+/*
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ *
+ * This file is part of Jalview.
+ *
+ * Jalview is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *
+ * Jalview is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ * PURPOSE. See the GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
+ */
package jalview.ext.ensembl;
+import java.io.IOException;
+import java.net.MalformedURLException;
+import java.net.URL;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.List;
+import java.util.Map;
+
+import org.json.simple.parser.ParseException;
+
import jalview.analysis.AlignmentUtils;
+import jalview.analysis.Dna;
+import jalview.bin.Console;
import jalview.datamodel.Alignment;
import jalview.datamodel.AlignmentI;
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.schemes.ResidueProperties;
+import jalview.util.Comparison;
import jalview.util.DBRefUtils;
+import jalview.util.IntRangeComparator;
import jalview.util.MapList;
-import jalview.util.MappingUtils;
-import jalview.util.StringUtils;
-
-import java.io.IOException;
-import java.net.MalformedURLException;
-import java.net.URL;
-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<String> CROSS_REFERENCES = Arrays
- .asList(new String[] { "CCDS", "Uniprot/SWISSPROT" });
-
- protected static final String CONSEQUENCE_TYPE = "consequence_type";
-
- protected static final String PARENT = "Parent";
-
- protected static final String ID = "ID";
-
- protected static final String NAME = "Name";
+ protected static final String DESCRIPTION = "description";
/*
* enum for 'type' parameter to the /sequence REST service
GENOMIC("genomic"),
/**
- * type=cdna to fetch dna including UTRs
+ * type=cdna to fetch coding dna including UTRs
*/
CDNA("cdna"),
}
/**
- * A comparator to sort ranges into ascending start position order
+ * Default constructor (to use rest.ensembl.org)
*/
- private class RangeSorter implements Comparator<int[]>
+ public EnsemblSeqProxy()
{
- boolean forwards;
-
- RangeSorter(boolean forward)
- {
- forwards = forward;
- }
-
- @Override
- public int compare(int[] o1, int[] o2)
- {
- return (forwards ? 1 : -1) * Integer.compare(o1[0], o2[0]);
- }
-
+ super();
}
/**
- * Constructor
+ * Constructor given the target domain to fetch data from
*/
- public EnsemblSeqProxy()
+ public EnsemblSeqProxy(String d)
{
+ super(d);
}
/**
// danger: accession separator used as a regex here, a string elsewhere
// in this case it is ok (it is just a space), but (e.g.) '\' would not be
- List<String> allIds = Arrays.asList(query
- .split(getAccessionSeparator()));
+ List<String> allIds = Arrays
+ .asList(query.split(getAccessionSeparator()));
AlignmentI alignment = null;
inProgress = true;
+ " chunks. Unexpected problem (" + r.getLocalizedMessage()
+ ")";
System.err.println(msg);
+ r.printStackTrace();
break;
}
}
* fetch and transfer genomic sequence features,
* fetch protein product and add as cross-reference
*/
- for (String accId : allIds)
+ for (int i = 0, n = allIds.size(); i < n; i++)
{
- addFeaturesAndProduct(accId, alignment);
+ addFeaturesAndProduct(allIds.get(i), alignment);
}
- for (SequenceI seq : alignment.getSequences())
+ List<SequenceI> seqs = alignment.getSequences();
+ for (int i = 0, n = seqs.size(); i < n; i++)
{
- getCrossReferences(seq);
+ getCrossReferences(seqs.get(i));
}
return alignment;
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();
