Fix for #929

josemduarte · josemduarte · commit f4b201bf3c0f · 2021-09-22T13:56:05.000-07:00
diff --git a/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/io/TestBondParsing.java b/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/io/TestBondParsing.java
@@ -31,4 +31,27 @@ public void testIssue943() throws Exception {
         assertEquals(3, zn.getAtom("ZN").getBonds().size());
 
     }
+
+    /**
+     * Integration test for SS bond parsing in PDB-format, where author chain ids and asym ids differ and can cause
+     * problems. See https://github.com/biojava/biojava/issues/929
+     */
+    @Test
+    public void testIssue929() throws Exception {
+        PDBFileReader reader = new PDBFileReader();
+        FileParsingParameters params = new FileParsingParameters();
+        params.setCreateAtomBonds(true);
+        reader.setFileParsingParameters(params);
+        Structure s = reader.getStructureById("1a4w");
+
+        Group cysB = s.getPolyChain("B").getAtomGroup(118);
+        Atom sgCysB = cysB.getAtom("SG");
+        assertEquals(2, sgCysB.getBonds().size());
+
+        Group cysA = s.getPolyChain("A").getAtomGroup(1);
+        Atom sgCysA = cysA.getAtom("SG");
+        assertEquals(2, sgCysA.getBonds().size());
+
+
+    }
 }
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
@@ -79,8 +79,8 @@ public class BondMaker {
 	 */
 	private static final double MAX_NUCLEOTIDE_BOND_LENGTH = 2.1;
 
-	private Structure structure;
-	private FileParsingParameters params;
+	private final Structure structure;
+	private final FileParsingParameters params;
 
 	public BondMaker(Structure structure, FileParsingParameters params) {
 		this.structure = structure;
@@ -282,7 +282,7 @@ private void trimBondLists() {
 
 	/**
 	 * Creates disulfide bond objects and references in the corresponding Atoms objects, given
-	 * a list of {@link SSBondImpl}s parsed from a PDB/mmCIF file.
+	 * a list of {@link SSBondImpl}s parsed from a PDB file.
 	 * @param disulfideBonds
 	 */
 	public void formDisulfideBonds(List<SSBondImpl> disulfideBonds) {
@@ -293,12 +293,24 @@ public void formDisulfideBonds(List<SSBondImpl> disulfideBonds) {
 
 	private void formDisulfideBond(SSBondImpl disulfideBond) {
 		try {
-			Map<Integer, Atom> a = getAtomFromRecord("SG", "",
-					disulfideBond.getChainID1(), disulfideBond.getResnum1(),
-					disulfideBond.getInsCode1());
-			Map<Integer, Atom> b = getAtomFromRecord("SG", "",
-					disulfideBond.getChainID2(), disulfideBond.getResnum2(),
-					disulfideBond.getInsCode2());
+			// The PDB format uses author chain ids to reference chains. But one author chain id corresponds to multiple asym ids,
+			// thus we need to grab all the possible asym ids (poly and nonpoly) and then try to find the atoms
+			// See issue https://github.com/biojava/biojava/issues/943
+			String polyChainId1 = structure.getPolyChainByPDB(disulfideBond.getChainID1()).getId();
+			String polyChainId2 = structure.getPolyChainByPDB(disulfideBond.getChainID2()).getId();
+			List<Chain> nonpolyChains1 = structure.getNonPolyChainsByPDB(disulfideBond.getChainID1());
+			List<Chain> nonpolyChains2 = structure.getNonPolyChainsByPDB(disulfideBond.getChainID2());
+
+			List<String> allChainIds1 = new ArrayList<>();
+			List<String> allChainIds2 = new ArrayList<>();
+			if (polyChainId1!=null) allChainIds1.add(polyChainId1);
+			if (polyChainId2!=null) allChainIds2.add(polyChainId2);
+			if (nonpolyChains1!=null) nonpolyChains1.forEach(npc -> allChainIds1.add(npc.getId()));
+			if (nonpolyChains2!=null) nonpolyChains2.forEach(npc -> allChainIds2.add(npc.getId()));
+
+			Map<Integer, Atom> a = getAtomFromRecordTryMultipleChainIds("SG", "", disulfideBond.getResnum1(), disulfideBond.getInsCode1(), allChainIds1);
+
+			Map<Integer, Atom> b = getAtomFromRecordTryMultipleChainIds("SG", "", disulfideBond.getResnum2(), disulfideBond.getInsCode2(), allChainIds2);
 
 			for(int i=0; i<structure.nrModels(); i++){
 				if(a.containsKey(i) && b.containsKey(i)){
