Package org.biojava.bio.dp

Class DP

java.lang.Object

org.biojava.bio.dp.DP

Direct Known Subclasses:

PairwiseDP, SingleDP

public abstract class DP
extends Object

Objects that can perform dymamic programming operations upon sequences with HMMs.

The three main DP operations are Forwards, Backwards and Viterbi. Forwards and Backwards calculate the probability of the sequences having been made in any way by the model. Viterbi finds the most supported way that the sequence could have been made.

Each of the functions can return the dynamic-programming matrix containing the intermediate results. This may be useful for model training, or for visualisation.

Each of the funcitons can be calculated using the model probabilities, the null-model probabilities or the odds (ratio between the two). For Forwards and Backwards, the odds calculations produce numbers with questionable basis in reality. For Viterbi with odds, you will recieve the path through the model that is most different from the null model, and supported by the probabilities.

Author:

Matthew Pocock, Thomas Down

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	DP.ReverseIterator

Constructor Summary

Constructors

Constructor	Description
DP()	This method will result in a DP with no model.
DP(MarkovModel model)

Method Summary

Modifier and Type	Method	Description
abstract double	backward(SymbolList[] symList, ScoreType scoreType)
abstract DPMatrix	backwardMatrix(SymbolList[] symList, DPMatrix matrix, ScoreType scoreType)
abstract DPMatrix	backwardMatrix(SymbolList[] symList, ScoreType scoreType)
static int[][]	backwardTransitions(MarkovModel model, State[] states)
static double[][]	backwardTransitionScores(MarkovModel model, State[] states, int[][] transitions, ScoreType scoreType)
static MarkovModel	flatView(MarkovModel model)
abstract double	forward(SymbolList[] symList, ScoreType scoreType)
abstract DPMatrix	forwardMatrix(SymbolList[] symList, DPMatrix matrix, ScoreType scoreType)
abstract DPMatrix	forwardMatrix(SymbolList[] symList, ScoreType scoreType)
DPMatrix	forwardsBackwards(SymbolList[] symList, ScoreType scoreType)
static int[][]	forwardTransitions(MarkovModel model, State[] states)	Returns a matrix for the specified States describing all valid Transitions between those States.
static double[][]	forwardTransitionScores(MarkovModel model, State[] states, int[][] transitions, ScoreType scoreType)	Compute the log(score) of all transitions between the specified States.
StatePath	generate(int length)	Generates an alignment from a model.
int[][]	getBackwardTransitions()
double[][]	getBackwardTransitionScores(ScoreType scoreType)
int	getDotStatesIndex()
int[][]	getForwardTransitions()
double[][]	getForwardTransitionScores(ScoreType scoreType)
MarkovModel	getModel()
State[]	getStates()
void	lockModel()
static double	scoreWeightMatrix(WeightMatrix matrix, SymbolList symList, int start)	Scores the SymbolList from symbol start to symbol (start+columns) with a weight matrix.
static double	scoreWeightMatrix(WeightMatrix matrix, SymbolList symList, ScoreType scoreType, int start)	Scores the SymbolList from symbol start to symbol (start+columns) with a weight matrix using a particular ScoreType.
void	setModel(MarkovModel model)
State[]	stateList(MarkovModel mm)
void	unlockModel()
void	update()
abstract StatePath	viterbi(SymbolList[] symList, ScoreType scoreType)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DP

public DP(MarkovModel model)

DP

public DP()

This method will result in a DP with no model. Use the setModel() method to set the model before use.

Method Detail

scoreWeightMatrix

public static double scoreWeightMatrix(WeightMatrix matrix,
                                       SymbolList symList,
                                       int start)
                                throws IllegalSymbolException

Scores the SymbolList from symbol start to symbol (start+columns) with a weight matrix.

Parameters:

matrix - the weight matrix used to evaluate the sequences

symList - the SymbolList to assess

start - the index of the first symbol in the window to evaluate

Returns:

the log probability or likelyhood of this weight matrix having generated symbols start to (start + columns) of symList

Throws:

IllegalSymbolException

scoreWeightMatrix

public static double scoreWeightMatrix(WeightMatrix matrix,
                                       SymbolList symList,
                                       ScoreType scoreType,
                                       int start)
                                throws IllegalSymbolException

Scores the SymbolList from symbol start to symbol (start+columns) with a weight matrix using a particular ScoreType.

This method allows you to use score types such as ScoreType.ODDS. The other scoreWeightMatrix methods gives a result similar or identical to ScoreType.PROBABILITY.

Parameters:

matrix - the weight matrix used to evaluate the sequences

symList - the SymbolList to assess

scoreType - the score type to apply

start - the index of the first symbol in the window to evaluate

Returns:

the sum of log scores of this weight matrix having generated symbols start to (start + columns) of symList

Throws:

IllegalSymbolException

Since:

1.4

flatView

public static MarkovModel flatView(MarkovModel model)
                            throws IllegalAlphabetException,
                                   IllegalSymbolException

Throws:

IllegalAlphabetException

IllegalSymbolException

stateList

public State[] stateList(MarkovModel mm)
                  throws IllegalSymbolException,
                         IllegalTransitionException,
                         BioException

Throws:

IllegalSymbolException

IllegalTransitionException

BioException

forwardTransitions

public static int[][] forwardTransitions(MarkovModel model,
                                         State[] states)
                                  throws IllegalSymbolException

Returns a matrix for the specified States describing all valid Transitions between those States.