+ EnsemblFeatures gffFetcher = new EnsemblFeatures(getDomain());
EnsemblFeatureType[] features = getFeaturesToFetch();
+
+ // Platform.timeCheck("ESP.getsequencerec1", Platform.TIME_MARK);
+
AlignmentI geneFeatures = gffFetcher.getSequenceRecords(accId,
features);
- if (geneFeatures.getHeight() > 0)
+ if (geneFeatures != null && geneFeatures.getHeight() > 0)
{
genomicSequence = geneFeatures.getSequenceAt(0);
}
+
+ // Platform.timeCheck("ESP.getsequencerec2", Platform.TIME_MARK);
+
if (genomicSequence != null)
{
/*
* transfer features to the query sequence
*/
- SequenceI querySeq = alignment.findName(accId);
+ SequenceI querySeq = alignment.findName(accId, true);
if (transferFeatures(accId, genomicSequence, querySeq))
{
* fetch and map protein product, and add it as a cross-reference
* of the retrieved sequence
*/
+ // Platform.timeCheck("ESP.transferFeatures", Platform.TIME_MARK);
addProteinProduct(querySeq);
}
}
} catch (IOException e)
{
- System.err.println("Error transferring Ensembl features: "
- + e.getMessage());
+ System.err.println(
+ "Error transferring Ensembl features: " + e.getMessage());
}
+ // Platform.timeCheck("ESP.addfeat done", Platform.TIME_MARK);
}
/**
String accId = querySeq.getName();
try
{
- AlignmentI protein = new EnsemblProtein().getSequenceRecords(accId);
+ System.out.println("Adding protein product for " + 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)
{
// clunky: ensure Uniprot xref if we have one is on mapped sequence
SequenceI ds = proteinSeq.getDatasetSequence();
- ds.setSourceDBRef(proteinSeq.getSourceDBRef());
+ // TODO: Verify ensp primary ref is on proteinSeq.getDatasetSequence()
Mapping map = new Mapping(ds, mapList);
- DBRefEntry dbr = new DBRefEntry(getDbSource(), getDbVersion(),
- accId, map);
+ DBRefEntry dbr = new DBRefEntry(getDbSource(),
+ getEnsemblDataVersion(), proteinSeq.getName(), map);
querySeq.getDatasetSequence().addDBRef(dbr);
-
+ List<DBRefEntry> uprots = DBRefUtils.selectRefs(ds.getDBRefs(),
+ new String[]
+ { DBRefSource.UNIPROT });
+ List<DBRefEntry> upxrefs = DBRefUtils
+ .selectRefs(querySeq.getDBRefs(), new String[]
+ { DBRefSource.UNIPROT });
+ if (uprots != null)
+ {
+ for (DBRefEntry up : uprots)
+ {
+ // locate local uniprot ref and map
+ List<DBRefEntry> upx = DBRefUtils.searchRefs(upxrefs,
+ up.getAccessionId());
+ DBRefEntry upxref;
+ if (upx.size() != 0)
+ {
+ upxref = upx.get(0);
+
+ if (upx.size() > 1)
+ {
+ Console.warn(
+ "Implementation issue - multiple uniprot acc on product sequence.");
+ }
+ }
+ else
+ {
+ upxref = new DBRefEntry(DBRefSource.UNIPROT,
+ getEnsemblDataVersion(), up.getAccessionId());
+ }
+
+ Mapping newMap = new Mapping(ds, mapList);
+ upxref.setVersion(getEnsemblDataVersion());
+ upxref.setMap(newMap);
+ if (upx.size() == 0)
+ {
+ // add the new uniprot ref
+ querySeq.getDatasetSequence().addDBRef(upxref);
+ }
+
+ }
+ }
+
/*
- * 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);
+ // JAL-3187 render on the fly instead
+ // AlignmentUtils.computeProteinFeatures(querySeq, proteinSeq, mapList);
}
} catch (Exception e)
{
*/
protected void getCrossReferences(SequenceI seq)
{
+
+ // Platform.timeCheck("ESP. getdataseq ", Platform.TIME_MARK);
+
while (seq.getDatasetSequence() != null)
{
seq = seq.getDatasetSequence();
}
- EnsemblXref xrefFetcher = new EnsemblXref();
- List<DBRefEntry> 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.
