import jalview.datamodel.DBRefEntry;
import jalview.datamodel.DBRefSource;
+import jalview.datamodel.Mapping;
import jalview.datamodel.PDBEntry;
import jalview.datamodel.SequenceI;
import java.util.ArrayList;
+import java.util.BitSet;
import java.util.HashMap;
import java.util.HashSet;
-import java.util.Hashtable;
import java.util.List;
import java.util.Map;
*/
public class DBRefUtils
{
+
+ public final static int DB_SOURCE = 1;
+ public final static int DB_VERSION = 2;
+ public final static int DB_ID = 4;
+ public final static int DB_MAP = 8;
+
+
+ public final static int SEARCH_MODE_NO_MAP_NO_VERSION = DB_SOURCE | DB_ID;
+ public final static int SEARCH_MODE_FULL = DB_SOURCE | DB_VERSION | DB_ID | DB_MAP;
+
/*
* lookup from lower-case form of a name to its canonical (standardised) form
*/
canonicalSourceNameLookup.put("uniprotkb/swiss-prot",
DBRefSource.UNIPROT);
canonicalSourceNameLookup.put("uniprotkb/trembl", DBRefSource.UNIPROT);
+
+ // Ensembl values for dbname in xref REST service:
+ canonicalSourceNameLookup.put("uniprot/sptrembl", DBRefSource.UNIPROT);
+ canonicalSourceNameLookup.put("uniprot/swissprot", DBRefSource.UNIPROT);
+
canonicalSourceNameLookup.put("pdb", DBRefSource.PDB);
canonicalSourceNameLookup.put("ensembl", DBRefSource.ENSEMBL);
+ // Ensembl Gn and Tr are for Ensembl genomic and transcript IDs as served
+ // from ENA.
+ canonicalSourceNameLookup.put("ensembl-tr", DBRefSource.ENSEMBL);
+ canonicalSourceNameLookup.put("ensembl-gn", DBRefSource.ENSEMBL);
+
+ // Make sure we have lowercase entries for all canonical string lookups
+// BH 2019.01.25 unnecessary -- they are all lower case already
+ //Set<String> keys = canonicalSourceNameLookup.keySet();
+// for (String k : keys)
+// {
+// canonicalSourceNameLookup.put(k.toLowerCase(),
+// canonicalSourceNameLookup.get(k));
+// }
dasCoordinateSystemsLookup.put("pdbresnum", DBRefSource.PDB);
dasCoordinateSystemsLookup.put("uniprot", DBRefSource.UNIPROT);
}
/**
+ * Returns those DBRefEntry objects whose source identifier (once converted to
+ * Jalview's canonical form) is in the list of sources to search for. Returns
+ * null if no matches found.
*
* @param dbrefs
- * array of DBRef objects to search
+ * DBRefEntry objects to search
* @param sources
- * String[] array of source DBRef IDs to retrieve
+ * array of sources to select
* @return
*/
- public static DBRefEntry[] selectRefs(DBRefEntry[] dbrefs,
+ public static List<DBRefEntry> selectRefs(List<DBRefEntry> dbrefs,
String[] sources)
{
if (dbrefs == null || sources == null)
{
return dbrefs;
}
+
+ // BH TODO
HashSet<String> srcs = new HashSet<String>();
for (String src : sources)
{
- srcs.add(src);
+ srcs.add(src.toUpperCase());
}
+ int nrefs = dbrefs.size();
List<DBRefEntry> res = new ArrayList<DBRefEntry>();
- for (DBRefEntry dbr : dbrefs)
+ for (int ib = 0; ib < nrefs; ib++)
{
+ DBRefEntry dbr = dbrefs.get(ib);
String source = getCanonicalName(dbr.getSource());
