2 // FORESTER -- software libraries and applications
3 // for evolutionary biology research and applications.
5 // Copyright (C) 2014 Christian M. Zmasek
8 // This library is free software; you can redistribute it and/or
9 // modify it under the terms of the GNU Lesser General Public
10 // License as published by the Free Software Foundation; either
11 // version 2.1 of the License, or (at your option) any later version.
13 // This library is distributed in the hope that it will be useful,
14 // but WITHOUT ANY WARRANTY; without even the implied warranty of
15 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 // Lesser General Public License for more details.
18 // You should have received a copy of the GNU Lesser General Public
19 // License along with this library; if not, write to the Free Software
20 // Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
22 // Contact: phylosoft @ gmail . com
23 // WWW: https://sites.google.com/site/cmzmasek/home/software/forester
25 package org.forester.evoinference.distance;
27 import java.math.RoundingMode;
28 import java.text.DecimalFormat;
29 import java.util.ArrayList;
30 import java.util.List;
31 import java.util.Map.Entry;
32 import java.util.SortedSet;
34 import org.forester.evoinference.matrix.distance.BasicSymmetricalDistanceMatrix;
35 import org.forester.phylogeny.Phylogeny;
36 import org.forester.phylogeny.PhylogenyNode;
37 import org.forester.util.ForesterUtil;
39 public final class NeighborJoiningR {
41 private final static DecimalFormat DF = new DecimalFormat( "0.00" );
42 private BasicSymmetricalDistanceMatrix _d;
43 private double[][] _d_values;
44 private final DecimalFormat _df;
45 private PhylogenyNode[] _external_nodes;
46 private int[] _mappings;
49 private final boolean _verbose;
53 private double _d_min; //TODO remove me
55 private NeighborJoiningR() {
60 private NeighborJoiningR( final boolean verbose, final int maximum_fraction_digits_for_distances ) {
61 if ( ( maximum_fraction_digits_for_distances < 1 ) || ( maximum_fraction_digits_for_distances > 9 ) ) {
62 throw new IllegalArgumentException( "maximum fraction digits for distances is out of range: "
63 + maximum_fraction_digits_for_distances );
66 _df = new DecimalFormat();
67 _df.setMaximumFractionDigits( maximum_fraction_digits_for_distances );
68 _df.setRoundingMode( RoundingMode.HALF_UP );
71 public final Phylogeny execute( final BasicSymmetricalDistanceMatrix distance ) {
73 final Phylogeny phylogeny = new Phylogeny();
75 System.out.println( "N=" + _n );
77 // Calculates the minimal distance.
78 // If more than one minimal distances, always the first found is used
79 final double m = updateM();
80 final int otu1 = _min_i;
81 final int otu2 = _min_j;
82 System.out.println( _min_i + " " + _min_j + " => " + DF.format( m ) + " (" + DF.format( _d_min ) + ")" );
83 // It is a condition that otu1 < otu2.
84 final PhylogenyNode node = new PhylogenyNode();
85 final double d = getDvalue( otu1, otu2 );
86 final double d1 = ( d / 2 ) + ( ( _r[ otu1 ] - _r[ otu2 ] ) / ( 2 * ( _n - 2 ) ) );
87 final double d2 = d - d1;
89 getExternalPhylogenyNode( otu1 ).setDistanceToParent( d1 );
90 getExternalPhylogenyNode( otu2 ).setDistanceToParent( d2 );
93 // yes, yes, slow but only grows with n (and not n^2 or worse)...
94 getExternalPhylogenyNode( otu1 ).setDistanceToParent( Double.parseDouble( _df.format( d1 ) ) );
95 getExternalPhylogenyNode( otu2 ).setDistanceToParent( Double.parseDouble( _df.format( d2 ) ) );
97 node.addAsChild( getExternalPhylogenyNode( otu1 ) );
98 node.addAsChild( getExternalPhylogenyNode( otu2 ) );
100 printProgress( otu1, otu2 );
102 calculateDistancesFromNewNode( otu1, otu2, d );
103 _external_nodes[ _mappings[ otu1 ] ] = node;
104 updateMappings( otu2 );
106 System.out.println( "-------------------------------------------------------------" );
107 System.out.println( "" );
109 final double d = getDvalue( 0, 1 ) / 2;
111 getExternalPhylogenyNode( 0 ).setDistanceToParent( d );
112 getExternalPhylogenyNode( 1 ).setDistanceToParent( d );
115 final double dd = Double.parseDouble( _df.format( d ) );
116 getExternalPhylogenyNode( 0 ).setDistanceToParent( dd );
117 getExternalPhylogenyNode( 1 ).setDistanceToParent( dd );
119 final PhylogenyNode root = new PhylogenyNode();
120 root.addAsChild( getExternalPhylogenyNode( 0 ) );
121 root.addAsChild( getExternalPhylogenyNode( 1 ) );
123 printProgress( 0, 1 );
125 phylogeny.setRoot( root );
126 phylogeny.setRooted( false );
130 public final List<Phylogeny> execute( final List<BasicSymmetricalDistanceMatrix> distances_list ) {
131 final List<Phylogeny> pl = new ArrayList<Phylogeny>();
132 for( final BasicSymmetricalDistanceMatrix distances : distances_list ) {
