bpp-phyl  2.1.0
bpp::MarkovModulatedSubstitutionModel Class Reference

Partial implementation of the Markov-modulated class of substitution models. More...

#include <Bpp/Phyl/Model/MarkovModulatedSubstitutionModel.h>

+ Inheritance diagram for bpp::MarkovModulatedSubstitutionModel:
+ Collaboration diagram for bpp::MarkovModulatedSubstitutionModel:

List of all members.

Public Member Functions

 MarkovModulatedSubstitutionModel (ReversibleSubstitutionModel *model, bool normalizeRateChanges, const std::string &prefix)
 Build a new MarkovModulatedSubstitutionModel object.
 MarkovModulatedSubstitutionModel (const MarkovModulatedSubstitutionModel &model)
MarkovModulatedSubstitutionModeloperator= (const MarkovModulatedSubstitutionModel &model)
virtual ~MarkovModulatedSubstitutionModel ()
MarkovModulatedSubstitutionModelclone () const =0
const AlphabetgetAlphabet () const
size_t getNumberOfStates () const
 Get the number of states.
const VdoublegetFrequencies () const
const Matrix< double > & getExchangeabilityMatrix () const
const Matrix< double > & getGenerator () const
const Matrix< double > & getPij_t (double t) const
const Matrix< double > & getdPij_dt (double t) const
const Matrix< double > & getd2Pij_dt2 (double t) const
const VdoublegetEigenValues () const
const VdoublegetIEigenValues () const
bool isDiagonalizable () const
bool isNonSingular () const
const Matrix< double > & getRowLeftEigenVectors () const
const Matrix< double > & getColumnRightEigenVectors () const
double freq (size_t i) const
double Sij (size_t i, size_t j) const
double Qij (size_t i, size_t j) const
double Pij_t (size_t i, size_t j, double t) const
double dPij_dt (size_t i, size_t j, double t) const
double d2Pij_dt2 (size_t i, size_t j, double t) const
double getInitValue (size_t i, int state) const throw (IndexOutOfBoundsException, BadIntException)
void setFreqFromData (const SequenceContainer &data, double pseudoCount=0)
 Set equilibrium frequencies equal to the frequencies estimated from the data.
const std::vector< int > & getAlphabetChars () const
int getAlphabetChar (size_t i) const
 Get the char in the alphabet corresponding to a given state in the model.
std::vector< size_t > getModelStates (int i) const
 Get the state in the model corresponding to a particular char in the alphabet.
const ReversibleSubstitutionModelgetNestedModel () const
size_t getRate (size_t i) const
 Get the rate category corresponding to a particular state in the compound model.
double getRate () const
 Get the rate.
void setRate (double rate)
 Set the rate of the model (must be positive).
double getScale () const
 Get the scalar product of diagonal elements of the generator and the frequencies vector. If the generator is normalized, then scale=1. Otherwise each element must be multiplied by 1/scale.
void setScale (double scale)
 Set the rate of the generator, defined as the scalar product of diagonal elements of the generator and the frequencies vector.
void enableEigenDecomposition (bool yn)
 Set if eigenValues and Vectors must be computed.
bool enableEigenDecomposition ()
 Tell if eigenValues and Vectors must be computed.
virtual void fireParameterChanged (const ParameterList &parameters)
 Tells the model that a parameter value has changed.
void setNamespace (const std::string &prefix)
virtual std::string getName () const =0
 Get the name of the model.
virtual void addRateParameter ()=0
virtual void setFreq (std::map< int, double > &frequencies)
 Set equilibrium frequencies.
virtual const FrequenciesSetgetFrequenciesSet () const
 If the model owns a FrequenciesSet, returns a pointer to it, otherwise return 0.
virtual size_t getNumberOfIndependentParameters () const =0
virtual void aliasParameters (const std::string &p1, const std::string &p2)=0
virtual void unaliasParameters (const std::string &p1, const std::string &p2)=0
virtual const ParameterListgetIndependentParameters () const =0
virtual std::vector< std::string > getAlias (const std::string &name) const =0
virtual bool hasParameter (const std::string &name) const =0
virtual const ParameterListgetParameters () const =0
virtual const ParametergetParameter (const std::string &name) const =0
virtual double getParameterValue (const std::string &name) const =0
virtual void setAllParametersValues (const ParameterList &parameters)=0
virtual void setParameterValue (const std::string &name, double value)=0
virtual void setParametersValues (const ParameterList &parameters)=0
virtual bool matchParametersValues (const ParameterList &parameters)=0
virtual size_t getNumberOfParameters () const =0
virtual std::string getNamespace () const =0
virtual std::string getParameterNameWithoutNamespace (const std::string &name) const =0
const ParameterListgetIndependentParameters () const
size_t getNumberOfIndependentParameters () const
void aliasParameters (const std::string &p1, const std::string &p2)
void unaliasParameters (const std::string &p1, const std::string &p2)
std::vector< std::string > getAlias (const std::string &name) const
std::string getFrom (const std::string &name) const
bool hasParameter (const std::string &name) const
const ParameterListgetParameters () const
const ParametergetParameter (const std::string &name) const
double getParameterValue (const std::string &name) const
void setAllParametersValues (const ParameterList &parameters)
void setParameterValue (const std::string &name, double value)
void setParametersValues (const ParameterList &parameters)
bool matchParametersValues (const ParameterList &parameters)
size_t getNumberOfParameters () const
std::string getNamespace () const
std::string getParameterNameWithoutNamespace (const std::string &name) const