- if (srcs.contains(source))
+ if (srcs.contains(source.toUpperCase()))
{
res.add(dbr);
}
if (res.size() > 0)
{
- DBRefEntry[] reply = new DBRefEntry[res.size()];
- return res.toArray(reply);
+ //List<DBRefEntry> reply = new DBRefEntry[res.size()];
+ return res;//.toArray(reply);
}
return null;
}
+ private static boolean selectRefsBS(List<DBRefEntry> dbrefs, String sourceKeys, BitSet bsSelect) {
+ if (dbrefs == null || sourceKeys == null)
+ {
+ return false;
+ }
+ for (int i = 0, n = dbrefs.size(); i < n; i++)
+ {
+ DBRefEntry dbr = dbrefs.get(i);
+ String sourceKey = dbr.getSourceKey();
+ if (sourceKeys.indexOf(sourceKey) < 0) {
+ bsSelect.clear(i);
+ }
+ }
+ return !bsSelect.isEmpty();
+ }
+
+
/**
* isDasCoordinateSystem
*
return false;
}
String coordsys = dasCoordinateSystemsLookup.get(string.toLowerCase());
- return coordsys == null ? false : coordsys.equals(dBRefEntry
- .getSource());
+ return coordsys == null ? false
+ : coordsys.equals(dBRefEntry.getSource());
}
/**
}
/**
- * Returns an array of those references that match the given entry, or null if
- * no matches. Currently uses a comparator which matches if
+ * Returns a (possibly empty) list of those references that match the given
+ * entry. Currently uses a comparator which matches if
* <ul>
* <li>database sources are the same</li>
* <li>accession ids are the same</li>
* Set of references to search
* @param entry
* pattern to match
+ * @param mode SEARCH_MODE_FULL for all; SEARCH_MODE_NO_MAP_NO_VERSION optional
* @return
*/
- public static DBRefEntry[] searchRefs(DBRefEntry[] ref, DBRefEntry entry)
+ public static List<DBRefEntry> searchRefs(List<DBRefEntry> ref,
+ DBRefEntry entry, int mode)
{
return searchRefs(ref, entry,
- matchDbAndIdAndEitherMapOrEquivalentMapList);
+ matchDbAndIdAndEitherMapOrEquivalentMapList, mode);
+ }
+
+ /**
+ * Returns a list of those references that match the given accession id
+ * <ul>
+ * <li>database sources are the same</li>
+ * <li>accession ids are the same</li>
+ * <li>both have no mapping, or the mappings are the same</li>
+ * </ul>
+ *
+ * @param refs
+ * Set of references to search
+ * @param accId
+ * accession id to match
+ * @return
+ */
+ public static List<DBRefEntry> searchRefs(List<DBRefEntry> refs, String accId)
+ {
+ List<DBRefEntry> rfs = new ArrayList<DBRefEntry>();
+ if (refs == null || accId == null)
+ {
+ return rfs;
+ }
+ for (int i = 0, n = refs.size(); i < n; i++)
+ {
+ DBRefEntry e = refs.get(i);
+ if (accId.equals(e.getAccessionId()))
+ {
+ rfs.add(e);
+ }
+ }
+ return rfs;
+// return searchRefs(refs, new DBRefEntry("", "", accId), matchId, SEARCH_MODE_FULL);
}
+
/**
- * Returns an array of those references that match the given entry, according
- * to the given comparator. Returns null if no matches.
+ * Returns a (possibly empty) list of those references that match the given
+ * entry, according to the given comparator.
