X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fext%2Fensembl%2FEnsemblSeqProxy.java;h=4af6525c7a3bdb97475f1766f63986206d12e57f;hb=a064561d8665ee9db217b17cda826fceac90cbbc;hp=0bfeda1121211a99e65976ef798a8853065f2182;hpb=949fed115506ff00c221669e096546f9c39a6ace;p=jalview.git diff --git a/src/jalview/ext/ensembl/EnsemblSeqProxy.java b/src/jalview/ext/ensembl/EnsemblSeqProxy.java index 0bfeda1..4af6525 100644 --- a/src/jalview/ext/ensembl/EnsemblSeqProxy.java +++ b/src/jalview/ext/ensembl/EnsemblSeqProxy.java @@ -11,12 +11,10 @@ import jalview.datamodel.SequenceI; import jalview.exceptions.JalviewException; import jalview.io.FastaFile; import jalview.io.FileParse; -import jalview.io.gff.SequenceOntology; -import jalview.schemes.ResidueProperties; +import jalview.io.gff.SequenceOntologyFactory; +import jalview.io.gff.SequenceOntologyI; import jalview.util.DBRefUtils; import jalview.util.MapList; -import jalview.util.MappingUtils; -import jalview.util.StringUtils; import java.io.IOException; import java.net.MalformedURLException; @@ -25,19 +23,19 @@ import java.util.ArrayList; import java.util.Arrays; import java.util.Collections; import java.util.Comparator; -import java.util.LinkedHashMap; import java.util.List; -import java.util.Map.Entry; /** * Base class for Ensembl sequence fetchers * + * @see http://rest.ensembl.org/documentation/info/sequence_id * @author gmcarstairs */ public abstract class EnsemblSeqProxy extends EnsemblRestClient { private static final List CROSS_REFERENCES = Arrays - .asList(new String[] { "CCDS" }); + .asList(new String[] { "CCDS", "Uniprot/SWISSPROT", + "Uniprot/SPTREMBL" }); protected static final String CONSEQUENCE_TYPE = "consequence_type"; @@ -45,33 +43,32 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient protected static final String ID = "ID"; - /* - * this needs special handling, as it isA sequence_variant in the - * Sequence Ontology, but behaves in Ensembl as if it isA transcript - */ - protected static final String NMD_VARIANT = "NMD_transcript_variant"; - protected static final String NAME = "Name"; + protected static final String DESCRIPTION = "description"; + + /* + * enum for 'type' parameter to the /sequence REST service + */ public enum EnsemblSeqType { /** - * type=genomic for the full dna including introns + * type=genomic to fetch full dna including introns */ GENOMIC("genomic"), /** - * type=cdna for transcribed dna including UTRs + * type=cdna to fetch dna including UTRs */ CDNA("cdna"), /** - * type=cds for coding dna excluding UTRs + * type=cds to fetch coding dna excluding UTRs */ CDS("cds"), /** - * type=protein for the peptide product sequence + * type=protein to fetch peptide product sequence */ PROTEIN("protein"); @@ -114,10 +111,19 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient } /** - * Constructor + * Default constructor (to use rest.ensembl.org) */ public EnsemblSeqProxy() { + super(); + } + + /** + * Constructor given the target domain to fetch data from + */ + public EnsemblSeqProxy(String d) + { + super(d); } /** @@ -127,7 +133,6 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient @Override public AlignmentI getSequenceRecords(String query) throws Exception { - long now = System.currentTimeMillis(); // TODO use a String... query vararg instead? // danger: accession separator used as a regex here, a string elsewhere @@ -156,17 +161,15 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient + " chunks. Unexpected problem (" + r.getLocalizedMessage() + ")"; System.err.println(msg); - if (alignment != null) - { - break; // return what we got - } - else - { - throw new JalviewException(msg, r); - } + break; } } + if (alignment == null) + { + return null; + } + /* * fetch and transfer genomic sequence features, * fetch