jalview/datamodel/MappedFeatures.java +++ b/src/jalview/datamodel/MappedFeatures.java @@ -5,175 +5,232 @@ import jalview.schemes.ResidueProperties; import jalview.util.MappingUtils; import jalview.util.StringUtils; -import java.util.ArrayList; +import java.util.HashSet; import java.util.List; +import java.util.Set; /** * A data bean to hold a list of mapped sequence features (e.g. CDS features - * mapped from protein), and the mapping between the sequences + * mapped from protein), and the mapping between the sequences. It also provides + * a method to derive peptide variants from codon variants. * * @author gmcarstairs */ public class MappedFeatures { + private static final String HGV_SP = "HGVSp"; + + private static final String CSQ = "CSQ"; + /* - * the mapping from CDS to peptide + * the mapping from one sequence to another */ public final Mapping mapping; /** - * the CDS sequence mapped to + * the sequence mapped from */ public final SequenceI fromSeq; /* - * the residue position in the peptide sequence + * features on the sequence mapped to that overlap the mapped positions */ - public final int fromPosition; + public final List features; /* - * the peptide residue at the position + * the residue position in the sequence mapped to */ - public final char fromResidue; + private final int toPosition; /* - * features on CDS that overlap the codon positions + * the residue at toPosition */ - public final List features; + private final char toResidue; + + /* + * if the mapping is 3:1 or 1:3 (peptide to CDS), this holds the + * mapped positions i.e. codon base positions in CDS; to + * support calculation of peptide variants from alleles + */ + private final int[] codonPos; + + private final char[] baseCodon; /** * Constructor * * @param theMapping + * @param from + * the sequence mapped from (e.g. CDS) * @param pos + * the residue position in the sequence mapped to * @param res + * the residue character at position pos * @param theFeatures + * list of mapped features found in the 'from' sequence at + * the mapped position(s) */ public MappedFeatures(Mapping theMapping, SequenceI from, int pos, - char res, - List theFeatures) + char res, List theFeatures) { mapping = theMapping; fromSeq = from; - fromPosition = pos; - fromResidue = res; + toPosition = pos; + toResidue = res; features = theFeatures; + + /* + * determine codon positions and canonical codon + * for a peptide-to-CDS mapping + */ + int[] codonIntervals = mapping.getMap().locateInFrom(toPosition, toPosition); + int[] codonPositions = codonIntervals == null ? null + : MappingUtils.flattenRanges(codonIntervals); + if (codonPositions != null && codonPositions.length == 3) + { + codonPos = codonPositions; + baseCodon = new char[3]; + int cdsStart = fromSeq.getStart(); + baseCodon[0] = Character + .toUpperCase(fromSeq.getCharAt(codonPos[0] - cdsStart)); + baseCodon[1] = Character + .toUpperCase(fromSeq.getCharAt(codonPos[1] - cdsStart)); + baseCodon[2] = Character + .toUpperCase(fromSeq.getCharAt(codonPos[2] - cdsStart)); + } + else + { + codonPos = null; + baseCodon = null; + } } /** - * Computes and returns a (possibly empty) list of HGVS notation peptide - * variants derived from codon allele variants + * Computes and returns comma-delimited HGVS notation peptide variants derived + * from codon allele variants. If no variants are found, answers an empty + * string. * + * @param sf + * a sequence feature (which must be one of those held in this + * object) * @return */ - public List findProteinVariants() + public String findProteinVariants(SequenceFeature sf) { - List vars = new ArrayList<>(); - if (features.isEmpty()) + if (!features.contains(sf) || baseCodon == null) { - return vars; + return ""; } /* - * determine canonical codon + * VCF data may already contain the protein consequence */ - int[] codonPos = MappingUtils.flattenRanges( - mapping.getMap().locateInFrom(fromPosition, fromPosition)); - if (codonPos.length != 3) + String hgvsp = sf.getValueAsString(CSQ, HGV_SP); + if (hgvsp != null) { - // error - return vars; + int colonPos = hgvsp.lastIndexOf(':'); + if (colonPos >= 0) + { + String var = hgvsp.substring(colonPos + 1); + if (var.contains("p.")) // sanity check + { + return var; + } + } } - final char[] baseCodon = new char[3]; - int cdsStart = fromSeq.getStart(); - baseCodon[0] = fromSeq.getCharAt(codonPos[0] - cdsStart); - baseCodon[1] = fromSeq.getCharAt(codonPos[1] - cdsStart); - baseCodon[2] = fromSeq.getCharAt(codonPos[2] - cdsStart); - for (SequenceFeature sf : features) + /* + * otherwise, compute codon and peptide variant + */ + int cdsPos = sf.getBegin(); + if (cdsPos != sf.getEnd()) { - /* - * VCF data may already contain the protein consequence - */ - String hgvsp = sf.getValueAsString("CSQ", "HGVSp"); - if (hgvsp != null) + // not handling multi-locus variant features + return ""; + } + if (cdsPos != codonPos[0] && cdsPos != codonPos[1] + && cdsPos != codonPos[2]) + { + // e.g. feature on intron within spliced codon! + return ""; + } + + String alls = (String) sf.getValue(Gff3Helper.ALLELES); + if (alls == null) + { + return ""; + } + + String from3 = StringUtils.toSentenceCase( + ResidueProperties.aa2Triplet.get(String.valueOf(toResidue))); + + /* + * make a peptide variant for each SNP allele + * e.g. C,G,T gives variants G and T for base C + */ + Set variantPeptides = new HashSet<>(); + String[] alleles = alls.toUpperCase().split(","); + StringBuilder vars = new StringBuilder(); + + for (String allele : alleles) + { + allele = allele.trim().toUpperCase(); + if (allele.length() > 1 || "-".equals(allele)) { - int colonPos = hgvsp.indexOf(':'); - if (colonPos >= 0) - { - String var = hgvsp.substring(colonPos + 1); - if (!vars.contains(var)) - { - vars.add(var); - } - continue; - } + continue; // multi-locus variant } + char[] variantCodon = new char[3]; + variantCodon[0] = baseCodon[0]; + variantCodon[1] = baseCodon[1]; + variantCodon[2] = baseCodon[2]; /* - * otherwise, compute codon and peptide variant + * poke variant base into canonical codon; + * ignore first 'allele' (canonical base) */ - // todo avoid duplication of code in AlignmentUtils.buildDnaVariantsMap - int cdsPos = sf.getBegin(); - if (cdsPos != sf.getEnd()) - { - // not handling multi-locus variant features - continue; - } - if (cdsPos != codonPos[0] && cdsPos != codonPos[1] - && cdsPos != codonPos[2]) + final int i = cdsPos == codonPos[0] ? 0 + : (cdsPos == codonPos[1] ? 1 : 2); + variantCodon[i] = allele.toUpperCase().charAt(0); + if (variantCodon[i] == baseCodon[i]) { - // e.g. feature on intron within spliced codon! continue; } - - String alls = (String) sf.getValue(Gff3Helper.ALLELES); - if (alls == null) + String codon = new String(variantCodon); + String peptide = ResidueProperties.codonTranslate(codon); + boolean synonymous = toResidue == peptide.charAt(0); + StringBuilder var = new StringBuilder(); + if (synonymous) { - continue; + /* + * synonymous variant notation e.g. c.1062C>A(p.=) + */ + var.append("c.").append(String.valueOf(cdsPos)) + .append(String.valueOf(baseCodon[i])).append(">") + .append(String.valueOf(variantCodon[i])) + .append("(p.=)"); } - String from3 = StringUtils.toSentenceCase( - ResidueProperties.aa2Triplet - .get(String.valueOf(fromResidue))); - - /* - * make a peptide variant for each SNP allele - * e.g. C,G,T gives variants G and T for base C - */ - String[] alleles = alls.toUpperCase().split(","); - for (String allele : alleles) + else { - allele = allele.trim().toUpperCase(); - if (allele.length() > 1 || "-".equals(allele)) - { - continue; // multi-locus variant - } - char[] variantCodon = new char[3]; - variantCodon[0] = baseCodon[0]; - variantCodon[1] = baseCodon[1]; - variantCodon[2] = baseCodon[2]; - /* - * poke variant base into canonical codon + * missense variant notation e.g. p.Arg355Met */ - int i = cdsPos == codonPos[0] ? 0 : (cdsPos == codonPos[1] ? 1 : 2); - variantCodon[i] = allele.toUpperCase().charAt(0); - String codon = new String(variantCodon); - String peptide = ResidueProperties.codonTranslate(codon); - if (fromResidue != peptide.charAt(0)) + String to3 = ResidueProperties.STOP.equals(peptide) ? "Ter" + : StringUtils.toSentenceCase( + ResidueProperties.aa2Triplet.get(peptide)); + var.append("p.").append(from3).append(String.valueOf(toPosition)) + .append(to3); + } + if (!variantPeptides.contains(peptide)) // duplicate consequence + { + variantPeptides.add(peptide); + if (vars.length() > 0) { - String to3 = ResidueProperties.STOP.equals(peptide) ? "STOP" - : StringUtils.toSentenceCase( - ResidueProperties.aa2Triplet.get(peptide)); - String var = "p." + from3 + fromPosition + to3; - if (!vars.contains(var)) - { - vars.add(var); - } + vars.append(","); } + vars.append(var); } } - return vars; + return vars.toString(); } }