methylseg.MethylSegHMM

class methylseg.MethylSegHMM(n_states)

Bases: object

Abstract base class for methylseg HMM backends.

Parameters:

n_states (int)

__init__(n_states)

Initialize an abstract HMM backend.

Parameters:

n_states (int) – Number of hidden methylation states the backend will model.

Methods

__init__(n_states)	Initialize an abstract HMM backend.
create_model()	Instantiate and initialize the backend-specific HMM object.
fit(emissions[, sample_info, chrom])	Fit backend-specific HMM parameters on the provided emissions.
format_fit(emissions)	Convert raw emissions into the representation expected by fit.
format_predict(emissions)	Convert raw emissions into the representation expected by predict.
predict(emissions)	Decode hidden states for the supplied emissions.

fit(emissions, sample_info=None, chrom=None)

Fit backend-specific HMM parameters on the provided emissions.

Parameters:

• emissions – Observation sequence or feature matrix already prepared for the backend.

• sample_info – Optional sample metadata used by backends that require genomic coordinates or other sample-level context.

• chrom – Chromosome label for single-chromosome fitting when relevant.

create_model()

Instantiate and initialize the backend-specific HMM object.

predict(emissions)

Decode hidden states for the supplied emissions.

Parameters:

emissions – Observation sequence or feature matrix prepared for prediction.

Returns:

Hidden-state assignments in backend-specific numeric form.

Return type:

numpy.ndarray

format_fit(emissions)

Convert raw emissions into the representation expected by fit.

Parameters:

emissions – Raw observation labels or features.

format_predict(emissions)

Convert raw emissions into the representation expected by predict.

Parameters:

emissions – Raw observation labels or features.