package jalview.ext.ensembl;
+import jalview.analysis.AlignmentUtils;
import jalview.datamodel.Alignment;
import jalview.datamodel.AlignmentI;
import jalview.datamodel.DBRefEntry;
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;
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<String> CROSS_REFERENCES = Arrays
+ .asList(new String[] { "CCDS", "Uniprot/SWISSPROT",
+ "Uniprot/SPTREMBL" });
+
protected static final String CONSEQUENCE_TYPE = "consequence_type";
protected static final String PARENT = "Parent";
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");
}
/**
- * 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);
}
/**
@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
+ " 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
addFeaturesAndProduct(accId, alignment);
}
- inProgress = false;
- System.out.println(getClass().getName() + " took "
- + (System.currentTimeMillis() - now) + "ms to fetch");
+ for (SequenceI seq : alignment.getSequences())
+ {
+ getCrossReferences(seq);
+ }
+
return alignment;
}
* 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);
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);
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)
{
}
/**
- * Returns a mapping from dna to protein by inspecting sequence features of
- * type "CDS" on the dna.
+ * Get database xrefs from Ensembl, and attach them to the sequence
*
- * @param dnaSeq
- * @param proteinSeq
- * @return
+ * @param seq
*/
- protected MapList mapCdsToProtein(SequenceI dnaSeq, SequenceI proteinSeq)
+ protected void getCrossReferences(SequenceI seq)
{
- SequenceFeature[] sfs = dnaSeq.getSequenceFeatures();
- if (sfs == null)
+ while (seq.getDatasetSequence() != null)
{
- return null;
+ seq = seq.getDatasetSequence();
}
- List<int[]> ranges = new ArrayList<int[]>(50);
- SequenceOntology so = SequenceOntology.getInstance();
-
- int mappedDnaLength = 0;
-
- /*
- * Map CDS columns of dna to peptide. 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.
- */
- boolean fivePrimeIncomplete = false;
- for (SequenceFeature sf : sfs)
+ EnsemblXref xrefFetcher = new EnsemblXref(getDomain());
+ List<DBRefEntry> xrefs = xrefFetcher.getCrossReferences(seq.getName(),
+ getCrossReferenceDatabases());
+ for (DBRefEntry xref : xrefs)
{
+ seq.addDBRef(xref);
/*
- * process a CDS feature (or a sub-type of CDS)
+ * Save any Uniprot xref to be the reference for SIFTS mapping
*/
- if (so.isA(sf.getType(), SequenceOntology.CDS))
+ if (DBRefSource.UNIPROT.equals(xref.getSource()))
{
- 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)
- {
- fivePrimeIncomplete = true;
- begin += phase;
- if (begin > end)
- {
- continue; // shouldn't happen?
- }
- }
- ranges.add(new int[] { begin, end });
- mappedDnaLength += Math.abs(end - begin) + 1;
+ seq.setSourceDBRef(xref);
}
}
- int proteinLength = proteinSeq.getLength();
- List<int[]> proteinRange = new ArrayList<int[]>();
- int proteinStart = 1;
- if (fivePrimeIncomplete && 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;
+ /**
+ * 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
+ */
+ protected List<String> getCrossReferenceDatabases()
+ {
+ return CROSS_REFERENCES;
}
/**
* 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));
* 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();
*/
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;
*/
Collections.sort(regions, new RangeSorter(direction == 1));
- List<int[]> to = new ArrayList<int[]>();
- to.add(new int[] { start, start + mappedLength - 1 });
+ List<int[]> to = Arrays.asList(new int[] { start,
+ start + mappedLength - 1 });
return new MapList(regions, to, 1, 1);
}
/*
* 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);
+ // }
+ // }
}
}
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;
}
/**
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))
{
}
/**
- * 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();
- }
-
- mapExonsToProtein(dnaSeq, peptide, dnaToProtein);
-
- LinkedHashMap<Integer, String[][]> variants = buildDnaVariantsMap(
- dnaSeq, dnaToProtein);
-
- /*
- * scan codon variations, compute peptide variants and add to peptide sequence
- */
- int count = 0;
- for (Entry<Integer, String[][]> variant : variants.entrySet())
- {
- int peptidePos = variant.getKey();
- String[][] codonVariants = variant.getValue();
- String residue = String.valueOf(peptide.getCharAt(peptidePos - 1)); // 0-based
- List<String> peptideVariants = computePeptideVariants(codonVariants,
- residue);
- if (!peptideVariants.isEmpty())
- {
- Collections.sort(peptideVariants);
- String desc = StringUtils.listToDelimitedString(peptideVariants,
- ", ");
- SequenceFeature sf = new SequenceFeature(
- SequenceOntology.SEQUENCE_VARIANT, desc, peptidePos,
- peptidePos, 0f, null);
- peptide.addSequenceFeature(sf);
- count++;
- }
- }
- return count;
- }
-
- /**
- * Transfers exon features to the corresponding mapped regions of the protein
- * sequence. This is useful because it allows visualisation of exon boundaries
- * on the peptide (using 'colour by label' for the exon name). Returns the
- * number of features written.
- *
- * @param dnaSeq
- * @param peptide
- * @param dnaToProtein
- */
- static int mapExonsToProtein(SequenceI dnaSeq, SequenceI peptide,
- MapList dnaToProtein)
- {
- SequenceFeature[] sfs = dnaSeq.getSequenceFeatures();
- if (sfs == null)
- {
- return 0;
- }
-
- SequenceOntology so = SequenceOntology.getInstance();
- int count = 0;
-
- for (SequenceFeature sf : sfs)
- {
- if (so.isA(sf.getType(), SequenceOntology.EXON))
- {
- int start = sf.getBegin();
- int end = sf.getEnd();
- int[] mapsTo = dnaToProtein.locateInTo(start, end);
- if (mapsTo != null)
- {
- SequenceFeature copy = new SequenceFeature(SequenceOntology.EXON,
- sf.getDescription(), mapsTo[0], mapsTo[1], 0f, null);
- peptide.addSequenceFeature(copy);
- 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<Integer, String[][]> 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<Integer, String[][]> variants = new LinkedHashMap<Integer, String[][]>();
- SequenceOntology so = SequenceOntology.getInstance();
-
- SequenceFeature[] dnaFeatures = dnaSeq.getSequenceFeatures();
- if (dnaFeatures == null)
- {
- return variants;
- }
-
- 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] - 1));
- 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 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 for codon positions 1, 2, 3
- * @param residue
- * the current residue translation
- * @return
- */
- static List<String> computePeptideVariants(
- String[][] codonVariants, String residue)
- {
- List<String> result = new ArrayList<String>();
- 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.equals(residue))
- {
- result.add(peptide);
- }
- }
- }
- }
- 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
*/
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);
}
}
git://source.jalview.org / jalview.git / blobdiff