-package jalview.analysis;\r
-\r
-import jalview.datamodel.*;\r
-import jalview.util.*;\r
-import jalview.schemes.ResidueProperties;\r
-import java.util.*;\r
-\r
-import jalview.io.NewickFile;\r
-\r
-public class NJTree {\r
-\r
- Vector cluster;\r
- SequenceI[] sequence;\r
-\r
- int done[];\r
- int noseqs;\r
- int noClus;\r
-\r
- float distance[][];\r
-\r
- int mini;\r
- int minj;\r
- float ri;\r
- float rj;\r
-\r
- Vector groups = new Vector();\r
- SequenceNode maxdist;\r
- SequenceNode top;\r
-\r
- float maxDistValue;\r
- float maxheight;\r
-\r
- int ycount;\r
-\r
- Vector node;\r
-\r
- String type;\r
- String pwtype;\r
-\r
- Object found = null;\r
- Object leaves = null;\r
-\r
- int start;\r
- int end;\r
-\r
- public NJTree(SequenceNode node) {\r
- top = node;\r
- maxheight = findHeight(top);\r
- }\r
-\r
- public String toString()\r
- {\r
- jalview.io.NewickFile fout = new jalview.io.NewickFile(getTopNode());\r
- return fout.print(false,true); // distances only\r
- }\r
-\r
- public NJTree(SequenceI[] seqs, NewickFile treefile) {\r
- top = treefile.getTree();\r
- maxheight = findHeight(top);\r
- SequenceIdMatcher algnIds = new SequenceIdMatcher(seqs);\r
-\r
- Vector leaves = new Vector();\r
- findLeaves(top, leaves);\r
-\r
- int i = 0;\r
- int namesleft = seqs.length;\r
-\r
- SequenceNode j;\r
- SequenceI nam;\r
- String realnam;\r
- while (i < leaves.size())\r
- {\r
- j = (SequenceNode) leaves.elementAt(i++);\r
- realnam = j.getName();\r
- nam = null;\r
- if (namesleft>-1)\r
- nam = algnIds.findIdMatch(realnam);\r
- if (nam != null) {\r
- j.setElement(nam);\r
- namesleft--;\r
- } else {\r
- j.setElement(new Sequence(realnam, "THISISAPLACEHLDER"));\r
- j.setPlaceholder(true);\r
-\r
- }\r
- }\r
- }\r
-\r
- public NJTree(SequenceI[] sequence,int start, int end) {\r
- this(sequence,"NJ","BL",start,end);\r
- }\r
-\r
- public NJTree(SequenceI[] sequence,String type,String pwtype,int start, int end ) {\r
-\r
- this.sequence = sequence;\r
- this.node = new Vector();\r
- this.type = type;\r
- this.pwtype = pwtype;\r
- this.start = start;\r
- this.end = end;\r
-\r
- if (!(type.equals("NJ"))) {\r
- type = "AV";\r
- }\r
-\r
- if (!(pwtype.equals("PID"))) {\r
- type = "BL";\r
- }\r
-\r
- int i=0;\r
-\r
- done = new int[sequence.length];\r
-\r
-\r
- while (i < sequence.length && sequence[i] != null) {\r
- done[i] = 0;\r
- i++;\r
- }\r
-\r
- noseqs = i++;\r
-\r
- distance = findDistances();\r
-\r
- makeLeaves();\r
-\r
- noClus = cluster.size();\r
-\r
- cluster();\r
-\r
- }\r
-\r
-\r
- public void cluster() {\r
-\r
- while (noClus > 2) {\r
- if (type.equals("NJ")) {\r
- float mind = findMinNJDistance();\r
- } else {\r
- float mind = findMinDistance();\r
- }\r
-\r
- Cluster c = joinClusters(mini,minj);\r
-\r
-\r
- done[minj] = 1;\r
-\r
- cluster.setElementAt(null,minj);\r
- cluster.setElementAt(c,mini);\r
-\r
- noClus--;\r
- }\r
-\r
- boolean onefound = false;\r
-\r
- int one = -1;\r
- int two = -1;\r
-\r
- for (int i=0; i < noseqs; i++) {\r
- if (done[i] != 1) {\r
- if (onefound == false) {\r
- two = i;\r
- onefound = true;\r
- } else {\r
- one = i;\r
- }\r
- }\r
- }\r
-\r
