2 // FORESTER -- software libraries and applications
3 // for evolutionary biology research and applications.
5 // Copyright (C) 2008-2009 Christian M. Zmasek
6 // Copyright (C) 2008-2009 Burnham Institute for Medical Research
9 // This library is free software; you can redistribute it and/or
10 // modify it under the terms of the GNU Lesser General Public
11 // License as published by the Free Software Foundation; either
12 // version 2.1 of the License, or (at your option) any later version.
14 // This library is distributed in the hope that it will be useful,
15 // but WITHOUT ANY WARRANTY; without even the implied warranty of
16 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 // Lesser General Public License for more details.
19 // You should have received a copy of the GNU Lesser General Public
20 // License along with this library; if not, write to the Free Software
21 // Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
23 // Contact: phylosoft @ gmail . com
24 // WWW: https://sites.google.com/site/cmzmasek/home/software/forester
26 package org.forester.evoinference.distance;
28 import java.math.RoundingMode;
29 import java.text.DecimalFormat;
30 import java.util.ArrayList;
31 import java.util.List;
33 import org.forester.evoinference.matrix.distance.BasicSymmetricalDistanceMatrix;
34 import org.forester.phylogeny.Phylogeny;
35 import org.forester.phylogeny.PhylogenyNode;
36 import org.forester.util.ForesterUtil;
38 public final class NeighborJoining {
40 private BasicSymmetricalDistanceMatrix _d;
41 private double[][] _d_values;
42 private double[][] _m_values;
45 private PhylogenyNode[] _external_nodes;
46 private int[] _mappings;
47 private final boolean _verbose;
48 private final DecimalFormat _df;
50 private NeighborJoining() {
55 private NeighborJoining( final boolean verbose, final int maximum_fraction_digits_for_distances ) {
56 if ( ( maximum_fraction_digits_for_distances < 1 ) || ( maximum_fraction_digits_for_distances > 9 ) ) {
57 throw new IllegalArgumentException( "maximum fraction digits for distances is out of range: "
58 + maximum_fraction_digits_for_distances );
61 _df = new DecimalFormat();
62 _df.setMaximumFractionDigits( maximum_fraction_digits_for_distances );
63 _df.setRoundingMode( RoundingMode.HALF_UP );
66 private final void printM() {
67 System.out.println( "M:" );
68 for( final double[] _m_value : _m_values ) {
69 for( int j = 0; j < _m_values.length; j++ ) {
70 System.out.print( _m_value[ j ] );
71 System.out.print( " " );
78 private final void printD() {
79 System.out.println( "D:" );
80 for( final double[] _d_value : _d_values ) {
81 for( int j = 0; j < _d_values.length; j++ ) {
82 System.out.print( _d_value[ j ] );
83 System.out.print( " " );
90 private final void calculateDistancesFromNewNode( final int otu1, final int otu2, final double d ) {
91 // final int otu1_m = _mappings[ otu1 ];
92 // final int otu2_m = _mappings[ otu2 ];
94 for( int i = 0; i < _n; ++i ) {
95 if ( ( i == otu1 ) || ( i == otu2 ) ) {
99 _d_values[ _mappings[ otu1 ] ][ _mappings[ i ] ] = ( getValueFromD( otu1, i ) + getValueFromD( i, otu2 ) - d ) / 2;
102 _d_values[ _mappings[ i ] ][ _mappings[ otu1 ] ] = ( getValueFromD( otu1, i ) + getValueFromD( i, otu2 ) - d ) / 2;
104 //i_m = _mappings[ i ];
105 //_d_values[ otu1_m ][ i_m ] = ( ( _d_values[ otu1_m ][ i_m ] + _d_values[ i_m ][ otu2_m ] ) - 2 ) / 2;
109 private final void calculateNetDivergences() {
112 for( int i = 0; i < _n; ++i ) {
114 //i_m = _mappings[ i ];
115 for( int n = 0; n < _n; ++n ) {
116 //d += _d_values[ i_m ][ _mappings[ n ] ];
118 d += getValueFromD( i, n );
125 public final Phylogeny execute( final BasicSymmetricalDistanceMatrix distance ) {
127 final Phylogeny phylogeny = new Phylogeny();
130 // Calculates the minimal distance.
131 // If more than one minimal distances, always the first found is used
132 // could randomize this, so that any would be returned in a randomized fashion...
133 double minimum = _m_values[ 0 ][ 1 ];
136 for( int j = 1; j < _n; ++j ) {
137 for( int i = 0; i < j; ++i ) {
138 if ( _m_values[ i ][ j ] < minimum ) {
139 minimum = _m_values[ i ][ j ];
145 // It is a condition that otu1 < otu2.