- *
- * @return
- */
- protected List<String> getCrossReferenceDatabases()
- {
- return CROSS_REFERENCES;
- }
-
- /**
- * 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<int[]> ranges = new ArrayList<int[]>(50);
-
- int mappedDnaLength = getCdsRanges(dnaSeq, ranges);
+ // Platform.timeCheck("ESP. getxref ", Platform.TIME_MARK);
- int proteinLength = proteinSeq.getLength();
- int proteinEnd = proteinLength;
- int proteinStart = 1;
+ EnsemblXref xrefFetcher = new EnsemblXref(getDomain(), getDbSource(),
+ getEnsemblDataVersion());
+ List<DBRefEntry> xrefs = xrefFetcher.getCrossReferences(seq.getName());
- /*
- * incomplete start codon may mean X at start of peptide
- * we ignore both for mapping purposes
- */
- if (proteinSeq.getCharAt(0) == 'X')
+ for (int i = 0, n = xrefs.size(); i < n; i++)
{
- proteinStart = 2;
- proteinLength--;
+ // Platform.timeCheck("ESP. getxref + " + (i) + "/" + n,
+ // Platform.TIME_MARK);
+ // BH 2019.01.25 this next method was taking 174 ms PER addition for a
+ // 266-reference example.
+ // DBRefUtils.ensurePrimaries(seq)
+ // was at the end of seq.addDBRef, so executed after ever addition!
+ // This method was moved to seq.getPrimaryDBRefs()
+ seq.addDBRef(xrefs.get(i));
}
- List<int[]> proteinRange = new ArrayList<int[]>();
+ // System.out.println("primaries are " + seq.getPrimaryDBRefs().toString());
/*
- * dna length should map to protein (or protein plus stop codon)
+ * and add a reference to itself
*/
- int codesForResidues = mappedDnaLength / 3;
- if (codesForResidues == (proteinLength + 1))
- {
- MappingUtils.unmapStopCodon(ranges, mappedDnaLength);
- codesForResidues--;
- }
- if (codesForResidues == proteinLength)
- {
- proteinRange.add(new int[] { proteinStart, proteinEnd });
- return new MapList(ranges, proteinRange, 3, 1);
- }
- return null;
- }
- /**
- * Adds CDS ranges to the ranges list, and returns the total length mapped
- * from.
- *
- * 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<int[]> ranges)
- {
- SequenceFeature[] sfs = dnaSeq.getSequenceFeatures();
- if (sfs == null)
- {
- return 0;
- }
- SequenceOntologyI so = SequenceOntologyFactory.getInstance();
- int mappedDnaLength = 0;
- for (SequenceFeature sf : sfs)
- {
- /*
- * process a CDS feature (or a sub-type of CDS)
- */
- if (so.isA(sf.getType(), SequenceOntologyI.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())
- {
- begin += phase;
- if (begin > end)
- {
- continue; // shouldn't happen?
- }
- }
- ranges.add(new int[] { begin, end });
- mappedDnaLength += Math.abs(end - begin) + 1;
- }
- }
- return mappedDnaLength;
+ // Platform.timeCheck("ESP. getxref self ", Platform.TIME_MARK);
+
+ DBRefEntry self = new DBRefEntry(getDbSource(), getEnsemblDataVersion(),
+ seq.getName());
+
+ // Platform.timeCheck("ESP. getxref self add ", Platform.TIME_MARK);
+
+ seq.addDBRef(self);
+
+ // Platform.timeCheck("ESP. seqprox done ", Platform.TIME_MARK);
}
/**
* @throws JalviewException
* @throws IOException
*/
- protected AlignmentI fetchSequences(List<String> ids, AlignmentI alignment)
- throws JalviewException, IOException
+ protected AlignmentI fetchSequences(List<String> ids,
+ AlignmentI alignment) throws JalviewException, IOException
{
if (!isEnsemblAvailable())