- Cluster c = joinClusters(one,two);\r
- top = (SequenceNode)(node.elementAt(one));\r
-\r
- reCount(top);\r
- findHeight(top);\r
- findMaxDist(top);\r
-\r
- }\r
-\r
- public Cluster joinClusters(int i, int j) {\r
-\r
- float dist = distance[i][j];\r
-\r
- int noi = ((Cluster)cluster.elementAt(i)).value.length;\r
- int noj = ((Cluster)cluster.elementAt(j)).value.length;\r
-\r
- int[] value = new int[noi + noj];\r
-\r
- for (int ii = 0; ii < noi;ii++) {\r
- value[ii] = ((Cluster)cluster.elementAt(i)).value[ii];\r
- }\r
-\r
- for (int ii = noi; ii < noi+ noj;ii++) {\r
- value[ii] = ((Cluster)cluster.elementAt(j)).value[ii-noi];\r
- }\r
-\r
- Cluster c = new Cluster(value);\r
-\r
- ri = findr(i,j);\r
- rj = findr(j,i);\r
-\r
- if (type.equals("NJ")) {\r
- findClusterNJDistance(i,j);\r
- } else {\r
- findClusterDistance(i,j);\r
- }\r
-\r
- SequenceNode sn = new SequenceNode();\r
-\r
- sn.setLeft((SequenceNode)(node.elementAt(i)));\r
- sn.setRight((SequenceNode)(node.elementAt(j)));\r
-\r
- SequenceNode tmpi = (SequenceNode)(node.elementAt(i));\r
- SequenceNode tmpj = (SequenceNode)(node.elementAt(j));\r
-\r
- if (type.equals("NJ")) {\r
- findNewNJDistances(tmpi,tmpj,dist);\r
- } else {\r
- findNewDistances(tmpi,tmpj,dist);\r
- }\r
-\r
- tmpi.setParent(sn);\r
- tmpj.setParent(sn);\r
-\r
- node.setElementAt(sn,i);\r
- return c;\r
- }\r
-\r
- public void findNewNJDistances(SequenceNode tmpi, SequenceNode tmpj, float dist) {\r
-\r
- float ih = 0;\r
- float jh = 0;\r
-\r
- SequenceNode sni = tmpi;\r
- SequenceNode snj = tmpj;\r
-\r
- tmpi.dist = (dist + ri - rj)/2;\r
- tmpj.dist = (dist - tmpi.dist);\r
-\r
- if (tmpi.dist < 0) {\r
- tmpi.dist = 0;\r
- }\r
- if (tmpj.dist < 0) {\r
- tmpj.dist = 0;\r
- }\r
- }\r
-\r
- public void findNewDistances(SequenceNode tmpi,SequenceNode tmpj,float dist) {\r
-\r
- float ih = 0;\r
- float jh = 0;\r
-\r
- SequenceNode sni = tmpi;\r
- SequenceNode snj = tmpj;\r
-\r
- while (sni != null) {\r
- ih = ih + sni.dist;\r
- sni = (SequenceNode)sni.left();\r
- }\r
-\r
- while (snj != null) {\r
- jh = jh + snj.dist;\r
- snj = (SequenceNode)snj.left();\r
- }\r
-\r
- tmpi.dist = (dist/2 - ih);\r
- tmpj.dist = (dist/2 - jh);\r
- }\r
-\r
-\r
-\r
- public void findClusterDistance(int i, int j) {\r
-\r
- int noi = ((Cluster)cluster.elementAt(i)).value.length;\r
- int noj = ((Cluster)cluster.elementAt(j)).value.length;\r
-\r
- // New distances from cluster to others\r
- float[] newdist = new float[noseqs];\r
-\r
- for (int l = 0; l < noseqs; l++) {\r
- if ( l != i && l != j) {\r
- newdist[l] = (distance[i][l] * noi + distance[j][l] * noj)/(noi + noj);\r
- } else {\r
- newdist[l] = 0;\r
- }\r
- }\r
-\r
- for (int ii=0; ii < noseqs;ii++) {\r
- distance[i][ii] = newdist[ii];\r
- distance[ii][i] = newdist[ii];\r
- }\r
- }\r
-\r
- public void findClusterNJDistance(int i, int j) {\r
-\r
- int noi = ((Cluster)cluster.elementAt(i)).value.length;\r
- int noj = ((Cluster)cluster.elementAt(j)).value.length;\r
-\r
- // New distances from cluster to others\r
- float[] newdist = new float[noseqs];\r
-\r
- for (int l = 0; l < noseqs; l++) {\r
- if ( l != i && l != j) {\r