Protected Member Functions

virtual void updateMatrices ()
virtual void updateRatesModel ()=0
 Update the rates vector, generator and equilibrium frequencies.
void addParameter_ (Parameter *parameter)
void addParameters_ (const ParameterList &parameters)
void deleteParameter_ (size_t index)
void resetParameters_ ()
ParametergetParameter_ (const std::string &name)
ParametergetParameter_ (size_t index)
const ParametergetParameter_ (size_t index) const
ParametergetParameterWithNamespace_ (const std::string &name)
const ParametergetParameterWithNamespace_ (const std::string &name) const
ParameterListgetParameters_ ()

Protected Attributes

ReversibleSubstitutionModelmodel_
size_t nbStates_
size_t nbRates_
RowMatrix< double > ratesGenerator_
std::vector< int > chars_
 The list of supported chars.
RowMatrix< double > generator_
 The generator matrix $Q$ of the model.
RowMatrix< double > exchangeability_
 The exchangeability matrix $S$ of the model.
RowMatrix< double > leftEigenVectors_
 The $U$ matrix made of left eigen vectors (by row).
RowMatrix< double > rightEigenVectors_
 The $U^-1$ matrix made of right eigen vectors (by column).
Vdouble eigenValues_
 The vector of real parts of eigen values.
Vdouble iEigenValues_
 The vector of imaginary parts of the eigen values (zero in case of reversible pmodel).
bool eigenDecompose_
 Tell if the eigen decomposition should be performed.
RowMatrix< double > pijt_
 These ones are for bookkeeping:
RowMatrix< double > dpijt_
RowMatrix< double > d2pijt_
Vdouble freq_
 The vector of equilibrium frequencies.
bool normalizeRateChanges_
std::string nestedPrefix_
Rate generator.

These variables must be initialized in the constructor of the derived class.

RowMatrix< double > rates_
RowMatrix< double > ratesExchangeability_
Vdouble ratesFreq_

Detailed Description

Partial implementation of the Markov-modulated class of substitution models.

This class wraps a substitution model and provide a matrix describing rate changes. The rate matrix must be initialized by derived classes of this class. Using these matrices, the diagonalization procedure of Galtier and Jean-Marie is used.

Such models can be described using two matrices: a substitution matrix, $M$, with size $m$, which is a "standard" substitution model of any alphabet type, and a rate matrix $G$ of size $g$. The generator of the markov-modulated model, $Q$ can be written using Kronecker matrix operands:

\[ Q=D_R \otimes M + G \otimes I_m, \]

where $D_R$ is the diagonal matrix of all rates, and $I_m$ is the identity matrix of size $m$.

This generator is normalized so that branch lengths are measured in unit of mean number of substitutions per site, where susbstitution here means "change of alphabet state". Rate changes are not counted.