{
inProgress = false;
throw new JalviewException("ENSEMBL Rest API not available.");
}
- FileParse fp = getSequenceReader(ids);
- FastaFile fr = new FastaFile(fp);
- if (fr.hasWarningMessage())
- {
- System.out.println(String.format(
- "Warning when retrieving %d ids %s\n%s", ids.size(),
- ids.toString(), fr.getWarningMessage()));
- }
- else if (fr.getSeqs().size() != ids.size())
+ // Platform.timeCheck("EnsemblSeqProx.fetchSeq ", Platform.TIME_MARK);
+
+ List<SequenceI> seqs = parseSequenceJson(ids);
+ if (seqs == null)
+ return alignment;
+
+ if (seqs.isEmpty())
{
- System.out.println(String.format(
- "Only retrieved %d sequences for %d query strings", fr
- .getSeqs().size(), ids.size()));
+ throw new IOException("No data returned for " + ids);
}
- if (fr.getSeqs().size() == 1 && fr.getSeqs().get(0).getLength() == 0)
+ if (seqs.size() != ids.size())
{
- /*
- * POST request has returned an empty FASTA file e.g. for invalid id
- */
- throw new IOException("No data returned for " + ids);
+ System.out.println(String.format(
+ "Only retrieved %d sequences for %d query strings",
+ seqs.size(), ids.size()));
}
- if (fr.getSeqs().size() > 0)
+ if (!seqs.isEmpty())
{
AlignmentI seqal = new Alignment(
- fr.getSeqsAsArray());
- for (SequenceI sq:seqal.getSequences())
+ 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")))
{
- DBRefUtils.parseToDbRef(sq, DBRefSource.ENSEMBL, "0", name);
+ // TODO JAL-3077 use true accession version in dbref
+ DBRefEntry dbref = DBRefUtils.parseToDbRef(seq, getDbSource(),
+ getEnsemblDataVersion(), name);
+ 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
+ */
+ // Platform.timeCheck("ENS seqproxy", Platform.TIME_MARK);
+ 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
+ }
+ // Platform.timeCheck("ENS seqproxy2", Platform.TIME_MARK);
+ return result;
+ }
+
+ /**
* Returns the URL for the REST call
*
* @return
* 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));
}
// @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"));
+
+ String objectType = getObjectType();
+ if (objectType != null)
+ {
+ urlstring.append("&").append(OBJECT_TYPE).append("=")
+ .append(objectType);
+ }
URL url = new URL(urlstring.toString());
return url;
}
/**
+ * Override this method to specify object_type request parameter
+ *
+ * @return
+ */
+ protected String getObjectType()
+ {
+ return null;
+ }
+
+ /**
* A sequence/id POST request currently allows up to 50 queries
*
* @see http://rest.ensembl.org/documentation/info/sequence_id_post
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();
- if (sfs == null)
+ 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())
{
System.out.println("Failed to identify target sequence for " + accId
* a final sort is needed since Ensembl returns CDS sorted within source
* (havana / ensembl_havana)
*/
- Collections.sort(regions, new RangeSorter(direction == 1));
-
- List<int[]> to = Arrays.asList(new int[] { start,
- start + mappedLength - 1 });
-
- return new MapList(regions, to, 1, 1);
- }
+ Collections.sort(regions, direction == 1 ? IntRangeComparator.ASCENDING
+ : IntRangeComparator.DESCENDING);
- /**
- * Answers true if the sequence being retrieved may occupy discontiguous
- * regions on the genomic sequence.
