import jalview.datamodel.DBRefEntry;
import jalview.datamodel.DBRefSource;
import jalview.datamodel.Mapping;
+import jalview.datamodel.Sequence;
import jalview.datamodel.SequenceFeature;
import jalview.datamodel.SequenceI;
import jalview.datamodel.features.SequenceFeatures;
import jalview.exceptions.JalviewException;
-import jalview.io.FastaFile;
-import jalview.io.FileParse;
+import jalview.io.gff.Gff3Helper;
import jalview.io.gff.SequenceOntologyFactory;
import jalview.io.gff.SequenceOntologyI;
import jalview.util.Comparison;
import jalview.util.DBRefUtils;
import jalview.util.IntRangeComparator;
import jalview.util.MapList;
+import jalview.util.Platform;
import java.io.IOException;
import java.net.MalformedURLException;
import java.util.Collections;
import java.util.List;
import java.util.Map;
-import java.util.Map.Entry;
+
+import org.json.simple.parser.ParseException;
/**
* Base class for Ensembl sequence fetchers
*/
public abstract class EnsemblSeqProxy extends EnsemblRestClient
{
- private static final String ALLELES = "alleles";
-
- protected static final String PARENT = "Parent";
-
- protected static final String ID = "ID";
-
- protected static final String NAME = "Name";
-
protected static final String DESCRIPTION = "description";
/*
try
{
/*
- * get 'dummy' genomic sequence with exon, cds and variation features
+ * get 'dummy' genomic sequence with gene, transcript,
+ * exon, cds and variation features
*/
SequenceI genomicSequence = null;
EnsemblFeatures gffFetcher = new EnsemblFeatures(getDomain());
/*
* transfer features to the query sequence
*/
- SequenceI querySeq = alignment.findName(accId);
+ SequenceI querySeq = alignment.findName(accId, true);
if (transferFeatures(accId, genomicSequence, querySeq))
{
String accId = querySeq.getName();
try
{
+ System.out.println("Adding protein product for " + accId);
AlignmentI protein = new EnsemblProtein(getDomain())
.getSequenceRecords(accId);
if (protein == null || protein.getHeight() == 0)
inProgress = false;
throw new JalviewException("ENSEMBL Rest API not available.");
}
- FileParse fp = getSequenceReader(ids);
- if (fp == null)
- {
- return alignment;
- }
+ Platform.timeCheck("EnsemblSeqProx.fetchSeq ", Platform.TIME_MARK);
+
+ List<SequenceI> seqs = parseSequenceJson(ids);
+ if (seqs == null)
+ return alignment;
- FastaFile fr = new FastaFile(fp);
- if (fr.hasWarningMessage())
+ if (seqs.isEmpty())
{
- System.out.println(
- String.format("Warning when retrieving %d ids %s\n%s",
- ids.size(), ids.toString(), fr.getWarningMessage()));
+ throw new IOException("No data returned for " + ids);
}
- else if (fr.getSeqs().size() != ids.size())
+
+ if (seqs.size() != ids.size())
{
System.out.println(String.format(
"Only retrieved %d sequences for %d query strings",
- fr.getSeqs().size(), ids.size()));
+ seqs.size(), ids.size()));
}
- if (fr.getSeqs().size() == 1 && fr.getSeqs().get(0).getLength() == 0)
+ if (!seqs.isEmpty())
{
- /*
- * POST request has returned an empty FASTA file e.g. for invalid id
- */
- throw new IOException("No data returned for " + ids);
- }
-
- if (fr.getSeqs().size() > 0)
- {
- AlignmentI seqal = new Alignment(fr.getSeqsAsArray());
- for (SequenceI sq : seqal.getSequences())
+ AlignmentI seqal = new Alignment(
+ seqs.toArray(new SequenceI[seqs.size()]));
+ for (SequenceI seq : seqs)
{
- if (sq.getDescription() == null)
+ if (seq.getDescription() == null)
{
- sq.setDescription(getDbName());
+ seq.setDescription(getDbName());
}
- String name = sq.getName();
+ String name = seq.getName();
if (ids.contains(name)
|| ids.contains(name.replace("ENSP", "ENST")))
{
- DBRefEntry dbref = DBRefUtils.parseToDbRef(sq, getDbSource(),
+ // TODO JAL-3077 use true accession version in dbref
+ DBRefEntry dbref = DBRefUtils.parseToDbRef(seq, getDbSource(),
getEnsemblDataVersion(), name);
- sq.addDBRef(dbref);
+ seq.addDBRef(dbref);
}
}
if (alignment == null)
}
/**
+ * Parses a JSON response for a single sequence ID query
+ *
+ * @param br
+ * @return a single jalview.datamodel.Sequence
+ * @see http://rest.ensembl.org/documentation/info/sequence_id
+ */
+ @SuppressWarnings("unchecked")
+ protected List<SequenceI> parseSequenceJson(List<String> ids)
+ {
+ List<SequenceI> result = new ArrayList<>();
+ try
+ {
