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;
32 import org.forester.evoinference.matrix.distance.BasicSymmetricalDistanceMatrix;
33 import org.forester.phylogeny.Phylogeny;
34 import org.forester.phylogeny.PhylogenyNode;
35 import org.forester.util.ForesterUtil;
37 public final class NeighborJoining {
39 private final static DecimalFormat DF = new DecimalFormat( "0.00000" );
40 private BasicSymmetricalDistanceMatrix _d;
41 private double[][] _d_values;
42 private final DecimalFormat _df;
43 private PhylogenyNode[] _external_nodes;
44 private int[] _mappings;
47 private final boolean _verbose;
51 private NeighborJoining() {
56 private NeighborJoining( final boolean verbose, final int maximum_fraction_digits_for_distances ) {
57 if ( ( maximum_fraction_digits_for_distances < 1 ) || ( maximum_fraction_digits_for_distances > 9 ) ) {
58 throw new IllegalArgumentException( "maximum fraction digits for distances is out of range: "
59 + maximum_fraction_digits_for_distances );
62 _df = new DecimalFormat();
63 _df.setMaximumFractionDigits( maximum_fraction_digits_for_distances );
64 _df.setRoundingMode( RoundingMode.HALF_UP );
67 public final Phylogeny execute( final BasicSymmetricalDistanceMatrix distance ) {
69 final Phylogeny phylogeny = new Phylogeny();
71 // Calculates the minimal distance.
72 // If more than one minimal distances, always the first found is used
74 final int otu1 = _min_i;
75 final int otu2 = _min_j;
76 System.out.println( _min_i + " " + _min_j );
77 // It is a condition that otu1 < otu2.
78 final PhylogenyNode node = new PhylogenyNode();
79 final double d = _d_values[ _mappings[ otu1 ] ][ _mappings[ otu2 ] ];
80 final double d1 = ( d / 2 ) + ( ( _r[ otu1 ] - _r[ otu2 ] ) / ( 2 * ( _n - 2 ) ) );
81 final double d2 = d - d1;
83 getExternalPhylogenyNode( otu1 ).setDistanceToParent( d1 );
84 getExternalPhylogenyNode( otu2 ).setDistanceToParent( d2 );
87 // yes, yes, slow but only grows with n (and not n^2 or worse)...
88 getExternalPhylogenyNode( otu1 ).setDistanceToParent( Double.parseDouble( _df.format( d1 ) ) );
89 getExternalPhylogenyNode( otu2 ).setDistanceToParent( Double.parseDouble( _df.format( d2 ) ) );
91 node.addAsChild( getExternalPhylogenyNode( otu1 ) );
92 node.addAsChild( getExternalPhylogenyNode( otu2 ) );
94 printProgress( otu1, otu2 );
96 calculateDistancesFromNewNode( otu1, otu2, d );
97 _external_nodes[ _mappings[ otu1 ] ] = node;
98 updateMappings( otu2 );
101 final double d = _d_values[ _mappings[ 0 ] ][ _mappings[ 1 ] ] / 2;
103 getExternalPhylogenyNode( 0 ).setDistanceToParent( d );
104 getExternalPhylogenyNode( 1 ).setDistanceToParent( d );
107 final double dd = Double.parseDouble( _df.format( d ) );
108 getExternalPhylogenyNode( 0 ).setDistanceToParent( dd );
109 getExternalPhylogenyNode( 1 ).setDistanceToParent( dd );
111 final PhylogenyNode root = new PhylogenyNode();
112 root.addAsChild( getExternalPhylogenyNode( 0 ) );
113 root.addAsChild( getExternalPhylogenyNode( 1 ) );
115 printProgress( 0, 1 );
117 phylogeny.setRoot( root );
118 phylogeny.setRooted( false );
122 public final List<Phylogeny> execute( final List<BasicSymmetricalDistanceMatrix> distances_list ) {
123 final List<Phylogeny> pl = new ArrayList<Phylogeny>();
124 for( final BasicSymmetricalDistanceMatrix distances : distances_list ) {
125 pl.add( execute( distances ) );
130 private final void calculateDistancesFromNewNode( final int otu1, final int otu2, final double d ) {
131 final int m_otu1 = _mappings[ otu1 ];
132 final int m_otu2 = _mappings[ otu2 ];
133 for( int i = 0; i < _n; ++i ) {
134 if ( ( i == otu1 ) || ( i == otu2 ) ) {
137 final int m_i = _mappings[ i ];