133 pl.add( execute( distances ) );
138 private final void calculateDistancesFromNewNode( final int otu1, final int otu2, final double d ) {
139 for( int i = 0; i < _n; ++i ) {
140 if ( ( i == otu1 ) || ( i == otu2 ) ) {
143 updateDvalue( otu1, otu2, i, d );
147 private final void updateDvalue( final int otu1, final int otu2, final int i, final double d ) {
148 setDvalue( otu1, i, ( getDvalue( otu1, i ) + getDvalue( i, otu2 ) - d ) / 2 );
151 private void setDvalue( final int i, final int j, final double d ) {
153 _d_values[ _mappings[ i ] ][ _mappings[ j ] ] = d;
155 _d_values[ _mappings[ j ] ][ _mappings[ i ] ] = d;
158 private double getDvalue( final int i, final int j ) {
160 return _d_values[ _mappings[ i ] ][ _mappings[ j ] ];
162 return _d_values[ _mappings[ j ] ][ _mappings[ i ] ];
165 private double getDvalueUnmapped( final int i, final int j ) {
167 return _d_values[ i ][ j ];
169 return _d_values[ j ][ i ];
172 private final void calculateNetDivergences() {
173 for( int i = 0; i < _n; ++i ) {
174 _r[ i ] = calculateNetDivergence( i );
178 private double calculateNetDivergence( final int i ) {
180 for( int n = 0; n < _n; ++n ) {
182 d += getDvalue( n, i );
188 private final PhylogenyNode getExternalPhylogenyNode( final int i ) {
189 return _external_nodes[ _mappings[ i ] ];
192 private final void initExternalNodes() {
193 _external_nodes = new PhylogenyNode[ _n ];
195 for( int i = 0; i < _n; ++i ) {
196 _external_nodes[ i ] = new PhylogenyNode();
197 id = _d.getIdentifier( i );
199 _external_nodes[ i ].setName( id );
202 _external_nodes[ i ].setName( Integer.toString( i ) );
208 private final void printProgress( final int otu1, final int otu2 ) {
209 System.out.println( "Node " + printProgressNodeToString( getExternalPhylogenyNode( otu1 ) ) + " joins "
210 + ( printProgressNodeToString( getExternalPhylogenyNode( otu2 ) ) ) );
213 private final String printProgressNodeToString( final PhylogenyNode n ) {
214 if ( n.isExternal() ) {
215 if ( ForesterUtil.isEmpty( n.getName() ) ) {
216 return Long.toString( n.getId() );
222 + ( ForesterUtil.isEmpty( n.getChildNode1().getName() ) ? n.getChildNode1().getId() : n.getChildNode1()
225 + ( ForesterUtil.isEmpty( n.getChildNode2().getName() ) ? n.getChildNode2().getId() : n.getChildNode2()
229 // only the values in the lower triangle are used.
230 // !matrix values will be changed!
231 private final void reset( final BasicSymmetricalDistanceMatrix distances ) {
232 _n = distances.getSize();
234 _r = new double[ _n ];
235 _mappings = new int[ _n ];
236 _d_values = _d.getValues();
238 _s.initialize( distances );
243 final private void printM() {
244 for( int j = 1; j < _n; ++j ) {
245 for( int i = 0; i < _n; ++i ) {
246 System.out.print( DF.format( _d_values[ _mappings[ i ] ][ _mappings[ j ] ] ) );
247 System.out.print( " " );
249 System.out.print( " " );
250 for( final Entry<Double, SortedSet<Integer>> entry : _s.getSentrySet( _mappings[ j ] ) ) {
251 final double key = entry.getKey();
252 final SortedSet<Integer> value = entry.getValue();
253 System.out.print( DF.format( key ) + "=" );
254 boolean first = true;
255 for( final Integer v : value ) {
257 System.out.print( "," );
260 System.out.print( v );
262 System.out.print( " " );
264 System.out.println();
268 private final double updateM() {
270 calculateNetDivergences();
271 Double min = Double.MAX_VALUE;
274 final int n_minus_2 = _n - 2;
275 for( int j = 1; j < _n; ++j ) {
276 final double r_j = _r[ j ];
277 final int m_j = _mappings[ j ];
280 X: for( final Entry<Double, SortedSet<Integer>> entry : _s.getSentrySet( m_j ) ) {
281 for( final int sorted_i : entry.getValue() ) {
282 //if ( counter_all >= j ) {
285 if ( _mappings[ counter ] == counter_all ) {
286 System.out.print( sorted_i + " " );
287 System.out.print( "(" + DF.format( getDvalueUnmapped( sorted_i, m_j ) ) + ") " );
288 final double m = getDvalueUnmapped( sorted_i, m_j ) - ( ( _r[ sorted_i ] + r_j ) / n_minus_2 );
290 _d_min = getDvalueUnmapped( sorted_i, m_j );
300 System.out.println();
302 for( int i = 0; i < j; ++i ) {
303 final double m = getDvalue( i, j ) - ( ( _r[ i ] + r_j ) / n_minus_2 );
306 _d_min = getDvalue( i, j );
312 System.out.println();
316 // otu2 will, in effect, be "deleted" from the matrix.
317 private final void updateMappings( final int otu2 ) {
318 for( int i = otu2; i < ( _mappings.length - 1 ); ++i ) {
319 _mappings[ i ] = _mappings[ i + 1 ];
323 public final static NeighborJoiningR createInstance() {
324 return new NeighborJoiningR();
327 public final static NeighborJoiningR createInstance( final boolean verbose,
328 final int maximum_fraction_digits_for_distances ) {
329 return new NeighborJoiningR( verbose, maximum_fraction_digits_for_distances );