- newdist[l] = (distance[i][l] + distance[j][l] - distance[i][j])/2;\r
- } else {\r
- newdist[l] = 0;\r
- }\r
- }\r
-\r
- for (int ii=0; ii < noseqs;ii++) {\r
- distance[i][ii] = newdist[ii];\r
- distance[ii][i] = newdist[ii];\r
- }\r
- }\r
-\r
- public float findr(int i, int j) {\r
-\r
- float tmp = 1;\r
- for (int k=0; k < noseqs;k++) {\r
- if (k!= i && k!= j && done[k] != 1) {\r
- tmp = tmp + distance[i][k];\r
- }\r
- }\r
-\r
- if (noClus > 2) {\r
- tmp = tmp/(noClus - 2);\r
- }\r
-\r
- return tmp;\r
- }\r
-\r
- public float findMinNJDistance() {\r
-\r
- float min = 100000;\r
-\r
- for (int i=0; i < noseqs-1; i++) {\r
- for (int j=i+1;j < noseqs;j++) {\r
- if (done[i] != 1 && done[j] != 1) {\r
- float tmp = distance[i][j] - (findr(i,j) + findr(j,i));\r
- if (tmp < min) {\r
-\r
- mini = i;\r
- minj = j;\r
-\r
- min = tmp;\r
-\r
- }\r
- }\r
- }\r
- }\r
- return min;\r
- }\r
-\r
- public float findMinDistance() {\r
-\r
- float min = 100000;\r
-\r
- for (int i=0; i < noseqs-1;i++) {\r
- for (int j = i+1; j < noseqs;j++) {\r
- if (done[i] != 1 && done[j] != 1) {\r
- if (distance[i][j] < min) {\r
- mini = i;\r
- minj = j;\r
-\r
- min = distance[i][j];\r
- }\r
- }\r
- }\r
- }\r
- return min;\r
- }\r
-\r
- public float[][] findDistances() {\r
-\r
- float[][] distance = new float[noseqs][noseqs];\r
- if (pwtype.equals("PID")) {\r
- for (int i = 0; i < noseqs-1; i++) {\r
- for (int j = i; j < noseqs; j++) {\r
- if (j==i) {\r
- distance[i][i] = 0;\r
- } else {\r
- distance[i][j] = 100-Comparison.PID(sequence[i], sequence[j]);\r
- distance[j][i] = distance[i][j];\r
- }\r
- }\r
- }\r
- } else if (pwtype.equals("BL")) {\r
- int maxscore = 0;\r
-\r
- for (int i = 0; i < noseqs-1; i++) {\r
- for (int j = i; j < noseqs; j++) {\r
- int score = 0;\r
- for (int k=0; k < sequence[i].getLength(); k++) {\r
- try{\r
- score +=\r
- ResidueProperties.getBLOSUM62(sequence[i].getSequence(k,\r
- k + 1),\r
- sequence[j].getSequence(k,\r
- k + 1));\r
- }catch(Exception ex){System.out.println("err creating BLOSUM62 tree");}\r
- }\r
- distance[i][j] = (float)score;\r
- if (score > maxscore) {\r
- maxscore = score;\r
- }\r
- }\r
- }\r
- for (int i = 0; i < noseqs-1; i++) {\r
- for (int j = i; j < noseqs; j++) {\r
- distance[i][j] = (float)maxscore - distance[i][j];\r
- distance[j][i] = distance[i][j];\r
- }\r
- }\r
- } else if (pwtype.equals("SW")) {\r
- float max = -1;\r
- for (int i = 0; i < noseqs-1; i++) {\r
- for (int j = i; j < noseqs; j++) {\r
- AlignSeq as = new AlignSeq(sequence[i],sequence[j],"pep");\r
- as.calcScoreMatrix();\r
- as.traceAlignment();\r
- as.printAlignment();\r
- distance[i][j] = (float)as.maxscore;\r
- if (max < distance[i][j]) {\r
- max = distance[i][j];\r
- }\r
- }\r
- }\r
- for (int i = 0; i < noseqs-1; i++) {\r
- for (int j = i; j < noseqs; j++) {\r
- distance[i][j] = max - distance[i][j];\r
- distance[j][i] = distance[i][j];\r
- }\r
- }\r
- }\r
-\r
- return distance;\r
- }\r
-\r
- public void makeLeaves() {\r
- cluster = new Vector();\r
-\r
- for (int i=0; i < noseqs; i++) {\r
- SequenceNode sn = new SequenceNode();\r
-\r
- sn.setElement(sequence[i]);\r
- sn.setName(sequence[i].getName());\r
