Public Member Functions
	ProbabilisticAssemblingModel (const string &folderpath, ModelBehaviour behav=PREDEFINED)
	Model constructor from the given folder with a model's parameters. More...

vector< prob_t >	computeFullProbabilities (const ClonesetView &repertoire, ErrorMode error_mode, SequenceType sequence_type=NUCLEOTIDE, MAAGComputeProbAction action=SUM_PROBABILITY) const
	Compute full assembling probabilities of clonotypes from the given repertoire. If some clonotypes have few alignments of some gene segments, than full probability of all combinations of gene segments will be computed and either summed or the max full probability will be chosen. More...

MAAGRepertoire	buildGraphs (const ClonesetView &repertoire, MetadataMode save_metadata, ErrorMode error_mode, SequenceType sequence_type=NUCLEOTIDE, bool verbose=true) const
	Build a set of MAAGs from the given cloneset. More...

void	updateEventProbabilities (MAAGRepertoire *repertoire, bool verbose=true)
	Update event probabilities in the given MAAG repertoire with new ones.

Cloneset	generateSequences (size_t count=1, bool verbose=true) const
	Generate artificial repertoire of sequences using this model. More...

const VDJRecombinationGenes &	gene_segments () const
	Access to the gene segments table. More...

const ModelParameterVector &	event_probabilities () const
	Access to vector of probabilities.

void	updateModelParameterVector (const ModelParameterVector &vec)
	Set a new event probabilities vector to this model. More...

void	updateGeneUsage (const ClonesetView &cloneset)
	Given the cloneset, compute its gene usage on out-of-frames and update the event probability vector with new gene probabilities.

bool	save (const string &folderpath) const
	Save the model to the given folder. More...

bool	status () const
	Return if the model is in the ready-to-work state. More...

Recombination	recombination () const

string	name () const

Protected Member Functions
	ProbabilisticAssemblingModel ()
	Private default constructor.

void	parseModelConfig (const string &folderpath)
	Parse JSON file with model parameters.

void	make_builder ()
	Make MAAGBuilder with this model's parameters (event probabilities and gene segments).

void	make_assembler ()
	Make ClonotypeAssembler with this model's parameters (event probabilities and gene segments).

void	save_vj (const string &folderpath) const

void	save_vdj (const string &folderpath) const

Protected Attributes
bool	_status

bool	_verbose

ModelBehaviour	_behaviour

Json::Value	_config

Recombination	_recomb

string	_model_path

unique_ptr< ModelParser >	_parser

unique_ptr< VDJRecombinationGenes >	_genes

unique_ptr< ModelParameterVector >	_param_vec

unique_ptr< MAAGBuilder >	_builder

unique_ptr< ClonotypeAssembler >	_generator

seq_len_t	_min_D_len

Constructor & Destructor Documentation

ymir::ProbabilisticAssemblingModel::ProbabilisticAssemblingModel	(	const string &	folderpath,
		ModelBehaviour	behav = `PREDEFINED`
	)

inline

Model constructor from the given folder with a model's parameters.

Parameters

folderpath Path to a folder with a model's parameters.

behav Model's behaviour. PREDEFINED means that each probability will stay constant for each input data unless the model probabilities vector updated with new parameters (either all parameters or gene usage only). EMPTY means that the model will load only gene segments sequences and generate vector of uniformly distributed marginal probabilities. This model could be used as an input model for the statistical inference of the model's parameters.

Member Function Documentation

MAAGRepertoire ymir::ProbabilisticAssemblingModel::buildGraphs	(	const ClonesetView &	repertoire,
		MetadataMode	save_metadata,
		ErrorMode	error_mode,
		SequenceType	sequence_type = `NUCLEOTIDE`,
		bool	verbose = `true`
	)		const

inline

Build a set of MAAGs from the given cloneset.

Parameters

repertoire	Cloneset.
save_metadatae	If true than build full MAAG (with additional matrices for event indices).
aminoacid	If true than build graphs of aminoacid CDR3 sequences, traslating the given repertoire first.

Returns: Set of MAAGs.

vector<prob_t> ymir::ProbabilisticAssemblingModel::computeFullProbabilities	(	const ClonesetView &	repertoire,
		ErrorMode	error_mode,
		SequenceType	sequence_type = `NUCLEOTIDE`,
		MAAGComputeProbAction	action = `SUM_PROBABILITY`
	)		const

inline

Compute full assembling probabilities of clonotypes from the given repertoire. If some clonotypes have few alignments of some gene segments, than full probability of all combinations of gene segments will be computed and either summed or the max full probability will be chosen.

Parameters

repertoire	Repertoire with alignments and event indices.
aminoacid	What probabilities will be computed - of nucleotide or amino acid sequences. If this parameter is false and some clone will have amino acid sequence, than error message will be generated to the output and probability of this clone will set to zero.

Returns: Vector of full assembling probabilities.

const VDJRecombinationGenes& ymir::ProbabilisticAssemblingModel::gene_segments ( ) const

inline

Access to the gene segments table.

Returns: Const reference to the gene segments table.

Cloneset ymir::ProbabilisticAssemblingModel::generateSequences	(	size_t	count = `1`,
		bool	verbose = `true`
	)		const

inline

Generate artificial repertoire of sequences using this model.

Parameters

count	Number of generated sequences.
merge	If true than return only unique clonotypes with counts of occurrences for each clonotype.

Returns: Artificial repertoire.

bool ymir::ProbabilisticAssemblingModel::save ( const string & folderpath ) const

inline

Save the model to the given folder.

Parameters

folderpath Path to the model folder.

Returns: True if all is ok.

bool ymir::ProbabilisticAssemblingModel::status ( ) const

inline

Return if the model is in the ready-to-work state.

Returns: True or false.

void ymir::ProbabilisticAssemblingModel::updateModelParameterVector ( const ModelParameterVector & vec )

inline

Set a new event probabilities vector to this model.

Parameters

vec	Vector with new event probabilities.

The documentation for this class was generated from the following file:

Ymir/src/probabilisticassemblingmodel.h

Public Member Functions

Protected Member Functions

Protected Attributes

Constructor & Destructor Documentation

Member Function Documentation