Fix bug in distance matrix calculation

lafita · lafita · commit 9cabf4101873 · 2015-11-17T15:21:11.000+01:00
diff --git a/biojava-phylo/src/main/java/demo/DemoDistanceTree.java b/biojava-phylo/src/main/java/demo/DemoDistanceTree.java
@@ -23,7 +23,7 @@
 
 /**
  * This demo contains the CookBook example to create a phylogenetic tree from a
- * given multiple sequence alignment (MSA).
+ * multiple sequence alignment (MSA).
  * 
  * @author Scooter Willis
  * @author Aleix Lafita
diff --git a/biojava-phylo/src/main/java/org/biojava/nbio/phylo/DistanceMatrixCalculator.java b/biojava-phylo/src/main/java/org/biojava/nbio/phylo/DistanceMatrixCalculator.java
@@ -263,7 +263,7 @@ public static <C extends Sequence<D>, D extends Compound> DistanceMatrix fractio
 	 * itself). Calculation of the score is as follows:
 	 * 
 	 * <pre>
-	 * Ds = maxScore - sum(M<sub>ai,bi</sub>)
+	 * Ds = maxScore - sum<sub>i</sub>(M<sub>ai,bi</sub>)
 	 * </pre>
 	 * 
 	 * It is recommended to use the method
@@ -307,17 +307,20 @@ public static <C extends Sequence<D>, D extends Compound> DistanceMatrix dissimi
 
 			// Obtain the similarity scores
 			for (int j = i; j < n; j++) {
+
 				double score = 0;
 				loopcount++;
+
 				for (int k = 0; k < end; k++) {
 					if (Comparison.isGap(sequenceString[i].charAt(k))
 							|| Comparison.isGap(sequenceString[j].charAt(k)))
 						continue;
-					// TODO bug
-					score += M.getValue(seqs.get(i).getCompoundAt(k),
-							seqs.get(j).getCompoundAt(k));
+					score += M.getValue(seqs.get(i).getCompoundAt(k + 1),
+							seqs.get(j).getCompoundAt(k + 1));
 				}
-				DM.setValue(i, j, score);
+
+				if (i != j)
+					DM.setValue(i, j, score);
 
 				if (score > maxscore) {
 					maxscore = score;
diff --git a/biojava-phylo/src/main/java/org/biojava/nbio/phylo/DistanceTreeEvaluator.java b/biojava-phylo/src/main/java/org/biojava/nbio/phylo/DistanceTreeEvaluator.java
@@ -116,7 +116,7 @@ public static double evaluate(Phylogeny tree, DistanceMatrix matrix) {
 		logger.info("Average tree distance: {}", averageTreeDistance);
 		logger.info("Average LS error: {}", averageTreeErrorDistance);
 
-		return Math.sqrt((averageTreeErrorDistance)) / (averageMatrixDistance);
+		return Math.sqrt(averageTreeErrorDistance) / averageMatrixDistance;
 	}
 
 	private static double getNodeDistance(PhylogenyNode parentNode,

Original file line number	Diff line number	Diff line change
`@@ -23,7 +23,7 @@`
`23`	`23`
`24`	`24`	`/**`
`25`	`25`	`* This demo contains the CookBook example to create a phylogenetic tree from a`
`26`		`- * given multiple sequence alignment (MSA).`
	`26`	`+ * multiple sequence alignment (MSA).`
`27`	`27`	`*`
`28`	`28`	`* @author Scooter Willis`
`29`	`29`	`* @author Aleix Lafita`
Original file line number	Diff line number	Diff line change
`@@ -116,7 +116,7 @@ public static double evaluate(Phylogeny tree, DistanceMatrix matrix) {`
`116`	`116`	`logger.info("Average tree distance: {}", averageTreeDistance);`
`117`	`117`	`logger.info("Average LS error: {}", averageTreeErrorDistance);`
`118`	`118`
`119`		`- return Math.sqrt((averageTreeErrorDistance)) / (averageMatrixDistance);`
	`119`	`+ return Math.sqrt(averageTreeErrorDistance) / averageMatrixDistance;`
`120`	`120`	`}`
`121`	`121`
`122`	`122`	`private static double getNodeDistance(PhylogenyNode parentNode,`