- */
- protected boolean isSpliceable()
- {
- return true;
+ List<int[]> to = Arrays
+ .asList(new int[]
+ { start, start + mappedLength - 1 });
+
+ return new MapList(regions, to, 1, 1);
}
/**
- * 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);
+
+ int bhtest = 0;
/**
* Transfers the sequence feature to the target sequence, locating its start
* @param mapping
* mapping from the sequence feature's coordinates to the target
* sequence
+ * @param forwardStrand
*/
protected void transferFeature(SequenceFeature sf,
- SequenceI targetSequence, MapList mapping)
+ SequenceI targetSequence, MapList mapping, boolean forwardStrand)
{
int start = sf.getBegin();
int end = sf.getEnd();
int[] mappedRange = mapping.locateInTo(start, end);
-
+
if (mappedRange != null)
{
- SequenceFeature copy = new SequenceFeature(sf);
- copy.setBegin(Math.min(mappedRange[0], mappedRange[1]));
- copy.setEnd(Math.max(mappedRange[0], mappedRange[1]));
+ // Platform.timeCheck(null, Platform.TIME_SET);
+ String group = sf.getFeatureGroup();
+ if (".".equals(group))
+ {
+ group = getDbSource();
+ }
+ int newBegin = Math.min(mappedRange[0], mappedRange[1]);
+ int newEnd = Math.max(mappedRange[0], mappedRange[1]);
+ // Platform.timeCheck(null, Platform.TIME_MARK);
+ bhtest++;
+ // 280 ms/1000 here:
+ SequenceFeature copy = new SequenceFeature(sf, newBegin, newEnd,
+ group, sf.getScore());
+ // 0.175 ms here:
targetSequence.addSequenceFeature(copy);
/*
- * for sequence_variant, make an additional feature with consequence
+ * for sequence_variant on reverse strand, have to convert the allele
+ * values to their complements
*/
- // 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);
- // }
- // }
+ if (!forwardStrand && SequenceOntologyFactory.getInstance()
+ .isA(sf.getType(), SequenceOntologyI.SEQUENCE_VARIANT))
+ {
+ reverseComplementAlleles(copy);
+ }
+ }
+ }
+
+ /**
+ * Change the 'alleles' value of a feature by converting to complementary
+ * bases, and also update the feature description to match
+ *
+ * @param sf
+ */
+ static void reverseComplementAlleles(SequenceFeature sf)
+ {
+ final String alleles = (String) sf.getValue(Gff3Helper.ALLELES);
+ if (alleles == null)
+ {
+ return;
+ }
+ StringBuilder complement = new StringBuilder(alleles.length());
+ for (String allele : alleles.split(","))
+ {
+ reverseComplementAllele(complement, allele);
+ }
+ String comp = complement.toString();
+ sf.setValue(Gff3Helper.ALLELES, comp);
+ sf.setDescription(comp);
+ }
+
+ /**
+ * Makes the 'reverse complement' of the given allele and appends it to the
+ * buffer, after a comma separator if not the first
+ *
+ * @param complement
+ * @param allele
+ */
+ static void reverseComplementAllele(StringBuilder complement,
+ String allele)
+ {
+ if (complement.length() > 0)
+ {
+ complement.append(",");
+ }
+
+ /*
+ * some 'alleles' are actually descriptive terms
+ * e.g. HGMD_MUTATION, PhenCode_variation
+ * - we don't want to 'reverse complement' these
+ */
+ if (!Comparison.isNucleotideSequence(allele, true))
+ {
+ complement.append(allele);
+ }
+ else
+ {
+ for (int i = allele.length() - 1; i >= 0; i--)
+ {
+ complement.append(Dna.getComplement(allele.charAt(i)));
+ }
}
}
}
// long start = System.currentTimeMillis();
- SequenceFeature[] sfs = sourceSequence.getSequenceFeatures();
- MapList mapping = getGenomicRangesFromFeatures(sourceSequence, accessionId,
- targetSequence.getStart());
+ List<SequenceFeature> sfs = sourceSequence.getFeatures()
+ .getPositionalFeatures();
+ MapList mapping = getGenomicRangesFromFeatures(sourceSequence,
+ accessionId, targetSequence.getStart());
if (mapping == null)
{
return false;
}
+ // Platform.timeCheck("ESP. xfer " + sfs.size(), Platform.TIME_MARK);
+
boolean result = transferFeatures(sfs, targetSequence, mapping,
accessionId);
- // System.out.println("transferFeatures (" + (sfs.length) + " --> "
- // + targetSequence.getSequenceFeatures().length + ") 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;
}
* converted using the mapping. Features which do not overlap are ignored.
* Features whose parent is not the specified identifier are also ignored.