*
* @param refs
* an array of database references to search
* @param entry
* an entry to compare against
* @param comparator
+ * @param mode SEARCH_MODE_FULL for all; SEARCH_MODE_NO_MAP_NO_VERSION optional
* @return
*/
- static DBRefEntry[] searchRefs(DBRefEntry[] refs, DBRefEntry entry,
- DbRefComp comparator)
+ static List<DBRefEntry> searchRefs(List<DBRefEntry> refs, DBRefEntry entry,
+ DbRefComp comparator, int mode)
{
+ List<DBRefEntry> rfs = new ArrayList<DBRefEntry>();
if (refs == null || entry == null)
{
- return null;
+ return rfs;
}
- List<DBRefEntry> rfs = new ArrayList<DBRefEntry>();
- for (int i = 0; i < refs.length; i++)
+ for (int i = 0, n = refs.size(); i < n; i++)
{
- if (comparator.matches(entry, refs[i]))
+ DBRefEntry e = refs.get(i);
+ if (comparator.matches(entry, e, SEARCH_MODE_FULL))
{
- rfs.add(refs[i]);
+ rfs.add(e);
}
}
- return rfs.size() == 0 ? null : rfs.toArray(new DBRefEntry[rfs.size()]);
+ return rfs;
}
interface DbRefComp
{
- public boolean matches(DBRefEntry refa, DBRefEntry refb);
- }
+ default public boolean matches(DBRefEntry refa, DBRefEntry refb) {
+ return matches(refa, refb, SEARCH_MODE_FULL);
+ };
+
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode);
+}
/**
* match on all non-null fields in refa
// TODO unused - remove?
public static DbRefComp matchNonNullonA = new DbRefComp()
{
- @Override
- public boolean matches(DBRefEntry refa, DBRefEntry refb)
+ @Override
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
{
- if (refa.getSource() == null
- || refb.getSource().equals(refa.getSource()))
+ if ((mode & DB_SOURCE) != 0 && refa.getSource() == null
+ || DBRefUtils.getCanonicalName(refb.getSource()).equals(
+ DBRefUtils.getCanonicalName(refa.getSource())))
{
- if (refa.getVersion() == null
+ if ((mode & DB_VERSION) != 0 && refa.getVersion() == null
|| refb.getVersion().equals(refa.getVersion()))
{
- if (refa.getAccessionId() == null
+ if ((mode & DB_ID) != 0 && refa.getAccessionId() == null
|| refb.getAccessionId().equals(refa.getAccessionId()))
{
- if (refa.getMap() == null
- || (refb.getMap() != null && refb.getMap().equals(
- refa.getMap())))
+ if ((mode & DB_MAP) != 0 && refa.getMap() == null || (refb.getMap() != null
+ && refb.getMap().equals(refa.getMap())))
{
return true;
}
public static DbRefComp matchEitherNonNull = new DbRefComp()
{
@Override
- public boolean matches(DBRefEntry refa, DBRefEntry refb)
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
{
- if (nullOrEqual(refa.getSource(), refb.getSource())
+ if (nullOrEqualSource(refa.getSource(), refb.getSource())
&& nullOrEqual(refa.getVersion(), refb.getVersion())
&& nullOrEqual(refa.getAccessionId(), refb.getAccessionId())
&& nullOrEqual(refa.getMap(), refb.getMap()))
}
return false;
}
+
};
/**
public static DbRefComp matchDbAndIdAndEitherMap = new DbRefComp()
{
@Override
- public boolean matches(DBRefEntry refa, DBRefEntry refb)
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
{
if (refa.getSource() != null && refb.getSource() != null
- && refb.getSource().equals(refa.getSource()))
+ && DBRefUtils.getCanonicalName(refb.getSource()).equals(
+ DBRefUtils.getCanonicalName(refa.getSource())))
{
// We dont care about version
if (refa.getAccessionId() != null && refb.getAccessionId() != null
- // FIXME should be && not || here?