Galtier N. and Jean-Marie A., Markov-modulated Markov chains and the covarion process of molecular evolution (2004). _Journal of Computational Biology_, 11:727-33.

Definition at line 75 of file MarkovModulatedSubstitutionModel.h.


Constructor & Destructor Documentation

bpp::MarkovModulatedSubstitutionModel::MarkovModulatedSubstitutionModel ( ReversibleSubstitutionModel model,
bool  normalizeRateChanges,
const std::string &  prefix 
) [inline]

Build a new MarkovModulatedSubstitutionModel object.

Parameters:
modelThe substitution model to use. Can be of any alphabet type, and will be owned by this instance.
normalizeRateChangesTells if the branch lengths must be computed in terms of rate and state
prefixThe parameter namespace to be forwarded to the AbstractParametrizable constructor. changes instead of state change only. NB: In most cases, this parameter should be set to false.

Definition at line 164 of file MarkovModulatedSubstitutionModel.h.

References bpp::AbstractParameterAliasable::addParameters_(), bpp::ParameterAliasable::getIndependentParameters(), model_, nestedPrefix_, and bpp::ParameterAliasable::setNamespace().


Member Function Documentation

double bpp::MarkovModulatedSubstitutionModel::d2Pij_dt2 ( size_t  i,
size_t  j,
double  t 
) const [inline, virtual]
Returns:
The second order derivative of the probability of change from state i to state j with respect to time t, at time t.
See also:
getd2Pij_dt2(), getStates()

Implements bpp::SubstitutionModel.

Definition at line 220 of file MarkovModulatedSubstitutionModel.h.

References getd2Pij_dt2().

double bpp::MarkovModulatedSubstitutionModel::dPij_dt ( size_t  i,
size_t  j,
double  t 
) const [inline, virtual]
Returns:
The first order derivative of the probability of change from state i to state j with respect to time t, at time t.
See also:
getdPij_dt(), getStates()

Implements bpp::SubstitutionModel.

Definition at line 219 of file MarkovModulatedSubstitutionModel.h.

References getdPij_dt().

Set if eigenValues and Vectors must be computed.

Implements bpp::SubstitutionModel.

Definition at line 280 of file MarkovModulatedSubstitutionModel.h.

References eigenDecompose_.

Tell if eigenValues and Vectors must be computed.

Implements bpp::SubstitutionModel.

Definition at line 282 of file MarkovModulatedSubstitutionModel.h.

References eigenDecompose_.

virtual void bpp::MarkovModulatedSubstitutionModel::fireParameterChanged ( const ParameterList parameters) [inline, virtual]

Tells the model that a parameter value has changed.

This updates the matrices consequently.

Reimplemented from bpp::AbstractParameterAliasable.

Reimplemented in bpp::G2001.

Definition at line 289 of file MarkovModulatedSubstitutionModel.h.

References bpp::ParameterAliasable::matchParametersValues(), model_, updateMatrices(), and updateRatesModel().

double bpp::MarkovModulatedSubstitutionModel::freq ( size_t  i) const [inline, virtual]
Returns:
Equilibrium frequency associated to character i.
See also:
getFrequencies(), getStates()

Implements bpp::SubstitutionModel.

Definition at line 214 of file MarkovModulatedSubstitutionModel.h.

References freq_.

Returns:
Get the alphabet associated to this model.

Implements bpp::SubstitutionModel.

Definition at line 191 of file MarkovModulatedSubstitutionModel.h.

References bpp::SubstitutionModel::getAlphabet(), and model_.

int bpp::MarkovModulatedSubstitutionModel::getAlphabetChar ( size_t  i) const [inline, virtual]

Get the char in the alphabet corresponding to a given state in the model.

In most cases, this method will return i.

Parameters:
iThe index of the state.
Returns:
The corresponding state in the alphabet.
See also:
MarkovModulatedSubstitutionModel
getStates()

Implements bpp::SubstitutionModel.

Definition at line 235 of file MarkovModulatedSubstitutionModel.h.