- node.addElement(sn);\r
-\r
- int[] value = new int[1];\r
- value[0] = i;\r
-\r
- Cluster c = new Cluster(value);\r
- cluster.addElement(c);\r
- }\r
- }\r
-\r
- public Vector findLeaves(SequenceNode node, Vector leaves) {\r
- if (node == null) {\r
- return leaves;\r
- }\r
-\r
- if (node.left() == null && node.right() == null) {\r
- leaves.addElement(node);\r
- return leaves;\r
- } else {\r
- findLeaves((SequenceNode)node.left(),leaves);\r
- findLeaves((SequenceNode)node.right(),leaves);\r
- }\r
- return leaves;\r
- }\r
-\r
- public Object findLeaf(SequenceNode node, int count) {\r
- found = _findLeaf(node,count);\r
-\r
- return found;\r
- }\r
- public Object _findLeaf(SequenceNode node,int count) {\r
- if (node == null) {\r
- return null;\r
- }\r
- if (node.ycount == count) {\r
- found = node.element();\r
- return found;\r
- } else {\r
- _findLeaf((SequenceNode)node.left(),count);\r
- _findLeaf((SequenceNode)node.right(),count);\r
- }\r
-\r
- return found;\r
- }\r
-\r
- public void printNode(SequenceNode node) {\r
- if (node == null) {\r
- return;\r
- }\r
- if (node.left() == null && node.right() == null) {\r
- System.out.println("Leaf = " + ((SequenceI)node.element()).getName());\r
- System.out.println("Dist " + ((SequenceNode)node).dist);\r
- System.out.println("Boot " + node.getBootstrap());\r
- } else {\r
- System.out.println("Dist " + ((SequenceNode)node).dist);\r
- printNode((SequenceNode)node.left());\r
- printNode((SequenceNode)node.right());\r
- }\r
- }\r
- public void findMaxDist(SequenceNode node) {\r
- if (node == null) {\r
- return;\r
- }\r
- if (node.left() == null && node.right() == null) {\r
-\r
- float dist = ((SequenceNode)node).dist;\r
- if (dist > maxDistValue) {\r
- maxdist = (SequenceNode)node;\r
- maxDistValue = dist;\r
- }\r
- } else {\r
- findMaxDist((SequenceNode)node.left());\r
- findMaxDist((SequenceNode)node.right());\r
- }\r
- }\r
- public Vector getGroups() {\r
- return groups;\r
- }\r
- public float getMaxHeight() {\r
- return maxheight;\r
- }\r
- public void groupNodes(SequenceNode node, float threshold) {\r
- if (node == null) {\r
- return;\r
- }\r
-\r
- if (node.height/maxheight > threshold) {\r
- groups.addElement(node);\r
- } else {\r
- groupNodes((SequenceNode)node.left(),threshold);\r
- groupNodes((SequenceNode)node.right(),threshold);\r
- }\r
- }\r
-\r
- public float findHeight(SequenceNode node) {\r
-\r
- if (node == null) {\r
- return maxheight;\r
- }\r
-\r
- if (node.left() == null && node.right() == null) {\r
- node.height = ((SequenceNode)node.parent()).height + node.dist;\r
-\r
- if (node.height > maxheight) {\r
- return node.height;\r
- } else {\r
- return maxheight;\r
- }\r
- } else {\r
- if (node.parent() != null) {\r
- node.height = ((SequenceNode)node.parent()).height + node.dist;\r
- } else {\r
- maxheight = 0;\r
- node.height = (float)0.0;\r
- }\r
-\r
- maxheight = findHeight((SequenceNode)(node.left()));\r
- maxheight = findHeight((SequenceNode)(node.right()));\r
- }\r
- return maxheight;\r
- }\r
- public SequenceNode reRoot() {\r
- if (maxdist != null) {\r
- ycount = 0;\r
- float tmpdist = maxdist.dist;\r