+ // FIXME should be && not || here?
|| refb.getAccessionId().equals(refa.getAccessionId()))
{
if ((refa.getMap() == null || refb.getMap() == null)
- || (refa.getMap() != null && refb.getMap() != null && refb
- .getMap().equals(refa.getMap())))
+ || (refa.getMap() != null && refb.getMap() != null
+ && refb.getMap().equals(refa.getMap())))
{
return true;
}
public static DbRefComp matchDbAndIdAndComplementaryMapList = new DbRefComp()
{
@Override
- public boolean matches(DBRefEntry refa, DBRefEntry refb)
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
{
if (refa.getSource() != null && refb.getSource() != null
- && refb.getSource().equals(refa.getSource()))
+ && DBRefUtils.getCanonicalName(refb.getSource()).equals(
+ DBRefUtils.getCanonicalName(refa.getSource())))
{
// We dont care about version
if (refa.getAccessionId() != null && refb.getAccessionId() != null
if ((refa.getMap() == null && refb.getMap() == null)
|| (refa.getMap() != null && refb.getMap() != null))
{
- if ((refb.getMap().getMap() == null && refa.getMap().getMap() == null)
+ if ((refb.getMap().getMap() == null
+ && refa.getMap().getMap() == null)
|| (refb.getMap().getMap() != null
- && refa.getMap().getMap() != null && refb
- .getMap().getMap().getInverse()
- .equals(refa.getMap().getMap())))
+ && refa.getMap().getMap() != null
+ && refb.getMap().getMap().getInverse()
+ .equals(refa.getMap().getMap())))
{
return true;
}
public static DbRefComp matchDbAndIdAndEquivalentMapList = new DbRefComp()
{
@Override
- public boolean matches(DBRefEntry refa, DBRefEntry refb)
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
{
if (refa.getSource() != null && refb.getSource() != null
- && refb.getSource().equals(refa.getSource()))
+ && DBRefUtils.getCanonicalName(refb.getSource()).equals(
+ DBRefUtils.getCanonicalName(refa.getSource())))
{
// We dont care about version
// if ((refa.getVersion()==null || refb.getVersion()==null)
{
return true;
}
- if (refa.getMap() != null
- && refb.getMap() != null
- && ((refb.getMap().getMap() == null && refa.getMap()
- .getMap() == null) || (refb.getMap().getMap() != null
- && refa.getMap().getMap() != null && refb
- .getMap().getMap().equals(refa.getMap().getMap()))))
+ if (refa.getMap() != null && refb.getMap() != null
+ && ((refb.getMap().getMap() == null
+ && refa.getMap().getMap() == null)
+ || (refb.getMap().getMap() != null
+ && refa.getMap().getMap() != null
+ && refb.getMap().getMap()
+ .equals(refa.getMap().getMap()))))
{
return true;
}
};
/**
- * accession ID and DB must be identical. Version is ignored. No map on either
- * or map but no maplist on either or maplist of map on a is equivalent to the
- * maplist of map on b.
+ * accession ID and DB must be identical, or null on a. Version is ignored. No
+ * map on either or map but no maplist on either or maplist of map on a is
+ * equivalent to the maplist of map on b.
*/
public static DbRefComp matchDbAndIdAndEitherMapOrEquivalentMapList = new DbRefComp()
{
@Override
- public boolean matches(DBRefEntry refa, DBRefEntry refb)
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
{
if (refa.getSource() != null && refb.getSource() != null
- && refb.getSource().equals(refa.getSource()))
+ && DBRefUtils.getCanonicalName(refb.getSource()).equals(
+ DBRefUtils.getCanonicalName(refa.getSource())))
{
// We dont care about version
- if (refa.getAccessionId() != null && refb.getAccessionId() != null
- && refb.getAccessionId().equals(refa.getAccessionId()))
+
+ if (refa.getAccessionId() == null
+ || refa.getAccessionId().equals(refb.getAccessionId()))
{
if (refa.getMap() == null || refb.getMap() == null)
{
return true;
}
if ((refa.getMap() != null && refb.getMap() != null)
- && (refb.getMap().getMap() == null && refa.getMap()
- .getMap() == null)
+ && (refb.getMap().getMap() == null
+ && refa.getMap().getMap() == null)
|| (refb.getMap().getMap() != null
- && refa.getMap().getMap() != null && (refb
- .getMap().getMap().equals(refa.getMap().getMap()))))
- { // getMap().getMap().containsEither(false,refa.getMap().getMap())
+ && refa.getMap().getMap() != null
+ && (refb.getMap().getMap()
+ .equals(refa.getMap().getMap()))))
+ {
return true;
}
}
};
/**
+ * accession ID only must be identical.