References chars_.

const std::vector<int>& bpp::MarkovModulatedSubstitutionModel::getAlphabetChars ( ) const [inline, virtual]
Returns:
The supported states of the model, as a vector of int codes.
See also:
Alphabet

Implements bpp::SubstitutionModel.

Definition at line 230 of file MarkovModulatedSubstitutionModel.h.

References chars_.

Returns:
A matrix with right eigen vectors. Each column in the matrix stands for an eigen vector.

Implements bpp::SubstitutionModel.

Definition at line 212 of file MarkovModulatedSubstitutionModel.h.

References rightEigenVectors_.

const Matrix< double > & MarkovModulatedSubstitutionModel::getd2Pij_dt2 ( double  t) const [virtual]
Returns:
All second order derivatives of the probability of change from state i to state j with respect to time t, at time t.
See also:
d2Pij_dt2()

Implements bpp::SubstitutionModel.

Definition at line 196 of file MarkovModulatedSubstitutionModel.cpp.

References d2pijt_, eigenValues_, bpp::VectorTools::exp(), leftEigenVectors_, bpp::MatrixTools::mult(), rightEigenVectors_, and bpp::NumTools::sqr().

Referenced by d2Pij_dt2().

const Matrix< double > & MarkovModulatedSubstitutionModel::getdPij_dt ( double  t) const [virtual]
Returns:
Get all first order derivatives of the probability of change from state i to state j with respect to time t, at time t.
See also:
dPij_dt()

Implements bpp::SubstitutionModel.

Definition at line 190 of file MarkovModulatedSubstitutionModel.cpp.

References dpijt_, eigenValues_, bpp::VectorTools::exp(), leftEigenVectors_, bpp::MatrixTools::mult(), and rightEigenVectors_.

Referenced by dPij_dt().

Returns:
A vector with all real parts of the eigen values of the generator of this model;

Implements bpp::SubstitutionModel.

Definition at line 205 of file MarkovModulatedSubstitutionModel.h.

References eigenValues_.

const Matrix<double>& bpp::MarkovModulatedSubstitutionModel::getExchangeabilityMatrix ( ) const [inline, virtual]
Returns:
The matrix of exchangeability terms. It is recommended that exchangeability matrix be normalized so that the normalized generator be obtained directly by the dot product $S . \pi$.

Implements bpp::SubstitutionModel.

Definition at line 197 of file MarkovModulatedSubstitutionModel.h.

References exchangeability_.

Returns:
A vector of all equilibrium frequencies.
See also:
freq()

Implements bpp::SubstitutionModel.

Definition at line 195 of file MarkovModulatedSubstitutionModel.h.

References freq_.

const Matrix<double>& bpp::MarkovModulatedSubstitutionModel::getGenerator ( ) const [inline, virtual]
Returns:
The normalized Markov generator matrix, i.e. all normalized rates of changes from state i to state j. The generator is normalized so that (i) $ \forall i; \sum_j Q_{i,j} = 0 $, meaning that $ $ \forall i; Q_{i,i} = -\sum_{j \neq i}Q_{i,j}$, and (ii) $ \sum_i Q_{i,i} \times \pi_i = -1$. This means that, under normalization, the mean rate of replacement at equilibrium is 1 and that time $t$ are measured in units of expected number of changes per site. Additionnaly, the rate_ attibute provides the possibility to increase or decrease this mean rate.

See Kosiol and Goldman (2005), Molecular Biology And Evolution 22(2) 193-9.

See also:
Qij()

Implements bpp::SubstitutionModel.

Definition at line 199 of file MarkovModulatedSubstitutionModel.h.

References generator_.

Returns:
A vector with all imaginary parts of the eigen values of the generator of this model;

Implements bpp::SubstitutionModel.

Definition at line 206 of file MarkovModulatedSubstitutionModel.h.

References iEigenValues_.

double MarkovModulatedSubstitutionModel::getInitValue ( size_t  i,
int  state 
) const throw (IndexOutOfBoundsException, BadIntException) [virtual]

This method is used to initialize likelihoods in reccursions. It typically sends 1 if i = state, 0 otherwise, where i is one of the possible states of the alphabet allowed in the model and state is the observed state in the considered sequence/site.