146 final PhylogenyNode node = new PhylogenyNode();
147 final double d = getValueFromD( otu1, otu2 );
148 final double d1 = ( d / 2 ) + ( ( _r[ otu1 ] - _r[ otu2 ] ) / ( 2 * ( _n - 2 ) ) );
149 final double d2 = d - d1;
151 getExternalPhylogenyNode( otu1 ).setDistanceToParent( d1 );
152 getExternalPhylogenyNode( otu2 ).setDistanceToParent( d2 );
155 // yes, yes, slow but only grows with n (and not n^2 or worse)...
156 getExternalPhylogenyNode( otu1 ).setDistanceToParent( Double.parseDouble( _df.format( d1 ) ) );
157 getExternalPhylogenyNode( otu2 ).setDistanceToParent( Double.parseDouble( _df.format( d2 ) ) );
159 node.addAsChild( getExternalPhylogenyNode( otu1 ) );
160 node.addAsChild( getExternalPhylogenyNode( otu2 ) );
162 printProgress( otu1, otu2 );
164 calculateDistancesFromNewNode( otu1, otu2, d );
165 _external_nodes[ _mappings[ otu1 ] ] = node;
166 updateMappings( otu2 );
169 final double d = getValueFromD( 0, 1 ) / 2;
171 getExternalPhylogenyNode( 0 ).setDistanceToParent( d );
172 getExternalPhylogenyNode( 1 ).setDistanceToParent( d );
175 final double dd = Double.parseDouble( _df.format( d ) );
176 getExternalPhylogenyNode( 0 ).setDistanceToParent( dd );
177 getExternalPhylogenyNode( 1 ).setDistanceToParent( dd );
179 final PhylogenyNode root = new PhylogenyNode();
180 root.addAsChild( getExternalPhylogenyNode( 0 ) );
181 root.addAsChild( getExternalPhylogenyNode( 1 ) );
183 printProgress( 0, 1 );
185 phylogeny.setRoot( root );
186 phylogeny.setRooted( false );
190 public final List<Phylogeny> execute( final List<BasicSymmetricalDistanceMatrix> distances_list ) {
191 final List<Phylogeny> pl = new ArrayList<Phylogeny>();
192 for( final BasicSymmetricalDistanceMatrix distances : distances_list ) {
193 pl.add( execute( distances ) );
198 private final PhylogenyNode getExternalPhylogenyNode( final int i ) {
199 return _external_nodes[ _mappings[ i ] ];
202 private final double getValueFromD( final int otu1, final int otu2 ) {
204 return _d_values[ _mappings[ otu2 ] ][ _mappings[ otu1 ] ];
206 return _d_values[ _mappings[ otu1 ] ][ _mappings[ otu2 ] ];
209 private final void initExternalNodes() {
210 _external_nodes = new PhylogenyNode[ _n ];
212 for( int i = 0; i < _n; ++i ) {
213 _external_nodes[ i ] = new PhylogenyNode();
214 id = _d.getIdentifier( i );
216 _external_nodes[ i ].setName( id );
219 _external_nodes[ i ].setName( "" + i );
225 private final void printProgress( final int otu1, final int otu2 ) {
226 final PhylogenyNode n1 = getExternalPhylogenyNode( otu1 );
227 final PhylogenyNode n2 = getExternalPhylogenyNode( otu2 );
228 System.out.println( "Node " + ( ForesterUtil.isEmpty( n1.getName() ) ? n1.getId() : n1.getName() ) + " joins "
229 + ( ForesterUtil.isEmpty( n2.getName() ) ? n2.getId() : n2.getName() ) );
232 // only the values in the lower triangle are used.
233 // !matrix values will be changed!
234 private final void reset( final BasicSymmetricalDistanceMatrix distances ) {
235 _n = distances.getSize();
237 _r = new double[ _n ];
238 _mappings = new int[ _n ];
239 _d_values = _d.getValues();
240 _m_values = new double[ _n ][ _n ];
244 private final void updateM() {
245 calculateNetDivergences();
248 final int _n_2 = _n - 2;
249 for( int j = 1; j < _n; ++j ) {
251 // j_m = _mappings[ j ];
252 for( int i = 0; i < j; ++i ) {
253 _m_values[ i ][ j ] = getValueFromD( i, j ) - ( ( _r[ i ] + _r[ j ] ) / ( _n - 2 ) );
254 //_m_values[ i ][ j ] = _d_values[ _mappings[ i ] ][ j_m ] - ( ( _r[ i ] + r_j ) / ( _n_2 ) );
259 // otu2 will, in effect, be "deleted" from the matrix.
260 private final void updateMappings( final int otu2 ) {
261 for( int i = otu2; i < ( _mappings.length - 1 ); ++i ) {
262 _mappings[ i ] = _mappings[ i + 1 ];
266 public final static NeighborJoining createInstance() {
267 return new NeighborJoining();
270 public final static NeighborJoining createInstance( final boolean verbose,
271 final int maximum_fraction_digits_for_distances ) {
272 return new NeighborJoining( verbose, maximum_fraction_digits_for_distances );