+ */
+ public static DbRefComp matchId = new DbRefComp()
+ {
+ @Override
+ public boolean matches(DBRefEntry refa, DBRefEntry refb, int mode)
+ {
+ if (refa.getAccessionId() != null && refb.getAccessionId() != null
+ && refb.getAccessionId().equals(refa.getAccessionId()))
+ {
+ return true;
+ }
+ return false;
+ }
+ };
+
+ /**
* Parses a DBRefEntry and adds it to the sequence, also a PDBEntry if the
* database is PDB.
* <p>
PDBEntry pdbr = new PDBEntry();
pdbr.setId(pdbid);
pdbr.setType(PDBEntry.Type.PDB);
- pdbr.setProperty(new Hashtable());
pdbr.setChainCode(chaincode);
- // pdbr.getProperty().put("CHAIN", chaincode);
seq.addPDBId(pdbr);
}
else
{
return true;
}
- return (o1 == null ? o2.equals(o1) : o1.equals(o2));
+ return o1.equals(o2);
+ }
+
+ /**
+ * canonicalise source string before comparing. null is always wildcard
+ *
+ * @param o1
+ * - null or source string to compare
+ * @param o2
+ * - null or source string to compare
+ * @return true if either o1 or o2 are null, or o1 equals o2 under
+ * DBRefUtils.getCanonicalName
+ * (o1).equals(DBRefUtils.getCanonicalName(o2))
+ */
+ public static boolean nullOrEqualSource(String o1, String o2)
+ {
+ if (o1 == null || o2 == null)
+ {
+ return true;
+ }
+ return DBRefUtils.getCanonicalName(o1)
+ .equals(DBRefUtils.getCanonicalName(o2));
+ }
+
+ /**
+ * Selects just the DNA or protein references from a set of references
+ *
+ * @param selectDna
+ * if true, select references to 'standard' DNA databases, else to
+ * 'standard' peptide databases
+ * @param refs
+ * a set of references to select from
+ * @return
+ */
+ public static List<DBRefEntry> selectDbRefs(boolean selectDna,
+ List<DBRefEntry> refs)
+ {
+ return selectRefs(refs,
+ selectDna ? DBRefSource.DNACODINGDBS : DBRefSource.PROTEINDBS);
+ // could attempt to find other cross
+ // refs here - ie PDB xrefs
+ // (not dna, not protein seq)
+ }
+
+ /**
+ * Returns the (possibly empty) list of those supplied dbrefs which have the
+ * specified source database, with a case-insensitive match of source name
+ *
+ * @param dbRefs
+ * @param source
+ * @return
+ */
+ public static List<DBRefEntry> searchRefsForSource(List<DBRefEntry> dbRefs,
+ String source)
+ {
+ List<DBRefEntry> matches = new ArrayList<DBRefEntry>();
+ if (dbRefs != null && source != null)
+ {
+ for (DBRefEntry dbref : dbRefs)
+ {
+ if (source.equalsIgnoreCase(dbref.getSource()))
+ {
+ matches.add(dbref);
+ }
+ }
+ }
+ return matches;
+ }
+
+ /**
+ * promote direct database references to primary for nucleotide or protein
+ * sequences if they have an appropriate primary ref
+ * <table>
+ * <tr>
+ * <th>Seq Type</th>
+ * <th>Primary DB</th>
+ * <th>Direct which will be promoted</th>
+ * </tr>
+ * <tr align=center>
+ * <td>peptides</td>
+ * <td>Ensembl</td>
+ * <td>Uniprot</td>
+ * </tr>