Parameters:
ithe index of the state in the model.
stateAn observed state in the sequence/site.
Returns:
1 or 0 depending if the two states are compatible.
Exceptions:
IndexOutOfBoundsExceptionif array position is out of range.
BadIntExceptionif states are not allowed in the associated alphabet.
See also:
getStates();

Implements bpp::SubstitutionModel.

Definition at line 204 of file MarkovModulatedSubstitutionModel.cpp.

References bpp::Alphabet::getAlias(), and bpp::BadIntException::getAlphabet().

std::vector<size_t> bpp::MarkovModulatedSubstitutionModel::getModelStates ( int  i) const [inline, virtual]

Get the state in the model corresponding to a particular char in the alphabet.

Parameters:
iThe alphabet char to check.
Returns:
A vector of indices of model states.

Implements bpp::SubstitutionModel.

Definition at line 240 of file MarkovModulatedSubstitutionModel.h.

References bpp::SubstitutionModel::getModelStates(), model_, nbRates_, and nbStates_.

virtual std::string bpp::SubstitutionModel::getName ( ) const [pure virtual, inherited]

Get the name of the model.

Returns:
The name of this model.

Implemented in bpp::YpR_Gen, bpp::YpR_Sym, bpp::F84, bpp::HKY85, bpp::JCprot, bpp::T92, bpp::K80, bpp::RN95, bpp::CodonDistanceFrequenciesSubstitutionModel, bpp::GTR, bpp::TN93, bpp::RE08, bpp::MixtureOfSubstitutionModels, bpp::JCnuc, bpp::CodonDistancePhaseFrequenciesSubstitutionModel, bpp::CodonDistanceFitnessPhaseFrequenciesSubstitutionModel, bpp::RN95s, bpp::BinarySubstitutionModel, bpp::CodonDistanceSubstitutionModel, bpp::UserProteinSubstitutionModel, bpp::MixtureOfASubstitutionModel, bpp::LLG08_EX3, bpp::WAG01, bpp::WordSubstitutionModel, bpp::LLG08_EHO, bpp::LLG08_UL3, bpp::SSR, bpp::L95, bpp::DSO78, bpp::JTT92, bpp::LGL08_CAT, bpp::LLG08_EX2, bpp::LLG08_UL2, bpp::gBGC, bpp::LG08, bpp::CodonRateFrequenciesSubstitutionModel, bpp::G2001, bpp::YNGKP_M1, bpp::YNGKP_M8, bpp::TS98, bpp::YNGKP_M7, bpp::GY94, bpp::YNGKP_M2, bpp::YNGKP_M3, bpp::TripletSubstitutionModel, bpp::YN98, bpp::CodonRateSubstitutionModel, bpp::MG94, bpp::Coala, bpp::LLG08_EX3::EmbeddedModel, bpp::LLG08_EHO::EmbeddedModel, bpp::LLG08_UL3::EmbeddedModel, bpp::LGL08_CAT::EmbeddedModel, bpp::LLG08_EX2::EmbeddedModel, and bpp::LLG08_UL2::EmbeddedModel.

Referenced by bpp::AbstractCodonFitnessSubstitutionModel::AbstractCodonFitnessSubstitutionModel(), bpp::AbstractSubstitutionModel::fireParameterChanged(), bpp::gBGC::getName(), bpp::MixtureOfSubstitutionModels::getSubmodelNumbers(), bpp::BppOSubstitutionModelFormat::readMixed_(), bpp::AbstractWordSubstitutionModel::updateMatrices(), bpp::AbstractBiblioSubstitutionModel::updateMatrices(), and bpp::BppOSubstitutionModelFormat::write().

Get the number of states.

For most models, this equals the size of the alphabet.

Returns:
The number of different states in the model.

Implements bpp::SubstitutionModel.

Definition at line 193 of file MarkovModulatedSubstitutionModel.h.

References nbRates_, and nbStates_.

const Matrix< double > & MarkovModulatedSubstitutionModel::getPij_t ( double  t) const [virtual]
Returns:
All probabilities of change from state i to state j during time t.
See also:
Pij_t()

Implements bpp::SubstitutionModel.