protein product and add as cross-reference @@ -181,8 +184,6 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient getCrossReferences(seq); } - System.out.println(getClass().getName() + " took " - + (System.currentTimeMillis() - now) + "ms to fetch"); return alignment; } @@ -208,7 +209,7 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient * get 'dummy' genomic sequence with exon, cds and variation features */ SequenceI genomicSequence = null; - EnsemblOverlap gffFetcher = new EnsemblOverlap(); + EnsemblFeatures gffFetcher = new EnsemblFeatures(getDomain()); EnsemblFeatureType[] features = getFeaturesToFetch(); AlignmentI geneFeatures = gffFetcher.getSequenceRecords(accId, features); @@ -258,10 +259,11 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient String accId = querySeq.getName(); try { - AlignmentI protein = new EnsemblProtein().getSequenceRecords(accId); + AlignmentI protein = new EnsemblProtein(getDomain()) + .getSequenceRecords(accId); if (protein == null || protein.getHeight() == 0) { - System.out.println("Failed to retrieve protein for " + accId); + System.out.println("No protein product found for " + accId); return; } SequenceI proteinSeq = protein.getSequenceAt(0); @@ -272,19 +274,22 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient proteinSeq.createDatasetSequence(); querySeq.createDatasetSequence(); - MapList mapList = mapCdsToProtein(querySeq, proteinSeq); + MapList mapList = AlignmentUtils.mapCdsToProtein(querySeq, proteinSeq); if (mapList != null) { - Mapping map = new Mapping(proteinSeq.getDatasetSequence(), mapList); + // clunky: ensure Uniprot xref if we have one is on mapped sequence + SequenceI ds = proteinSeq.getDatasetSequence(); + ds.setSourceDBRef(proteinSeq.getSourceDBRef()); + Mapping map = new Mapping(ds, mapList); DBRefEntry dbr = new DBRefEntry(getDbSource(), getDbVersion(), accId, map); querySeq.getDatasetSequence().addDBRef(dbr); /* - * compute peptide variants from dna variants and add as - * sequence features on the protein sequence ta-da + * copy exon features to protein, compute peptide variants from dna + * variants and add as features on the protein sequence ta-da */ - computeProteinFeatures(querySeq, proteinSeq, mapList); + AlignmentUtils.computeProteinFeatures(querySeq, proteinSeq, mapList); } } catch (Exception e) { @@ -295,8 +300,7 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient } /** - * Get Uniprot and PDB xrefs from Ensembl, and attach them to the protein - * sequence + * Get database xrefs from Ensembl, and attach them to the sequence * * @param seq */ @@ -307,17 +311,26 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient seq = seq.getDatasetSequence(); } - EnsemblXref xrefFetcher = new EnsemblXref(); + EnsemblXref xrefFetcher = new EnsemblXref(getDomain()); List xrefs = xrefFetcher.getCrossReferences(seq.getName(), getCrossReferenceDatabases()); for (DBRefEntry xref : xrefs) { seq.addDBRef(xref); + /* + * Save any Uniprot xref to be the reference for SIFTS mapping + */ + if (DBRefSource.UNIPROT.equals(xref.getSource())) + { + seq.setSourceDBRef(xref); + } } } /** * Returns a list of database names to be used when fetching cross-references. + * Specifically, the names are used to filter data returned by the Ensembl + * xrefs REST service on the value in field 'dbname'. * * @return */ @@ -327,101 +340,6 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient } /** - * Returns a mapping from dna to protein by inspecting sequence features of - * type "CDS" on the dna. - * - * @param dnaSeq - * @param proteinSeq - * @return - */ - protected MapList mapCdsToProtein(SequenceI dnaSeq, SequenceI proteinSeq) - { - List ranges = new ArrayList(50); - - int mappedDnaLength = getCdsRanges(dnaSeq, ranges); - - int proteinLength = proteinSeq.getLength(); - List proteinRange = new ArrayList(); - int proteinStart = 1; - - /* - * incomplete start codon may mean X at start of peptide - * we ignore both for mapping purposes - */ - if (proteinSeq.getCharAt(0) == 'X') - { - proteinStart = 2; - proteinLength--; - } - proteinRange.add(new int[] { proteinStart, proteinLength }); - - /* - * dna length should map to protein (or protein plus stop codon) - */ - int codesForResidues = mappedDnaLength / 3; - if (codesForResidues == proteinLength - || codesForResidues == (proteinLength + 1)) - { - return new MapList(ranges, proteinRange, 3, 1); - } - return null; - } - - /** - * Adds CDS ranges to the ranges list, and returns the total length mapped. - * - * No need to worry about reverse strand dna here since the retrieved sequence - * is as transcribed (reverse complement for reverse strand), i.e in the same - * sense as the peptide. - * - * @param dnaSeq - * @param ranges - * @return - */ - protected int getCdsRanges(SequenceI dnaSeq, List ranges) - { - SequenceFeature[] sfs = dnaSeq.getSequenceFeatures(); - if (sfs == null) - { - return 0; - } - int mappedDnaLength = 0; - for (SequenceFeature sf : sfs) - { - /* - * process a CDS feature (or a sub-type of CDS) - */ - if (SequenceOntology.getInstance().isA(sf.getType(), SequenceOntology.CDS)) - { - int phase = 0; - try { - phase = Integer.parseInt(sf.getPhase()); - } catch (NumberFormatException e) - { - // 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 (ranges.isEmpty() && phase > 0) - { - begin += phase; - if (begin > end) - { - continue; // shouldn't happen? - } - } - ranges.add(new int[] { begin, end }); - mappedDnaLength += Math.abs(end - begin) + 1; - } - } - return mappedDnaLength; - } - - /** * Fetches sequences for the list of accession ids and adds them to the * alignment. Returns the extended (or created) alignment. * @@ -505,7 +423,7 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient * multiple ids go in the POST body instead */ StringBuffer urlstring = new StringBuffer(128); - urlstring.append(SEQUENCE_ID_URL); + urlstring.append(getDomain() + "/sequence/id"); if (ids.size() == 1) { urlstring.append("/").append(ids.get(0)); @@ -579,7 +497,7 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient * the start position of the sequence we are mapping to * @return */ - protected MapList getGenomicRanges(SequenceI sourceSequence, + protected MapList getGenomicRangesFromFeatures(SequenceI sourceSequence, String accId, int start) { SequenceFeature[] sfs = sourceSequence.getSequenceFeatures(); @@ -605,11 +523,12 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient */ if (identifiesSequence(sf, accId)) { - int strand = sf.getStrand(); - - if (directionSet && strand != direction) - { - // abort - mix of forward and backward + int strand = sf.getStrand(); + strand = strand == 0 ? 1 : strand; // treat unknown as forward + + if (directionSet && strand != direction) + { + // abort - mix of forward and backward System.err.println("Error: forward and backward strand for " + accId); return null; @@ -654,8 +573,8 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient */ Collections.sort(regions, new RangeSorter(direction == 1)); - List to = new ArrayList(); - to.add(new int[] { start, start + mappedLength - 1 }); + List to = Arrays.asList(new int[] { start, + start + mappedLength - 1 }); return new MapList(regions, to, 1, 1); } @@ -710,17 +629,18 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient /* * for sequence_variant, make an additional feature with consequence */ - if (SequenceOntology.getInstance().isSequenceVariant(sf.getType())) - { - String consequence = (String) sf.getValue(CONSEQUENCE_TYPE); - if (consequence != null) - { - SequenceFeature sf2 = new