-\r
- // New top\r
- SequenceNode sn = new SequenceNode();\r
- sn.setParent(null);\r
-\r
- // New right hand of top\r
- SequenceNode snr = (SequenceNode)maxdist.parent();\r
- changeDirection(snr,maxdist);\r
- System.out.println("Printing reversed tree");\r
- printN(snr);\r
- snr.dist = tmpdist/2;\r
- maxdist.dist = tmpdist/2;\r
-\r
- snr.setParent(sn);\r
- maxdist.setParent(sn);\r
-\r
- sn.setRight(snr);\r
- sn.setLeft(maxdist);\r
-\r
- top = sn;\r
-\r
- ycount = 0;\r
- reCount(top);\r
- findHeight(top);\r
-\r
- }\r
- return top;\r
- }\r
- public static void printN(SequenceNode node) {\r
- if (node == null) {\r
- return;\r
- }\r
-\r
- if (node.left() != null && node.right() != null) {\r
- printN((SequenceNode)node.left());\r
- printN((SequenceNode)node.right());\r
- } else {\r
- System.out.println(" name = " + ((SequenceI)node.element()).getName());\r
- }\r
- System.out.println(" dist = " + ((SequenceNode)node).dist + " " + ((SequenceNode)node).count + " " + ((SequenceNode)node).height);\r
- }\r
-\r
- public void reCount(SequenceNode node) {\r
- ycount = 0;\r
- _reCount(node);\r
- }\r
- public void _reCount(SequenceNode node) {\r
- if (node == null) {\r
- return;\r
- }\r
-\r
- if (node.left() != null && node.right() != null) {\r
- _reCount((SequenceNode)node.left());\r
- _reCount((SequenceNode)node.right());\r
-\r
- SequenceNode l = (SequenceNode)node.left();\r
- SequenceNode r = (SequenceNode)node.right();\r
-\r
- ((SequenceNode)node).count = l.count + r.count;\r
- ((SequenceNode)node).ycount = (l.ycount + r.ycount)/2;\r
-\r
- } else {\r
- ((SequenceNode)node).count = 1;\r
- ((SequenceNode)node).ycount = ycount++;\r
- }\r
-\r
- }\r
- public void swapNodes(SequenceNode node) {\r
- if (node == null) {\r
- return;\r
- }\r
- SequenceNode tmp = (SequenceNode)node.left();\r
-\r
- node.setLeft(node.right());\r
- node.setRight(tmp);\r
- }\r
- public void changeDirection(SequenceNode node, SequenceNode dir) {\r
- if (node == null) {\r
- return;\r
- }\r
- if (node.parent() != top) {\r
- changeDirection((SequenceNode)node.parent(), node);\r
-\r
- SequenceNode tmp = (SequenceNode)node.parent();\r
-\r
- if (dir == node.left()) {\r
- node.setParent(dir);\r
- node.setLeft(tmp);\r
- } else if (dir == node.right()) {\r
- node.setParent(dir);\r
- node.setRight(tmp);\r
- }\r
-\r
- } else {\r
- if (dir == node.left()) {\r
- node.setParent(node.left());\r
-\r
- if (top.left() == node) {\r
- node.setRight(top.right());\r
- } else {\r
- node.setRight(top.left());\r
- }\r
- } else {\r
- node.setParent(node.right());\r
-\r
- if (top.left() == node) {\r
- node.setLeft(top.right());\r
- } else {\r
- node.setLeft(top.left());\r
- }\r
- }\r
- }\r
- }\r
- public void setMaxDist(SequenceNode node) {\r
- this.maxdist = maxdist;\r
- }\r
- public SequenceNode getMaxDist() {\r
- return maxdist;\r
- }\r
- public SequenceNode getTopNode() {\r
- return top;\r
- }\r
-\r
-}\r
-\r
-\r
-\r
-class Cluster {\r
-\r
- int[] value;\r
-\r
- public Cluster(int[] value) {\r
- this.value = value;\r
- }\r
-\r
-}\r
-\r
+/*
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ *
+ * This file is part of Jalview.