140 _d_values[ m_otu1 ][ m_i ] = ( _d_values[ m_otu1 ][ m_i ] + _d_values[ m_i ][ m_otu2 ] - d ) / 2;
141 System.out.print( DF.format( _d_values[ m_otu1 ][ m_i ] ) );
144 _d_values[ m_otu1 ][ m_i ] = ( _d_values[ m_otu1 ][ m_i ] + _d_values[ m_otu2 ][ m_i ] - d ) / 2;
145 System.out.print( DF.format( _d_values[ m_otu1 ][ m_i ] ) );
151 _d_values[ m_i ][ m_otu1 ] = ( _d_values[ m_i ][ m_otu1 ] + _d_values[ m_i ][ m_otu2 ] - d ) / 2;
152 System.out.print( DF.format( _d_values[ m_i ][ m_otu1 ] ) );
156 _d_values[ m_i ][ m_otu1 ] = ( _d_values[ m_i ][ m_otu1 ] + _d_values[ m_otu2 ][ m_i ] - d ) / 2;
157 System.out.print( DF.format( _d_values[ m_otu1 ][ m_i ] ) );
161 System.out.print( " " );
165 private final void calculateNetDivergences() {
167 for( int i = 0; i < _n; ++i ) {
169 final int m_i = _mappings[ i ];
170 for( int n = 0; n < _n; ++n ) {
173 d += _d_values[ _mappings[ n ] ][ m_i ];
176 d += _d_values[ m_i ][ _mappings[ n ] ];
184 private final PhylogenyNode getExternalPhylogenyNode( final int i ) {
185 return _external_nodes[ _mappings[ i ] ];
188 private final void initExternalNodes() {
189 _external_nodes = new PhylogenyNode[ _n ];
191 for( int i = 0; i < _n; ++i ) {
192 _external_nodes[ i ] = new PhylogenyNode();
193 id = _d.getIdentifier( i );
195 _external_nodes[ i ].setName( id );
198 _external_nodes[ i ].setName( Integer.toString( i ) );
204 private final void printProgress( final int otu1, final int otu2 ) {
205 System.out.println( "Node " + printProgressNodeToString( getExternalPhylogenyNode( otu1 ) ) + " joins "
206 + ( printProgressNodeToString( getExternalPhylogenyNode( otu2 ) ) ) );
209 private final String printProgressNodeToString( final PhylogenyNode n ) {
210 if ( n.isExternal() ) {
211 if ( ForesterUtil.isEmpty( n.getName() ) ) {
212 return Long.toString( n.getId() );
218 + ( ForesterUtil.isEmpty( n.getChildNode1().getName() ) ? n.getChildNode1().getId() : n.getChildNode1()
221 + ( ForesterUtil.isEmpty( n.getChildNode2().getName() ) ? n.getChildNode2().getId() : n.getChildNode2()
225 // only the values in the lower triangle are used.
226 // !matrix values will be changed!
227 private final void reset( final BasicSymmetricalDistanceMatrix distances ) {
228 _n = distances.getSize();
230 _r = new double[ _n ];
231 _mappings = new int[ _n ];
232 _d_values = _d.getValues();
236 private final void updateM() {
237 calculateNetDivergences();
238 Double min = Double.MAX_VALUE;
241 final int n_minus_2 = _n - 2;
242 for( int j = 1; j < _n; ++j ) {
243 final double r_j = _r[ j ];
244 final int m_j = _mappings[ j ];
245 for( int i = 0; i < j; ++i ) {
246 final double m = _d_values[ _mappings[ i ] ][ m_j ] - ( ( _r[ i ] + r_j ) / n_minus_2 );
254 for( int j = 1; j < _n; ++j ) {
255 final double r_j = _r[ j ];
256 final int m_j = _mappings[ j ];
257 for( int i = 0; i < j; ++i ) {
258 System.out.print( i );
259 System.out.print( "->" );
260 System.out.print( DF.format( _r[ i ] ) );
261 System.out.print( " " );
263 System.out.println();
267 // otu2 will, in effect, be "deleted" from the matrix.
268 private final void updateMappings( final int otu2 ) {
269 for( int i = otu2; i < ( _mappings.length - 1 ); ++i ) {
270 System.out.print( _mappings[ i ] );
271 _mappings[ i ] = _mappings[ i + 1 ];
272 System.out.println( "----->" + _mappings[ i ] );
274 for( int i = 0; i < _mappings.length; ++i ) {
275 System.out.println( i + "-->" + _mappings[ i ] );
279 public final static NeighborJoining createInstance() {
280 return new NeighborJoining();
283 public final static NeighborJoining createInstance( final boolean verbose,
284 final int maximum_fraction_digits_for_distances ) {
285 return new NeighborJoining( verbose, maximum_fraction_digits_for_distances );