Create a test for the new parameters of the SmithWaterman3Daligner biojava#552

lafita · lafita · commit 7be152f6d014 · 2017-03-21T13:00:12.000+01:00
diff --git a/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/align/AlignmentToolsTest.java b/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/align/AlignmentToolsTest.java
@@ -443,4 +443,5 @@ public boolean containsKey(Object key) {
 			return true;
 		}
 	}
+	
 }
diff --git a/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/align/seq/TestSmithWaterman3Daligner.java b/biojava-integrationtest/src/test/java/org/biojava/nbio/structure/test/align/seq/TestSmithWaterman3Daligner.java
@@ -0,0 +1,67 @@
+package org.biojava.nbio.structure.test.align.seq;
+
+import static org.junit.Assert.*;
+
+import java.io.IOException;
+
+import org.biojava.nbio.structure.Atom;
+import org.biojava.nbio.structure.Structure;
+import org.biojava.nbio.structure.StructureException;
+import org.biojava.nbio.structure.StructureTools;
+import org.biojava.nbio.structure.align.model.AFPChain;
+import org.biojava.nbio.structure.align.seq.SmithWaterman3DParameters;
+import org.biojava.nbio.structure.align.seq.SmithWaterman3Daligner;
+import org.junit.Test;
+
+/**
+ * Test the superposition based on a sequence alignment on different cases.
+ * 
+ * @author Aleix Lafita
+ *
+ */
+public class TestSmithWaterman3Daligner {
+
+	/**
+	 * Test the changes in the alignment by the max RMSD parameter. Use the
+	 * chain A and B of hemoglobin, because they have very similar sequences,
+	 * but some columns of the alignment have to be removed for a better
+	 * superposition (lower RMSD).
+	 */
+	@Test
+	public void testMaxRMSD() throws StructureException, IOException {
+
+		Structure s1 = StructureTools.getStructure("1A3N.A");
+		Structure s2 = StructureTools.getStructure("1A3N.B");
+
+		Atom[] ca1 = StructureTools.getRepresentativeAtomArray(s1);
+		Atom[] ca2 = StructureTools.getRepresentativeAtomArray(s2);
+
+		SmithWaterman3Daligner aligner = new SmithWaterman3Daligner();
+		SmithWaterman3DParameters params = new SmithWaterman3DParameters();
+
+		// Use no restriction on the RMSD
+		params.setMaxRmsd(99.0);
+
+		AFPChain afpChain = aligner.align(ca1, ca2, params);
+
+		assertEquals("RMSD is wrong", 1.39, afpChain.getTotalRmsdOpt(), 0.005);
+		assertEquals("Length is wrong", 137, afpChain.getOptLength());
+		
+		// Restrict it to 1A RMSD (18 columns have to be dropped)
+		params.setMaxRmsd(1.0);
+
+		afpChain = aligner.align(ca1, ca2, params);
+
+		assertTrue("RMSD is above the threshold", afpChain.getTotalRmsdOpt() < 1.0);
+		assertEquals("Length is wrong", 119, afpChain.getOptLength());
+		
+		// Restrict it to 0A RMSD (the minimum length is relevant )
+		params.setMaxRmsd(0.0);
+		params.setMinLen(30);
+
+		afpChain = aligner.align(ca1, ca2, params);
+
+		assertEquals("Length shoild be the minimum possible", 30, afpChain.getOptLength());
+
+	}
+}
diff --git a/biojava-structure/src/main/java/org/biojava/nbio/structure/StructureTools.java b/biojava-structure/src/main/java/org/biojava/nbio/structure/StructureTools.java
@@ -45,7 +45,6 @@
 import org.biojava.nbio.structure.contact.Grid;
 import org.biojava.nbio.structure.io.FileParsingParameters;
 import org.biojava.nbio.structure.io.PDBFileParser;
-import org.biojava.nbio.structure.io.mmcif.model.ChemComp;
 import org.biojava.nbio.structure.io.util.FileDownloadUtils;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
diff --git a/biojava-structure/src/main/java/org/biojava/nbio/structure/align/seq/SmithWaterman3Daligner.java b/biojava-structure/src/main/java/org/biojava/nbio/structure/align/seq/SmithWaterman3Daligner.java
@@ -71,8 +71,7 @@ public class SmithWaterman3Daligner extends AbstractStructureAlignment implement
 
 	/**
 	 *  version history:
-	 *  1.1 - Added more parameters to the command line, including maximum RMSD 
-	 *  		and minimum length
+	 *  1.1 - Added a maxRMSD and minLen parameters
 	 *  1.0 - Initial version
 	 */
 	private static final String version = "1.1";
@@ -147,7 +146,7 @@ public AFPChain align(Atom[] ca1, Atom[] ca2, Object parameters)
 		afpChain = convert(ca1,ca2,pair, smithWaterman);
 		
 		// Perform an iterative dropping of the columns
-		while (afpChain.getAlnLength() > params.getMinLen()
+		while (afpChain.getOptLength() > params.getMinLen()
 				&& afpChain.getTotalRmsdOpt() > params.getMaxRmsd()) {
 			afpChain = AlignmentTools.deleteHighestDistanceColumn(afpChain, ca1, ca2);
 		}
diff --git a/biojava-structure/src/main/java/org/biojava/nbio/structure/align/util/AlignmentTools.java b/biojava-structure/src/main/java/org/biojava/nbio/structure/align/util/AlignmentTools.java
@@ -1425,11 +1425,16 @@ public static AFPChain deleteHighestDistanceColumn(AFPChain afpChain,
 
 		for (int b = 0; b < optAln.length; b++) {
 			for (int p = 0; p < optAln[b][0].length; p++) {
+				Atom ca2clone = ca2[optAln[b][1][p]];
+				Calc.rotate(ca2clone, afpChain.getBlockRotationMatrix()[b]);
+				Calc.shift(ca2clone, afpChain.getBlockShiftVector()[b]);
+				
 				double distance = Calc.getDistance(ca1[optAln[b][0][p]],
-						ca2[optAln[b][1][p]]);
+						ca2clone);
 				if (distance > maxDistance) {
 					maxBlock = b;
 					maxPos = p;
+					maxDistance = distance;
 				}
 			}
 		}
@@ -1479,8 +1484,8 @@ public static AFPChain deleteColumn(AFPChain afpChain, Atom[] ca1,
 			if (p == pos)
 				continue;
 
-			newPos0[position] = optAln[block][0][position];
-			newPos1[position] = optAln[block][1][position];
+			newPos0[position] = optAln[block][0][p];
+			newPos1[position] = optAln[block][1][p];
 
 			position++;
 		}

Original file line number	Diff line number	Diff line change
`@@ -443,4 +443,5 @@ public boolean containsKey(Object key) {`
`443`	`443`	`return true;`
`444`	`444`	`}`
`445`	`445`	`}`
	`446`	`+`
`446`	`447`	`}`