+ /*
+ * for now, assumes only one sequence returned; refactor if needed
+ * in future to handle a JSONArray with more than one
+ */
+ Map<String, Object> val = (Map<String, Object>) getJSON(null, ids, -1, MODE_MAP, null);
+ if (val == null)
+ return null;
+ Object s = val.get("desc");
+ String desc = s == null ? null : s.toString();
+ s = val.get("id");
+ String id = s == null ? null : s.toString();
+ s = val.get("seq");
+ String seq = s == null ? null : s.toString();
+ Sequence sequence = new Sequence(id, seq);
+ if (desc != null)
+ {
+ sequence.setDescription(desc);
+ }
+ // todo JAL-3077 make a DBRefEntry with true accession version
+ // s = val.get("version");
+ // String version = s == null ? "0" : s.toString();
+ // DBRefEntry dbref = new DBRefEntry(getDbSource(), version, id);
+ // sequence.addDBRef(dbref);
+ result.add(sequence);
+ } catch (ParseException | IOException e)
+ {
+ System.err.println("Error processing JSON response: " + e.toString());
+ // ignore
+ }
+ return result;
+ }
+
+ /**
* Returns the URL for the REST call
*
* @return
}
// @see https://github.com/Ensembl/ensembl-rest/wiki/Output-formats
urlstring.append("?type=").append(getSourceEnsemblType().getType());
- urlstring.append(("&Accept=text/x-fasta"));
+ urlstring.append(("&Accept=application/json"));
+ urlstring.append(("&content-type=application/json"));
- Map<String, String> params = getAdditionalParameters();
- if (params != null)
+ String objectType = getObjectType();
+ if (objectType != null)
{
- for (Entry<String, String> entry : params.entrySet())
- {
- urlstring.append("&").append(entry.getKey()).append("=")
- .append(entry.getValue());
- }
+ urlstring.append("&").append(OBJECT_TYPE).append("=")
+ .append(objectType);
}
URL url = new URL(urlstring.toString());
}
/**
- * Override this method to add any additional x=y URL parameters needed
+ * Override this method to specify object_type request parameter
*
* @return
*/
- protected Map<String, String> getAdditionalParameters()
+ protected String getObjectType()
{
return null;
}
return false;
}
- @Override
- protected String getRequestMimeType(boolean multipleIds)
- {
- return multipleIds ? "application/json" : "text/x-fasta";
- }
-
- @Override
- protected String getResponseMimeType()
- {
- return "text/x-fasta";
- }
-
/**
*
* @return the configured sequence return type for this source
protected MapList getGenomicRangesFromFeatures(SequenceI sourceSequence,
String accId, int start)
{
- // SequenceFeature[] sfs = sourceSequence.getSequenceFeatures();
- List<SequenceFeature> sfs = sourceSequence.getFeatures()
- .getPositionalFeatures();
+ List<SequenceFeature> sfs = getIdentifyingFeatures(sourceSequence,
+ accId);
if (sfs.isEmpty())
{
return null;
* generously initial size for number of cds regions
* (worst case titin Q8WZ42 has c. 313 exons)
*/
- List<int[]> regions = new ArrayList<int[]>(100);
+ List<int[]> regions = new ArrayList<>(100);
int mappedLength = 0;
int direction = 1; // forward
boolean directionSet = false;
for (SequenceFeature sf : sfs)
{
+ 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;
+ }
+ direction = strand;
+ directionSet = true;
+
/*
- * accept the target feature type or a specialisation of it
- * (e.g. coding_exon for exon)
+ * add to CDS ranges, semi-sorted forwards/backwards
*/
- if (identifiesSequence(sf, accId))
+ if (strand < 0)
{
- 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;
- }
- direction = strand;
- directionSet = true;
-
- /*
- * add to CDS ranges, semi-sorted forwards/backwards
- */
- if (strand < 0)
- {
- regions.add(0, new int[] { sf.getEnd(), sf.getBegin() });
- }
- else
- {
- regions.add(new int[] { sf.getBegin(), sf.getEnd() });
- }
- mappedLength += Math.abs(sf.getEnd() - sf.getBegin() + 1);
-
- if (!isSpliceable())
- {
- /*
- * 'gene' sequence is contiguous so we can stop as soon as its
- * identifying feature has been found
- */
- break;
- }
+ regions.add(0, new int[] { sf.getEnd(), sf.getBegin() });
}
+ else
+ {
+ regions.add(new int[] { sf.getBegin(), sf.getEnd() });
+ }
+ mappedLength += Math.abs(sf.getEnd() - sf.getBegin() + 1);
}
if (regions.isEmpty())
}
/**
- * Answers true if the sequence being retrieved may occupy discontiguous
- * regions on the genomic sequence.