SequenceFeature("consequence", - consequence, copy.getBegin(), copy.getEnd(), 0f, - null); - targetSequence.addSequenceFeature(sf2); - } - } + // if (SequenceOntologyFactory.getInstance().isA(sf.getType(), + // SequenceOntologyI.SEQUENCE_VARIANT)) + // { + // String consequence = (String) sf.getValue(CONSEQUENCE_TYPE); + // if (consequence != null) + // { + // SequenceFeature sf2 = new SequenceFeature("consequence", + // consequence, copy.getBegin(), copy.getEnd(), 0f, + // null); + // targetSequence.addSequenceFeature(sf2); + // } + // } } } @@ -740,15 +660,22 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient return false; } + // long start = System.currentTimeMillis(); SequenceFeature[] sfs = sourceSequence.getSequenceFeatures(); - MapList mapping = getGenomicRanges(sourceSequence, accessionId, + MapList mapping = getGenomicRangesFromFeatures(sourceSequence, accessionId, targetSequence.getStart()); if (mapping == null) { return false; } - return transferFeatures(sfs, targetSequence, mapping, accessionId); + boolean result = transferFeatures(sfs, targetSequence, mapping, + accessionId); + // System.out.println("transferFeatures (" + (sfs.length) + " --> " + // + targetSequence.getSequenceFeatures().length + ") to " + // + targetSequence.getName() + // + " took " + (System.currentTimeMillis() - start) + "ms"); + return result; } /** @@ -850,7 +777,7 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient SequenceFeature[] sfs = sequence.getSequenceFeatures(); if (sfs != null) { - SequenceOntology so = SequenceOntology.getInstance(); + SequenceOntologyI so = SequenceOntologyFactory.getInstance(); for (SequenceFeature sf :sfs) { if (so.isA(sf.getType(), type)) { @@ -866,224 +793,6 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient } /** - * Maps exon features from dna to protein, and computes variants in peptide - * product generated by variants in dna, and adds them as sequence_variant - * features on the protein sequence. Returns the number of variant features - * added. - * - * @param dnaSeq - * @param peptide - * @param dnaToProtein - */ - static int computeProteinFeatures(SequenceI dnaSeq, - SequenceI peptide, MapList dnaToProtein) - { - while (dnaSeq.getDatasetSequence() != null) - { - dnaSeq = dnaSeq.getDatasetSequence(); - } - while (peptide.getDatasetSequence() != null) - { - peptide = peptide.getDatasetSequence(); - } - - AlignmentUtils.transferFeatures(dnaSeq, peptide, dnaToProtein, - SequenceOntology.EXON); - - LinkedHashMap variants = buildDnaVariantsMap( - dnaSeq, dnaToProtein); - - /* - * scan codon variations, compute peptide variants and add to peptide sequence - */ - int count = 0; - for (Entry variant : variants.entrySet()) - { - int peptidePos = variant.getKey(); - String[][] codonVariants = variant.getValue(); - String residue = String.valueOf(peptide.getCharAt(peptidePos - 1)); // 0-based - List peptideVariants = computePeptideVariants(codonVariants, - residue); - if (!peptideVariants.isEmpty()) - { - String desc = StringUtils.listToDelimitedString(peptideVariants, - ", "); - SequenceFeature sf = new SequenceFeature( - SequenceOntology.SEQUENCE_VARIANT, desc, peptidePos, - peptidePos, 0f, null); - peptide.addSequenceFeature(sf); - count++; - } - } - return count; - } - - /** - * Builds a map whose key is position in the protein sequence, and value is an - * array of all variants for the coding codon positions - * - * @param dnaSeq - * @param dnaToProtein - * @return - */ - static LinkedHashMap buildDnaVariantsMap( - SequenceI dnaSeq, MapList dnaToProtein) - { - /* - * map from peptide position to all variant features of the codon for it - * LinkedHashMap ensures we add the peptide features in sequence order - */ - LinkedHashMap variants = new LinkedHashMap(); - SequenceOntology so = SequenceOntology.getInstance(); - - SequenceFeature[] dnaFeatures = dnaSeq.getSequenceFeatures(); - if (dnaFeatures == null) - { - return variants; - } - - int dnaStart = dnaSeq.getStart(); - int[] lastCodon = null; - int lastPeptidePostion = 0; - - /* - * build a map of codon variations for peptides - */ - for (SequenceFeature sf : dnaFeatures) - { - int dnaCol = sf.getBegin(); - if (dnaCol != sf.getEnd()) - { - // not handling multi-locus variant features - continue; - } - if (so.isSequenceVariant(sf.getType())) - { - int[] mapsTo = dnaToProtein.locateInTo(dnaCol, dnaCol); - if (mapsTo == null) - { - // feature doesn't lie within coding region - continue; - } - int peptidePosition = mapsTo[0]; - String[][] codonVariants = variants.get(peptidePosition); - if (codonVariants == null) - { - codonVariants = new String[3][]; - variants.put(peptidePosition, codonVariants); - } - - /* - * extract dna variants to a string array - */ - String alls = (String) sf.getValue("alleles"); - if (alls == null) - { - continue; - } - String[] alleles = alls.split(","); - - /* - * get this peptides 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 this variant) for each codon position - */ - for (int codonPos = 0; codonPos < 3; codonPos++) - { - String nucleotide = String.valueOf(dnaSeq - .getCharAt(codon[codonPos] - dnaStart)); - if (codon[codonPos] == dnaCol) - { - /* - * record current dna base and its alleles - */ - String[] dnaVariants = new String[alleles.length + 1]; - dnaVariants[0] = nucleotide; - System.arraycopy(alleles, 0, dnaVariants, 1, alleles.length); - codonVariants[codonPos] = dnaVariants; - } - else if (codonVariants[codonPos] == null) - { - /* - * record current dna base only - * (at least until we find any variation and overwrite it) - */ - codonVariants[codonPos] = new String[] { nucleotide }; - } - } - } - } - return variants; - } - - /** - * Returns a sorted, non-redundant list of all peptide translations generated - * by the given dna variants, excluding the current residue value - * - * @param codonVariants - * an array of base values (acgtACGT) for codon positions 1, 2, 3 - * @param residue - * the current residue translation - * @return - */ - static List computePeptideVariants( - String[][] codonVariants, String residue) - { - List result = new ArrayList(); - for (String base1 : codonVariants[0]) - { - for (String base2 : codonVariants[1]) - { - for (String base3 : codonVariants[2]) - { - String codon = base1 + base2 + base3; - // TODO: report frameshift/insertion/deletion - // and multiple-base variants?! - String peptide = codon.contains("-") ? "-" : ResidueProperties - .codonTranslate(codon); - if (peptide != null && !result.contains(peptide) - && !peptide.equalsIgnoreCase(residue)) - { - result.add(peptide); - } - } - } - } - - /* - * sort alphabetically with STOP at the end - */ - Collections.sort(result, new Comparator() - { - - @Override - public int compare(String o1, String o2) - { - if ("STOP".equals(o1)) - { - return 1; - } - else if ("STOP".equals(o2)) - { - return -1; - } - else - { - return o1.compareTo(o2); - } - } - }); - return result; - } - - /** * Answers true if the feature type is either 'NMD_transcript_variant' or * 'transcript' or one of its sub-types in the Sequence Ontology. This is * needed because NMD_transcript_variant behaves like 'transcript' in Ensembl @@ -1095,7 +804,8 @@ public abstract class EnsemblSeqProxy extends EnsemblRestClient */ public static boolean isTranscript(String featureType) { - return NMD_VARIANT.equals(featureType) - || SequenceOntology.getInstance().isA(featureType, SequenceOntology.TRANSCRIPT); + return SequenceOntologyI.NMD_TRANSCRIPT_VARIANT.equals(featureType) + || SequenceOntologyFactory.getInstance().isA(featureType, + SequenceOntologyI.TRANSCRIPT); } }