1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
|
/*
Copyright (C) 2011 Tal Pupko TalP@tauex.tau.ac.il.
This program 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.
This program 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 this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "bblLS.h"
#include "numRec.h"
#include "likelihoodComputation.h"
#include "likelihoodComputationGL.h"
#include "gainLossOptions.h"
#include <cmath>
bblLS::bblLS()
{}
MDOUBLE bblLS::optimizeBranches(tree& tr, stochasticProcess* sp, const sequenceContainer &sc, Vdouble* weights, unObservableData* unObservableData_p,
const int outerIter,
const MDOUBLE epsilonOptimizationBranch, const int numIterations,
MDOUBLE curL)
{
_weights = weights;
MDOUBLE prevIterL = VERYSMALL;
if (curL == NULL)
_treeLikelihood = likelihoodComputation::getTreeLikelihoodAllPosAlphTheSame(tr,sc,*sp,_weights,unObservableData_p);
else
_treeLikelihood = curL;
LOGnOUT(4,<<"============================="<<endl;);
LOGnOUT(4,<<"ll before bbl = "<<_treeLikelihood<<endl;);
vector<tree::nodeP> nodesV;
tr.getAllNodes(nodesV,tr.getRoot());
int numberOfBranchs = nodesV.size();
MDOUBLE epsilonOptimizationIterFactor = numberOfBranchs/1.5; // (was 2) for 100 branches (~50 species) the epsilon for the entire iter is 50 times the one for branch
epsilonOptimizationIterFactor = max(5.0,epsilonOptimizationIterFactor);
MDOUBLE epsilonOptimizationIter = epsilonOptimizationBranch*epsilonOptimizationIterFactor; // for eBranch=0.2 next iteration only for 10 logL points
LOGnOUT(4,<<"BBL starts with epsilon branch= "<<epsilonOptimizationBranch<<" and epsilon iter="<<epsilonOptimizationIter<<endl;);
int iter;
for (iter = 1; iter <= numIterations; ++iter)
{
if (_treeLikelihood < prevIterL + epsilonOptimizationIter){
LOGnOUT(3,<<" BBL optimization converged. Iter= "<<iter<<" Likelihood="<<_treeLikelihood<<endl);
return _treeLikelihood; //likelihood converged
}
prevIterL = _treeLikelihood;
LOG(4,<<"---- BBL iteration: "<<iter<<endl;);
MDOUBLE paramFound;
MDOUBLE oldBl;
MDOUBLE newL;
for (int i=0; i<nodesV.size(); i++)
{
if (nodesV[i]->isRoot())
continue;
oldBl = nodesV[i]->dis2father();
if(gainLossOptions::_isBblForceFactorCorrection){
newL = -brent((oldBl+gainLossOptions::_minBranchLength)/gainLossOptions::_BblFactorCorrection,
oldBl,
(oldBl+gainLossOptions::_minBranchLength)*gainLossOptions::_BblFactorCorrection,
evalBranch(nodesV[i],&tr, sc, sp,_weights,unObservableData_p), epsilonOptimizationBranch, ¶mFound);
}
else{
newL = -brent(gainLossOptions::_minBranchLength, oldBl, gainLossOptions::_maxBranchLength, evalBranch(nodesV[i],&tr, sc, sp,_weights,unObservableData_p), epsilonOptimizationBranch, ¶mFound);
}
if (newL >= _treeLikelihood)
{
_treeLikelihood = newL;
nodesV[i]->setDisToFather(paramFound);
if(unObservableData_p) unObservableData_p->setLforMissingData(tr,sp);
LOGnOUT(4,<<"BL old... "<<oldBl<<" BL done... "<<nodesV[i]->dis2father()<<"...LL="<<_treeLikelihood<<"..."<<endl;);
}
else //likelihood went down!