*
- * @param features
+ * @param sfs
* @param targetSequence
* @param mapping
* @param parentId
* @return
*/
- protected boolean transferFeatures(SequenceFeature[] features,
+ protected boolean transferFeatures(List<SequenceFeature> sfs,
SequenceI targetSequence, MapList mapping, String parentId)
{
final boolean forwardStrand = mapping.isFromForwardStrand();
/*
- * sort features by start position (descending if reverse strand)
- * before transferring (in forwards order) to the target sequence
+ * sort features by start position (which corresponds to end
+ * position descending if reverse strand) so as to add them in
+ * 'forwards' order to the target sequence
*/
- Arrays.sort(features, new Comparator<SequenceFeature>()
- {
- @Override
- public int compare(SequenceFeature o1, SequenceFeature o2)
- {
- int c = Integer.compare(o1.getBegin(), o2.getBegin());
- return forwardStrand ? c : -c;
- }
- });
+ SequenceFeatures.sortFeatures(sfs, forwardStrand);
boolean transferred = false;
- for (SequenceFeature sf : features)
+
+ for (int i = 0, n = sfs.size(); i < n; i++)
{
+
+ // if ((i%1000) == 0) {
+ //// Platform.timeCheck("Feature " + bhtest, Platform.TIME_GET);
+ // Platform.timeCheck("ESP. xferFeature + " + (i) + "/" + n,
+ // Platform.TIME_MARK);
+ // }
+
+ SequenceFeature sf = sfs.get(i);
if (retainFeature(sf, parentId))
{
- transferFeature(sf, targetSequence, mapping);
+ transferFeature(sf, targetSequence, mapping, forwardStrand);
transferred = true;
}
}
+
return transferred;
}
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()
{
/**
* Returns a (possibly empty) list of features on the sequence which have the
- * specified sequence ontology type (or a sub-type of it), and the given
+ * specified sequence ontology term (or a sub-type of it), and the given
* identifier as parent
*
* @param sequence
- * @param type
+ * @param term
* @param parentId
* @return
*/
protected List<SequenceFeature> findFeatures(SequenceI sequence,
- String type, String parentId)
+ String term, String parentId)
{
- List<SequenceFeature> result = new ArrayList<SequenceFeature>();
-
- SequenceFeature[] sfs = sequence.getSequenceFeatures();
- if (sfs != null) {
- SequenceOntologyI so = SequenceOntologyFactory.getInstance();
- for (SequenceFeature sf :sfs) {
- if (so.isA(sf.getType(), type))
- {
- String parent = (String) sf.getValue(PARENT);
- if (parent.equals(parentId))
- {
- result.add(sf);
- }
- }
- }
- }
- return result;
- }
+ List<SequenceFeature> result = new ArrayList<>();
- /**
- * 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,
- SequenceOntologyI.EXON);
-
- 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())
- {
- String desc = StringUtils.listToDelimitedString(peptideVariants,
- ", ");
- SequenceFeature sf = new SequenceFeature(
- SequenceOntologyI.SEQUENCE_VARIANT, desc, peptidePos,
- peptidePos, 0f, null);
- peptide.addSequenceFeature(sf);
- count++;
- }
- }
-
- /*
- * ugly sort to get sequence features in start position order
- * - would be better to store in Sequence as a TreeSet instead?
- */
- Arrays.sort(peptide.getSequenceFeatures(),
- new Comparator<SequenceFeature>()
- {
- @Override
- public int compare(SequenceFeature o1, SequenceFeature o2)
- {
- int c = Integer.compare(o1.getBegin(), o2.getBegin());
- return c == 0 ? Integer.compare(o1.getEnd(), o2.getEnd())
- : c;
- }
- });
- 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[][]>();
- SequenceOntologyI so = SequenceOntologyFactory.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.isA(sf.getType(), SequenceOntologyI.SEQUENCE_VARIANT))
- {
- 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<String> computePeptideVariants(
- String[][] codonVariants, String residue)
- {
- List<String> result = new ArrayList<String>();
- for (String base1 : codonVariants[0])
+ List<SequenceFeature> sfs = sequence.getFeatures()
+ .getFeaturesByOntology(term);
+ for (SequenceFeature sf : sfs)
{
- for (String base2 : codonVariants[1])
+ String parent = (String) sf.getValue(PARENT);
+ if (parent != null && parent.equalsIgnoreCase(parentId))
{
- 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);
- }
- }
+ result.add(sf);
}
}
- /*
- * sort alphabetically with STOP at the end
- */
- Collections.sort(result, new Comparator<String>()
- {
-
- @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
+ * '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