Definition at line 181 of file MarkovModulatedSubstitutionModel.cpp.

References eigenValues_, bpp::VectorTools::exp(), leftEigenVectors_, bpp::MatrixTools::mult(), nbRates_, nbStates_, pijt_, and rightEigenVectors_.

Referenced by Pij_t().

size_t bpp::MarkovModulatedSubstitutionModel::getRate ( size_t  i) const [inline]

Get the rate category corresponding to a particular state in the compound model.

Parameters:
iThe state.
Returns:
The corresponding rate category.
See also:
getState;

Definition at line 259 of file MarkovModulatedSubstitutionModel.h.

References nbStates_.

double bpp::MarkovModulatedSubstitutionModel::getRate ( ) const [inline, virtual]

Get the rate.

Implements bpp::SubstitutionModel.

Reimplemented in bpp::G2001, and bpp::TS98.

Definition at line 264 of file MarkovModulatedSubstitutionModel.h.

References bpp::SubstitutionModel::getRate(), and model_.

const Matrix<double>& bpp::MarkovModulatedSubstitutionModel::getRowLeftEigenVectors ( ) const [inline, virtual]
Returns:
A matrix with left eigen vectors. Each row in the matrix stands for an eigen vector.

Implements bpp::SubstitutionModel.

Definition at line 211 of file MarkovModulatedSubstitutionModel.h.

References leftEigenVectors_.

double bpp::MarkovModulatedSubstitutionModel::getScale ( ) const [inline, virtual]

Get the scalar product of diagonal elements of the generator and the frequencies vector. If the generator is normalized, then scale=1. Otherwise each element must be multiplied by 1/scale.

Returns:
Minus the scalar product of diagonal elements and the frequencies vector.

Implements bpp::SubstitutionModel.

Definition at line 268 of file MarkovModulatedSubstitutionModel.h.

References bpp::MatrixTools::diag(), freq_, and generator_.

Returns:
True if the model is diagonalizable in R.

Implements bpp::SubstitutionModel.

Definition at line 208 of file MarkovModulatedSubstitutionModel.h.

bool bpp::MarkovModulatedSubstitutionModel::isNonSingular ( ) const [inline, virtual]
Returns:
True is the model is non-singular.

Implements bpp::SubstitutionModel.

Definition at line 209 of file MarkovModulatedSubstitutionModel.h.

double bpp::MarkovModulatedSubstitutionModel::Pij_t ( size_t  i,
size_t  j,
double  t 
) const [inline, virtual]
Returns:
The probability of change from state i to state j during time t.
See also:
getPij_t(), getStates()

Implements bpp::SubstitutionModel.

Definition at line 218 of file MarkovModulatedSubstitutionModel.h.

References getPij_t().

double bpp::MarkovModulatedSubstitutionModel::Qij ( size_t  i,
size_t  j 
) const [inline, virtual]
Returns:
The rate in the generator of change from state i to state j.
See also:
getStates();

Implements bpp::SubstitutionModel.

Definition at line 216 of file MarkovModulatedSubstitutionModel.h.

References generator_.

void bpp::MarkovModulatedSubstitutionModel::setFreqFromData ( const SequenceContainer data,
double  pseudoCount = 0 
) [inline, virtual]

Set equilibrium frequencies equal to the frequencies estimated from the data.

Parameters:
dataThe sequences to use.
pseudoCountA quantity $\psi$ to add to adjust the observed values in order to prevent issues due to missing states on small data set. The corrected frequencies shall be computed as

\[ \pi_i = \frac{n_i+\psi}{\sum_j (f_j+\psi)} \]

Implements bpp::SubstitutionModel.

Definition at line 224 of file MarkovModulatedSubstitutionModel.h.

References model_, bpp::SubstitutionModel::setFreqFromData(), and updateMatrices().

void MarkovModulatedSubstitutionModel::setNamespace ( const std::string &  prefix) [virtual]

Reimplemented from bpp::AbstractParameterAliasable.

Reimplemented in bpp::G2001.

Definition at line 221 of file MarkovModulatedSubstitutionModel.cpp.