+ *
+ * Jalview is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *
+ * Jalview is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ * PURPOSE. See the GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
+ */
+package jalview.analysis;
+
+import jalview.api.analysis.ScoreModelI;
+import jalview.api.analysis.SimilarityParamsI;
+import jalview.datamodel.SequenceNode;
+import jalview.viewmodel.AlignmentViewport;
+
+/**
+ * This class implements distance calculations used in constructing a Neighbour
+ * Joining tree
+ */
+public class NJTree extends TreeBuilder
+{
+ /**
+ * Constructor given a viewport, tree type and score model
+ *
+ * @param av
+ * the current alignment viewport
+ * @param sm
+ * a distance or similarity score model to use to compute the tree
+ * @param scoreParameters
+ */
+ public NJTree(AlignmentViewport av, ScoreModelI sm,
+ SimilarityParamsI scoreParameters)
+ {
+ super(av, sm, scoreParameters);
+ }
+
+ /**
+ * {@inheritDoc}
+ */
+ @Override
+ protected double findMinDistance()
+ {
+ double min = Double.MAX_VALUE;
+
+ for (int i = 0; i < (noseqs - 1); i++)
+ {
+ for (int j = i + 1; j < noseqs; j++)
+ {
+ if (!done.get(i) && !done.get(j))
+ {
+ double tmp = distances.getValue(i, j)
+ - (findr(i, j) + findr(j, i));
+
+ if (tmp < min)
+ {
+ mini = i;
+ minj = j;
+
+ min = tmp;
+ }
+ }
+ }
+ }
+
+ return min;
+ }
+
+ /**
+ * {@inheritDoc}
+ */
+ @Override
+ protected void findNewDistances(SequenceNode nodei, SequenceNode nodej,
+ double dist)
+ {
+ nodei.dist = ((dist + ri) - rj) / 2;
+ nodej.dist = (dist - nodei.dist);
+
+ if (nodei.dist < 0)
+ {
+ nodei.dist = 0;
+ }
+
+ if (nodej.dist < 0)
+ {
+ nodej.dist = 0;
+ }
+ }
+
+ /**
+ * Calculates and saves the distance between the combination of cluster(i) and
+ * cluster(j) and all other clusters. The new distance to cluster k is
+ * calculated as the average of the distances from i to k and from j to k,
+ * less half the distance from i to j.
+ *
+ * @param i
+ * @param j
+ */
+ @Override
+ protected void findClusterDistance(int i, int j)
+ {
+ // New distances from cluster i to others
+ double[] newdist = new double[noseqs];
+
+ double ijDistance = distances.getValue(i, j);
+ for (int l = 0; l < noseqs; l++)
+ {
+ if ((l != i) && (l != j))
+ {
+ newdist[l] = (distances.getValue(i, l) + distances.getValue(j, l)
+ - ijDistance) / 2;
+ }
+ else
+ {
+ newdist[l] = 0;
+ }
+ }
+
+ for (int ii = 0; ii < noseqs; ii++)
+ {
+ distances.setValue(i, ii, newdist[ii]);
+ distances.setValue(ii, i, newdist[ii]);
+ }
+ }
+}
git://source.jalview.org / jalview.git / blobdiff