add ability to pull gene and protein aliases from uniprot

emckee2006 · emckee2006 · commit 6c6cdaaef0bd · 2016-08-24T23:46:33.000-04:00
diff --git a/biojava-core/src/main/java/org/biojava/nbio/core/sequence/io/GenericInsdcHeaderFormat.java b/biojava-core/src/main/java/org/biojava/nbio/core/sequence/io/GenericInsdcHeaderFormat.java
@@ -26,6 +26,7 @@
 import org.biojava.nbio.core.sequence.Strand;
 import org.biojava.nbio.core.sequence.features.FeatureInterface;
 import org.biojava.nbio.core.sequence.features.Qualifier;
+import org.biojava.nbio.core.sequence.location.template.AbstractLocation;
 import org.biojava.nbio.core.sequence.location.template.Point;
 import org.biojava.nbio.core.sequence.template.AbstractSequence;
 import org.biojava.nbio.core.sequence.template.Compound;
@@ -220,7 +221,7 @@ private String _insdc_feature_location_string(FeatureInterface<AbstractSequence<
 
 	private String _insdc_location_string_ignoring_strand_and_subfeatures(
 			//SequenceLocation<AbstractSequence<C>, C> sequenceLocation,
-						org.biojava.nbio.core.sequence.location.template.AbstractLocation sequenceLocation,
+						AbstractLocation sequenceLocation,
 			int record_length) {
 	/*
 	if location.ref:
diff --git a/biojava-core/src/main/java/org/biojava/nbio/core/sequence/loader/UniprotProxySequenceReader.java b/biojava-core/src/main/java/org/biojava/nbio/core/sequence/loader/UniprotProxySequenceReader.java
@@ -337,10 +337,21 @@ public ArrayList<AccessionID> getAccessions() throws XPathExpressionException {
 
 	/**
 	 * Pull uniprot protein aliases associated with this sequence
+	 * Provided for backwards compatibility now that we support both
+	 * gene and protein aliases via separate methods.
 	 * @return
 	 * @throws XPathExpressionException
 	 */
 	public ArrayList<String> getAliases() throws XPathExpressionException {
+
+		return getProteinAliases();
+	}
+	/**
+	 * Pull uniprot protein aliases associated with this sequence
+	 * @return
+	 * @throws XPathExpressionException
+	 */
+	public ArrayList<String> getProteinAliases() throws XPathExpressionException {
 		ArrayList<String> aliasList = new ArrayList<String>();
 		if (uniprotDoc == null) {
 			return aliasList;
@@ -353,6 +364,32 @@ public ArrayList<String> getAliases() throws XPathExpressionException {
 			Element fullNameElement = XMLHelper.selectSingleElement(element, "fullName");
 			aliasList.add(fullNameElement.getTextContent());
 		}
+		keyWordElementList = XMLHelper.selectElements(proteinElement, "recommendedName");
+		for (Element element : keyWordElementList) {
+			Element fullNameElement = XMLHelper.selectSingleElement(element, "fullName");
+			aliasList.add(fullNameElement.getTextContent());
+		}
+
+		return aliasList;
+	}
+
+	/**
+	 * Pull uniprot gene aliases associated with this sequence
+	 * @return
+	 * @throws XPathExpressionException
+	 */
+	public ArrayList<String> getGeneAliases() throws XPathExpressionException {
+		ArrayList<String> aliasList = new ArrayList<String>();
+		if (uniprotDoc == null) {
+			return aliasList;
+		}
+		Element uniprotElement = uniprotDoc.getDocumentElement();
+		Element entryElement = XMLHelper.selectSingleElement(uniprotElement, "entry");
+		Element proteinElement = XMLHelper.selectSingleElement(entryElement, "gene");
+		ArrayList<Element> keyWordElementList = XMLHelper.selectElements(proteinElement, "name");
+		for (Element element : keyWordElementList) {
+			aliasList.add(element.getTextContent());
+		}
 
 		return aliasList;
 	}
diff --git a/biojava-structure/src/main/java/org/biojava/nbio/structure/io/PDBFileParser.java b/biojava-structure/src/main/java/org/biojava/nbio/structure/io/PDBFileParser.java