References model_, nestedPrefix_, and bpp::ParameterAliasable::setNamespace().

void bpp::MarkovModulatedSubstitutionModel::setRate ( double  rate) [inline, virtual]

Set the rate of the model (must be positive).

Parameters:
ratemust be positive.

Implements bpp::SubstitutionModel.

Reimplemented in bpp::G2001, and bpp::TS98.

Definition at line 266 of file MarkovModulatedSubstitutionModel.h.

References model_, and bpp::SubstitutionModel::setRate().

void bpp::MarkovModulatedSubstitutionModel::setScale ( double  scale) [inline, virtual]

Set the rate of the generator, defined as the scalar product of diagonal elements of the generator and the frequencies vector.

When the generator is normalized, scale=1. Otherwise each element is multiplied such that the correct scale is set.

Implements bpp::SubstitutionModel.

Definition at line 275 of file MarkovModulatedSubstitutionModel.h.

References model_, bpp::SubstitutionModel::setScale(), and updateMatrices().

double bpp::MarkovModulatedSubstitutionModel::Sij ( size_t  i,
size_t  j 
) const [inline, virtual]
Returns:
The exchangeability between state i and state j.

By definition Sij(i,j) = Sij(j,i).

Implements bpp::SubstitutionModel.

Definition at line 215 of file MarkovModulatedSubstitutionModel.h.

References exchangeability_.

virtual void bpp::MarkovModulatedSubstitutionModel::updateRatesModel ( ) [protected, pure virtual]

Update the rates vector, generator and equilibrium frequencies.

This method must be implemented by the derived class. It is called by the fireParameterChanged() method.

Implemented in bpp::G2001, and bpp::TS98.

Referenced by fireParameterChanged().


Member Data Documentation

std::vector<int> bpp::MarkovModulatedSubstitutionModel::chars_ [protected]

The list of supported chars.

Definition at line 100 of file MarkovModulatedSubstitutionModel.h.

Referenced by getAlphabetChar(), getAlphabetChars(), operator=(), and updateMatrices().

Tell if the eigen decomposition should be performed.

Definition at line 136 of file MarkovModulatedSubstitutionModel.h.

Referenced by enableEigenDecomposition(), and operator=().

The vector of real parts of eigen values.

Definition at line 125 of file MarkovModulatedSubstitutionModel.h.

Referenced by getd2Pij_dt2(), getdPij_dt(), getEigenValues(), getPij_t(), operator=(), and updateMatrices().

The exchangeability matrix $S$ of the model.

Definition at line 110 of file MarkovModulatedSubstitutionModel.h.

Referenced by getExchangeabilityMatrix(), operator=(), Sij(), and updateMatrices().

The vector of equilibrium frequencies.

Definition at line 148 of file MarkovModulatedSubstitutionModel.h.

Referenced by freq(), getFrequencies(), getScale(), operator=(), and updateMatrices().

The generator matrix $Q$ of the model.

Definition at line 105 of file MarkovModulatedSubstitutionModel.h.

Referenced by getGenerator(), getScale(), operator=(), Qij(), and updateMatrices().

The vector of imaginary parts of the eigen values (zero in case of reversible pmodel).

Definition at line 131 of file MarkovModulatedSubstitutionModel.h.

Referenced by getIEigenValues(), operator=(), and updateMatrices().

The $U$ matrix made of left eigen vectors (by row).

Definition at line 115 of file MarkovModulatedSubstitutionModel.h.

Referenced by getd2Pij_dt2(), getdPij_dt(), getPij_t(), getRowLeftEigenVectors(), operator=(), and updateMatrices().

These ones are for bookkeeping:

Definition at line 141 of file MarkovModulatedSubstitutionModel.h.

Referenced by getPij_t(), operator=(), and updateMatrices().

The $U^-1$ matrix made of right eigen vectors (by column).

Definition at line 120 of file MarkovModulatedSubstitutionModel.h.

Referenced by getColumnRightEigenVectors(), getd2Pij_dt2(), getdPij_dt(), getPij_t(), operator=(), and updateMatrices().


The documentation for this class was generated from the following files:
 All Classes Namespaces Files Functions Variables Typedefs Friends