+ * <tr align=center>
+ * <td>peptides</td>
+ * <td>Ensembl</td>
+ * <td>Uniprot</td>
+ * </tr>
+ * <tr align=center>
+ * <td>dna</td>
+ * <td>Ensembl</td>
+ * <td>ENA</td>
+ * </tr>
+ * </table>
+ *
+ * @param sequence
+ */
+ public static void ensurePrimaries(SequenceI sequence, List<DBRefEntry> pr)
+ {
+ if (pr.size() == 0)
+ {
+ // nothing to do
+ return;
+ }
+ int sstart = sequence.getStart();
+ int send = sequence.getEnd();
+ boolean isProtein = sequence.isProtein();
+ BitSet bsSelect = new BitSet();
+
+// List<DBRefEntry> selfs = new ArrayList<DBRefEntry>();
+// {
+
+// List<DBRefEntry> selddfs = selectDbRefs(!isprot, sequence.getDBRefs());
+// if (selfs == null || selfs.size() == 0)
+// {
+// // nothing to do
+// return;
+// }
+
+ List<DBRefEntry> dbrefs = sequence.getDBRefs();
+ bsSelect.set(0, dbrefs.size());
+
+ if (!selectRefsBS(dbrefs, isProtein ? DBRefSource.PROTEINDBSKEYS : DBRefSource.DNACODINGDBSKEYS, bsSelect))
+ return;
+
+// selfs.addAll(selfArray);
+// }
+
+ // filter non-primary refs
+ for (int ip = pr.size(); --ip >= 0;)
+ {
+ DBRefEntry p = pr.get(ip);
+ for (int i = bsSelect.nextSetBit(0); i >= 0; i = bsSelect.nextSetBit(i + 1)) {
+ if (dbrefs.get(i) == p)
+ bsSelect.clear(i);
+ }
+// while (selfs.contains(p))
+// {
+// selfs.remove(p);
+// }
+ }
+// List<DBRefEntry> toPromote = new ArrayList<DBRefEntry>();
+
+
+
+ for (int ip = pr.size(), keys = 0; --ip >= 0 && keys != DBRefSource.ALL_MASKS;)
+ {
+ DBRefEntry p = pr.get(ip);
+ if (isProtein)
+ {
+ switch (getCanonicalName(p.getSource()))
+ {
+ case DBRefSource.UNIPROT:
+ keys |= DBRefSource.UNIPROT_MASK;
+ break;
+ case DBRefSource.ENSEMBL:
+ keys |= DBRefSource.ENSEMBL_MASK;
+ break;
+ }
+ }
+ else
+ {
+ // TODO: promote transcript refs ??
+ }
+ if (keys == 0 || !selectRefsBS(dbrefs, DBRefSource.PROMTYPES[keys], bsSelect))
+ return;
+// if (candidates != null)
+ {
+ for (int ic = bsSelect.nextSetBit(0); ic >= 0; ic = bsSelect.nextSetBit(ic + 1))
+// for (int ic = 0, n = candidates.size(); ic < n; ic++)
+ {
+ DBRefEntry cand = dbrefs.get(ic);//candidates.get(ic);
+ if (cand.hasMap())
+ {
+ Mapping map = cand.getMap();
+ SequenceI cto = map.getTo();
+ if (cto != null
+ && cto != sequence)
+ {
+ // can't promote refs with mappings to other sequences
+ continue;
+ }
+ MapList mlist = map.getMap();
+ if (mlist.getFromLowest() != sstart
+ && mlist.getFromHighest() != send)
+ {
+ // can't promote refs with mappings from a region of this sequence
+ // - eg CDS
+ continue;
+ }
+ }
+ // and promote
+ cand.setVersion(p.getVersion() + " (promoted)");
+ bsSelect.clear(ic);
+ //selfs.remove(cand);
+// toPromote.add(cand);
+ if (!cand.isPrimaryCandidate())
+ {
+ System.out.println(
+ "Warning: Couldn't promote dbref " + cand.toString()
+ + " for sequence " + sequence.toString());
+ }
+ }
+ }
+ }
}
}