{
nodesV[i]->setDisToFather(oldBl); //return to previous BL
unObservableData_p->setLforMissingData(tr,sp);
LOGnOUT(4,<<"*** WARNING: L went down : "<<endl;);
LOGnOUT(4,<<" BL Found... "<<paramFound<<"...LL="<<newL<<"...";);
LOGnOUT(4,<<" BL old... "<<oldBl<<"...LL="<<_treeLikelihood<<"..."<<endl;);
}
}
string treeINodes = gainLossOptions::_outDir + "//" + "TheTree.INodes.iter" +int2string(outerIter)+ ".Inner"+ int2string(iter) + ".ph";
printTree (tr, treeINodes);
LOGnOUT(3,<<"BBL iter "<<iter<<"...LL="<<_treeLikelihood<<"..."<<endl;);
}
if (iter>numIterations)
LOGnOUT(4,<<" Too many="<<iter-1<<" iterations in BBL. Last optimized tree is used."<<endl);
return _treeLikelihood;
}
//////////////////////////////////////////////////////////////////////////
MDOUBLE bblLS::optimizeBranches(tree& tr, vector<vector<stochasticProcess*> >& spVVec,
const distribution * gainDist, const distribution * lossDist,
const sequenceContainer &sc,
Vdouble* weights, unObservableData* unObservableData_p,
const int outerIter,
const MDOUBLE epsilonOptimizationBranch , const int numIterations ,
MDOUBLE curL)
{
_weights = weights;
MDOUBLE prevIterL = VERYSMALL;
if (curL == NULL)
_treeLikelihood = likelihoodComputationGL::getTreeLikelihoodAllPosAlphTheSame(tr,sc,spVVec,gainDist,lossDist,weights,unObservableData_p);
else
_treeLikelihood = curL;
LOGnOUT(4,<<"============================="<<endl;);
LOGnOUT(4,<<"ll before bbl = "<<_treeLikelihood<<endl;);
vector<tree::nodeP> nodesV;
tr.getAllNodes(nodesV,tr.getRoot());
int numberOfBranchs = nodesV.size();
MDOUBLE epsilonOptimizationIterFactor = numberOfBranchs/2.0; // for 100 branches (~50 species) the epsilon for the entire iter is 50 times the one for branch
epsilonOptimizationIterFactor = max(5.0,epsilonOptimizationIterFactor);
MDOUBLE epsilonOptimizationIter = epsilonOptimizationBranch*epsilonOptimizationIterFactor; // for eBranch=0.2 next iteration only for 10 logL points
LOGnOUT(4,<<"BBL starts with epsilon branch= "<<epsilonOptimizationBranch<<" and epsilon iter="<<epsilonOptimizationIter<<endl;);
int iter;
for (iter = 1; iter <= numIterations; ++iter)
{
if (_treeLikelihood < prevIterL + epsilonOptimizationIter){
LOGnOUT(3,<<" BBL optimization converged. Iter= "<<iter<<" Likelihood="<<_treeLikelihood<<endl);
return _treeLikelihood; //likelihood converged
}
prevIterL = _treeLikelihood;
LOG(4,<<"---- BBL iteration: "<<iter<<endl;);
MDOUBLE paramFound;
MDOUBLE oldBl;
MDOUBLE newL;
for (int i=0; i<numberOfBranchs; i++)
{
if (nodesV[i]->isRoot())
continue;
oldBl = nodesV[i]->dis2father();
if(gainLossOptions::_isBblForceFactorCorrection){
newL = -brent((oldBl+gainLossOptions::_minBranchLength)/gainLossOptions::_BblFactorCorrection,
oldBl,
(oldBl+gainLossOptions::_minBranchLength)*gainLossOptions::_BblFactorCorrection, evalBranchSPvv(nodesV[i],&tr, sc, spVVec,gainDist,lossDist,weights,unObservableData_p), epsilonOptimizationBranch, ¶mFound);
}
else{
newL = -brent(gainLossOptions::_minBranchLength, oldBl, gainLossOptions::_maxBranchLength, evalBranchSPvv(nodesV[i],&tr, sc, spVVec,gainDist,lossDist,weights,unObservableData_p), epsilonOptimizationBranch, ¶mFound);
}
if (newL >= _treeLikelihood)
{
_treeLikelihood = newL;
nodesV[i]->setDisToFather(paramFound);
if(unObservableData_p) unObservableData_p->setLforMissingData(tr,spVVec,gainDist,lossDist);
LOGnOUT(4,<<"BL old... "<<oldBl<<" BL done... "<<nodesV[i]->dis2father()<<"...LL="<<_treeLikelihood<<"..."<<endl;);
}
else //likelihood went down!