The matrix is 2-dimensional. The primary array has an element corresponding to every State in the states argument. That element is itself an array the elements of which identify the States that can reach that State. The source States are identified by their index within the states [] array.

Parameters:

model - MarkovModel to be analysed.

states - The States for which the transition matrix is to be determined.

Throws:

IllegalSymbolException

forwardTransitionScores

public static double[][] forwardTransitionScores(MarkovModel model,
                                                 State[] states,
                                                 int[][] transitions,
                                                 ScoreType scoreType)

Compute the log(score) of all transitions between the specified States. The layout of the array is identical to that of the transitions array.

Note that all parameters MUST be consistent with each other!!!!

Parameters:

model - The model for which the data is to be computed.

states - The States within that model for which scores are required.

transitions - The transition matrix obtained by calling forwardTransitions() with the above argument values.

scoreType - The type of score to be evaluated.

backwardTransitions

public static int[][] backwardTransitions(MarkovModel model,
                                          State[] states)
                                   throws IllegalSymbolException

Throws:

IllegalSymbolException

backwardTransitionScores

public static double[][] backwardTransitionScores(MarkovModel model,
                                                  State[] states,
                                                  int[][] transitions,
                                                  ScoreType scoreType)

getDotStatesIndex

public int getDotStatesIndex()

getModel

public MarkovModel getModel()

getStates

public State[] getStates()

getForwardTransitions

public int[][] getForwardTransitions()

getForwardTransitionScores

public double[][] getForwardTransitionScores(ScoreType scoreType)

getBackwardTransitions

public int[][] getBackwardTransitions()

getBackwardTransitionScores

public double[][] getBackwardTransitionScores(ScoreType scoreType)

lockModel

public void lockModel()

unlockModel

public void unlockModel()

update

public void update()

setModel

public void setModel(MarkovModel model)

forward

public abstract double forward(SymbolList[] symList,
                               ScoreType scoreType)
                        throws IllegalSymbolException,
                               IllegalAlphabetException,
                               IllegalTransitionException

Throws:

IllegalSymbolException

IllegalAlphabetException

IllegalTransitionException

backward

public abstract double backward(SymbolList[] symList,
                                ScoreType scoreType)
                         throws IllegalSymbolException,
                                IllegalAlphabetException,
                                IllegalTransitionException

Throws:

IllegalSymbolException

IllegalAlphabetException

IllegalTransitionException

forwardMatrix

public abstract DPMatrix forwardMatrix(SymbolList[] symList,
                                       ScoreType scoreType)
                                throws IllegalSymbolException,
                                       IllegalAlphabetException,
                                       IllegalTransitionException

Throws:

IllegalSymbolException

IllegalAlphabetException

IllegalTransitionException

backwardMatrix

public abstract DPMatrix backwardMatrix(SymbolList[] symList,
                                        ScoreType scoreType)
                                 throws IllegalSymbolException,
                                        IllegalAlphabetException,
                                        IllegalTransitionException

Throws:

IllegalSymbolException

IllegalAlphabetException

IllegalTransitionException

forwardMatrix

public abstract DPMatrix forwardMatrix(SymbolList[] symList,
                                       DPMatrix matrix,
                                       ScoreType scoreType)
                                throws IllegalArgumentException,
                                       IllegalSymbolException,
                                       IllegalAlphabetException,
                                       IllegalTransitionException

Throws:

IllegalArgumentException

IllegalSymbolException

IllegalAlphabetException

IllegalTransitionException

backwardMatrix

public abstract DPMatrix backwardMatrix(SymbolList[] symList,
                                        DPMatrix matrix,
                                        ScoreType scoreType)
                                 throws IllegalArgumentException,
                                        IllegalSymbolException,
                                        IllegalAlphabetException,
                                        IllegalTransitionException

Throws:

IllegalArgumentException

IllegalSymbolException

IllegalAlphabetException

IllegalTransitionException

viterbi

public abstract StatePath viterbi(SymbolList[] symList,
                                  ScoreType scoreType)
                           throws IllegalSymbolException,
                                  IllegalArgumentException,
                                  IllegalAlphabetException,
                                  IllegalTransitionException

Throws:

IllegalSymbolException

IllegalArgumentException

IllegalAlphabetException

IllegalTransitionException

forwardsBackwards

public DPMatrix forwardsBackwards(SymbolList[] symList,
                                  ScoreType scoreType)
                           throws BioException

Throws:

BioException

generate

public StatePath generate(int length)
                   throws IllegalSymbolException,
                          BioException

Generates an alignment from a model.

If the length is set to -1 then the model length will be sampled using the model's transition to the end state. If the length is fixed using length, then the transitions to the end state are implicitly invoked.

Parameters:

length - the length of the sequence to generate

Returns:

a StatePath generated at random

Throws:

IllegalSymbolException

BioException