- */
- protected boolean isSpliceable()
- {
- return true;
- }
-
- /**
- * Returns true if the sequence feature marks positions of the genomic
+ * Answers a list of sequence features that mark positions of the genomic
* sequence feature which are within the sequence being retrieved. For
* example, an 'exon' feature whose parent is the target transcript marks the
- * cdna positions of the transcript.
+ * cdna positions of the transcript. For a gene sequence, this is trivially
+ * just the 'gene' feature with matching gene id.
*
- * @param sf
+ * @param seq
* @param accId
* @return
*/
- protected abstract boolean identifiesSequence(SequenceFeature sf,
- String accId);
+ protected abstract List<SequenceFeature> getIdentifyingFeatures(
+ SequenceI seq, String accId);
/**
* Transfers the sequence feature to the target sequence, locating its start
*/
static void reverseComplementAlleles(SequenceFeature sf)
{
- final String alleles = (String) sf.getValue(ALLELES);
+ final String alleles = (String) sf.getValue(Gff3Helper.ALLELES);
if (alleles == null)
{
return;
reverseComplementAllele(complement, allele);
}
String comp = complement.toString();
- sf.setValue(ALLELES, comp);
+ sf.setValue(Gff3Helper.ALLELES, comp);
sf.setDescription(comp);
/*
String atts = sf.getAttributes();
if (atts != null)
{
- atts = atts.replace(ALLELES + "=" + alleles, ALLELES + "=" + comp);
+ atts = atts.replace(Gff3Helper.ALLELES + "=" + alleles,
+ Gff3Helper.ALLELES + "=" + comp);
sf.setAttributes(atts);
}
}
return false;
}
- long start = System.currentTimeMillis();
- // SequenceFeature[] sfs = sourceSequence.getSequenceFeatures();
+// long start = System.currentTimeMillis();
List<SequenceFeature> sfs = sourceSequence.getFeatures()
.getPositionalFeatures();
MapList mapping = getGenomicRangesFromFeatures(sourceSequence,
boolean result = transferFeatures(sfs, targetSequence, mapping,
accessionId);
- System.out.println("transferFeatures (" + (sfs.size()) + " --> "
- + targetSequence.getFeatures().getFeatureCount(true) + ") to "
- + targetSequence.getName() + " took "
- + (System.currentTimeMillis() - start) + "ms");
+// System.out.println("transferFeatures (" + (sfs.size()) + " --> "
+// + targetSequence.getFeatures().getFeatureCount(true) + ") to "
+// + targetSequence.getName() + " took "
+// + (System.currentTimeMillis() - start) + "ms");
return result;
}
protected boolean featureMayBelong(SequenceFeature sf, String identifier)
{
String parent = (String) sf.getValue(PARENT);
- // using contains to allow for prefix "gene:", "transcript:" etc
- if (parent != null && !parent.contains(identifier))
+ if (parent != null
+ && !parent.equalsIgnoreCase(identifier))
{
// this genomic feature belongs to a different transcript
return false;
return true;
}
+ /**
+ * Answers a short description of the sequence fetcher
+ */
@Override
public String getDescription()
{
protected List<SequenceFeature> findFeatures(SequenceI sequence,
String term, String parentId)
{
- List<SequenceFeature> result = new ArrayList<SequenceFeature>();
+ List<SequenceFeature> result = new ArrayList<>();
List<SequenceFeature> sfs = sequence.getFeatures()
.getFeaturesByOntology(term);
for (SequenceFeature sf : sfs)
{
String parent = (String) sf.getValue(PARENT);
- if (parent != null && parent.equals(parentId))
+ if (parent != null && parent.equalsIgnoreCase(parentId))
{
result.add(sf);
}
/**
* 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
+ * 'transcript' (or one of its sub-types in the Sequence Ontology). This is
+ * because NMD_transcript_variant behaves like 'transcript' in Ensembl
* although strictly speaking it is not (it is a sub-type of
* sequence_variant).
+ * <p>
+ * (This test was needed when fetching transcript features as GFF. As we are
+ * now fetching as JSON, all features have type 'transcript' so the check for
+ * NMD_transcript_variant is redundant. Left in for any future case arising.)
*
* @param featureType
* @return