Fixing issues with bonds in 3th3 and better logging

josemduarte · josemduarte · commit a50248e39110 · 2016-03-09T12:00:23.000-08:00
diff --git a/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/io/Test3th3.java b/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/io/Test3th3.java
@@ -0,0 +1,55 @@
+package org.biojava.nbio.structure.test.io;
+
+import static org.junit.Assert.*;
+
+import java.io.IOException;
+
+import org.biojava.nbio.structure.Chain;
+import org.biojava.nbio.structure.Group;
+import org.biojava.nbio.structure.ResidueNumber;
+import org.biojava.nbio.structure.Structure;
+import org.biojava.nbio.structure.StructureException;
+import org.biojava.nbio.structure.StructureIO;
+import org.biojava.nbio.structure.align.util.AtomCache;
+import org.biojava.nbio.structure.io.FileParsingParameters;
+import org.junit.Test;
+
+/**
+ * Test for a difficult for a file some sugar molecules covalently bound to residues
+ * having same residue number for sugar and residue (chain T, residue 201)
+ * @author Jose Duarte
+ *
+ */
+public class Test3th3 {
+
+	@Test 
+	public void test3th3() throws StructureException, IOException {
+		AtomCache cache = new AtomCache();
+		
+		FileParsingParameters params = cache.getFileParsingParams();
+		params.setUseInternalChainId(false);
+		params.setCreateAtomBonds(true);
+		StructureIO.setAtomCache(cache);
+		
+		Structure s = StructureIO.getStructure("3th3");
+		
+		Chain c = s.getChainByPDB("T");
+		
+		ResidueNumber rn = ResidueNumber.fromString("201");
+		rn.setChainId("T");
+		
+		Group g = c.getGroupByPDB(rn);
+				
+		assertEquals("LYS", g.getPDBName());
+		
+		int count = 0;
+		
+		for (Group gr : c.getAtomGroups()) {
+			if (gr.getResidueNumber().equals(rn)) count++;
+		}
+		
+		assertEquals(2, count);
+	}
+
+	
+}
diff --git a/biojava-structure/src/main/java/org/biojava/nbio/structure/ChainImpl.java b/biojava-structure/src/main/java/org/biojava/nbio/structure/ChainImpl.java
@@ -284,7 +284,11 @@ public void addGroup(Group group) {
 		 */
 
 			// this check is to give in this case the entry priority that is an AminoAcid / comes first...
+			// a good example of same residue number for 2 residues is 3th3, chain T, residue 201 (a LYS and a sugar BGC covalently attached to it) - JD 2016-03-09
 			if (  pdbResnumMap.containsKey(pdbResnum)) {
+				
+				logger.warn("Adding residue {}({}) to chain {} but a residue with same residue number is already present: {}({}). Will add only the aminoacid residue (if any) to the lookup, lookups for that residue number won't work properly.", 
+						pdbResnum, group.getPDBName(), getChainID(), groups.get(pdbResnumMap.get(pdbResnum)).getResidueNumber(), groups.get(pdbResnumMap.get(pdbResnum)).getPDBName()); 
 				if ( group instanceof AminoAcid)
 					pdbResnumMap.put(pdbResnum,pos);
 			} else
diff --git a/biojava-structure/src/main/java/org/biojava/nbio/structure/io/BondMaker.java b/biojava-structure/src/main/java/org/biojava/nbio/structure/io/BondMaker.java
@@ -347,32 +347,46 @@ public void formBondsFromStructConn(List<StructConn> structConn) {
 			String altLoc2 = "";
 			if (!conn.getPdbx_ptnr2_label_alt_id().equals("?")) altLoc2 = conn.getPdbx_ptnr2_label_alt_id();
 
+			// TODO: when issue 220 is implemented, add robust symmetry handling to allow bonds between symmetry-related molecules.
+			if (!conn.getPtnr1_symmetry().equals(symop) || !conn.getPtnr2_symmetry().equals(symop) ) {
+				logger.info("Skipping bond between atoms {}(residue {}{}) and {}(residue {}{}) belonging to different symmetry partners, because it is not supported yet", 
+						atomName1, seqId1, insCode1, atomName2, seqId2, insCode2);
+				continue;
+			}
+
+			
+			String altLocStr1 = altLoc1.isEmpty()? "" : "(alt loc "+altLoc1+")";
+			String altLocStr2 = altLoc2.isEmpty()? "" : "(alt loc "+altLoc2+")";
+			
 			Atom a1 = null;
 			Atom a2 = null;
 
 			try {
 				a1 = getAtomFromRecord(atomName1, altLoc1, resName1, chainId1, seqId1, insCode1);
 
 			} catch (StructureException e) {
-				String altLocStr1 = altLoc1.isEmpty()? "" : "(alt loc "+altLoc1+")";
+				
 				logger.warn("Could not find atom specified in struct_conn record: {}{}({}) in chain {}, atom {} {}", seqId1, insCode1, resName1, chainId1, atomName1, altLocStr1);
 				continue;
 			}
 			try {
 				a2 = getAtomFromRecord(atomName2, altLoc2, resName2, chainId2, seqId2, insCode2);
 			} catch (StructureException e) {
-				String altLocStr2 = altLoc2.isEmpty()? "" : "(alt loc "+altLoc2+")";
+				
 				logger.warn("Could not find atom specified in struct_conn record: {}{}({}) in chain {}, atom {} {}", seqId2, insCode2, resName2, chainId2, atomName2, altLocStr2);
 				continue;
 			}
 
-
-			// TODO: when issue 220 is implemented, add robust symmetry handling to allow bonds between symmetry-related molecules.
-			if (!conn.getPtnr1_symmetry().equals(symop) || !conn.getPtnr2_symmetry().equals(symop) ) {
-				logger.info("Skipping bond between atoms {}({}) and {}({}) belonging to different symmetry partners, because it is not supported yet", 
-						a1.getPDBserial(), a1.getName(), a2.getPDBserial(), a2.getName());
+			if (a1==null) {
+				// we couldn't find the atom, something must be wrong with the file 
+				logger.warn("Could not find atom {} {} from residue {}{}({}) in chain {} to create bond specified in struct_conn", atomName1, altLocStr1, seqId1, insCode1, resName1, chainId1);
 				continue;
 			}
+			if (a2==null) {
+				// we couldn't find the atom, something must be wrong with the file 
+				logger.warn("Could not find atom {} {} from residue {}{}({}) in chain {} to create bond specified in struct_conn", atomName2, altLocStr2, seqId2, insCode2, resName2, chainId2);
+				continue;				
+			}
 
 			// assuming order 1 for all bonds, no information is provided by struct_conn
 			Bond bond = new BondImpl(a1, a2, 1);