{
nodesV[i]->setDisToFather(oldBl); //return to previous BL
if(unObservableData_p) unObservableData_p->setLforMissingData(tr,spVVec,gainDist,lossDist);
LOGnOUT(4,<<"*** WARNING: L went down: "<<endl;);
LOGnOUT(4,<<" BL Found... "<<paramFound<<"...LL="<<newL<<"...";);
LOGnOUT(4,<<" BL old... "<<oldBl<<"...LL="<<_treeLikelihood<<"..."<<endl;);
}
}
string treeINodes = gainLossOptions::_outDir + "//" + "TheTree.INodes.iter" +int2string(outerIter)+ ".Inner"+ int2string(iter) + ".ph";
printTree (tr, treeINodes);
LOGnOUT(3,<<"BBL iter "<<iter<<"...LL="<<_treeLikelihood<<"..."<<endl;);
}
if (iter>numIterations)
LOGnOUT(4,<<" Too many="<<iter-1<<" iterations in BBL. Last optimized tree is used."<<endl);
return _treeLikelihood;
}
//////////////////////////////////////////////////////////////////////////
MDOUBLE evalBranch::operator()(MDOUBLE x)
{
_pNode->setDisToFather(x);
if(_unObservableData_p)_unObservableData_p->setLforMissingData(*_tr,_sp);
MDOUBLE LL = likelihoodComputation::getTreeLikelihoodAllPosAlphTheSame(*_tr,_sc,*_sp,_weights,_unObservableData_p);
return -LL;
}
//////////////////////////////////////////////////////////////////////////
MDOUBLE evalBranchSPvv::operator()(MDOUBLE x)
{
_pNode->setDisToFather(x);
if(_unObservableData_p) _unObservableData_p->setLforMissingData(*_tr,_spVVec,_gainDist,_lossDist);
MDOUBLE LL = likelihoodComputationGL::getTreeLikelihoodAllPosAlphTheSame(*_tr,_sc,_spVVec,_gainDist,_lossDist,_weights,_unObservableData_p);
return -LL;
}
//////////////////////////////////////////////////////////////////////////
MDOUBLE evalBranchProportionExponent::operator()(MDOUBLE x)
{
MDOUBLE factorBL = pow(10,x);
_tr->multipleAllBranchesByFactor(factorBL);
if(_unObservableData_p)_unObservableData_p->setLforMissingData(*_tr,_sp);
MDOUBLE LL = likelihoodComputation::getTreeLikelihoodAllPosAlphTheSame(*_tr,_sc,*_sp,_weights,_unObservableData_p);
_tr->multipleAllBranchesByFactor(1/factorBL);
LOG(5,<<"Branch factor val = "<<factorBL<<" logL = "<<LL<<endl);
return -LL;
}
//////////////////////////////////////////////////////////////////////////
MDOUBLE evalBranchProportionExponentSPvv::operator()(MDOUBLE x)
{
MDOUBLE factorBL = pow(10,x);
_tr->multipleAllBranchesByFactor(factorBL);
if(_unObservableData_p) _unObservableData_p->setLforMissingData(*_tr,_spVVec,_gainDist,_lossDist);
MDOUBLE LL = likelihoodComputationGL::getTreeLikelihoodAllPosAlphTheSame(*_tr,_sc,_spVVec,_gainDist,_lossDist,_weights,_unObservableData_p);
LOG(5,<<"Branch factor val = "<<factorBL<<" logL = "<<LL<<endl);
_tr->multipleAllBranchesByFactor(1/factorBL);
return -LL;
}
|
