Types¶

Core data structures and type definitions used throughout pyprego.

pyprego.types ¶

Core type definitions for pyprego.

This module defines the data structures used throughout the package. PSSM matrices and spatial models are represented as pandas DataFrames to keep things simple, inspectable, and consistent with the R prego package.

NUCLEOTIDES `module-attribute` ¶

NUCLEOTIDES = ('A', 'C', 'G', 'T')

RegressionResult `dataclass` ¶

Container for the output of :func:pyprego.regression.regress_pwm.

ATTRIBUTE	DESCRIPTION
`pssm`	PSSM DataFrame (pos, A, C, G, T) for the inferred motif. TYPE: `DataFrame`
`spat`	Spatial model DataFrame (bin, spat_factor). TYPE: `DataFrame`
`pred`	Predicted PWM score for each input sequence. TYPE: `ndarray`
`consensus`	Consensus sequence derived from the PSSM. TYPE: `str`
`r2`	R-squared of prediction vs response (continuous response). TYPE: `float \| None`
`ks`	KS statistic (binary response). TYPE: `float \| None`
`seed_motif`	The seed motif / kmer that initialised the regression. TYPE: `str \| None`
`bidirect`	Whether the model is bidirectional (uses reverse complement). TYPE: `bool`
`spat_min`	Minimum spatial position used. TYPE: `int`
`spat_max`	Maximum spatial position used. TYPE: `int \| None`
`seq_length`	Length of input sequences. TYPE: `int \| None`
`_predict_fn`	Internal prediction function (set after fitting). TYPE: `Callable \| None`

predict ¶

predict(sequences: list[str] | ndarray) -> np.ndarray

Predict PWM scores for new sequences.

PARAMETER	DESCRIPTION
`sequences`	DNA sequences to score. TYPE: `list[str] \| ndarray`

RETURNS	DESCRIPTION
`ndarray`	Predicted scores, one per sequence.

RAISES	DESCRIPTION
`RuntimeError`	If the model has not been fitted yet.

to_dict ¶

to_dict() -> dict[str, Any]

Serialise the result to a plain dictionary (for YAML/JSON export).

pssm_dataframe ¶

pssm_dataframe(matrix: ndarray) -> pd.DataFrame

Create a PSSM DataFrame from a (L, 4) NumPy array.

PARAMETER	DESCRIPTION
`matrix`	Array of shape (L, 4) with columns ordered A, C, G, T. TYPE: `ndarray`

RETURNS	DESCRIPTION
`DataFrame`	DataFrame with columns `pos`, `A`, `C`, `G`, `T`.

RAISES	DESCRIPTION
`ValueError`	If matrix does not have exactly 4 columns.

pssm_to_array ¶

pssm_to_array(pssm: DataFrame) -> np.ndarray

Extract the (L, 4) NumPy array from a PSSM DataFrame.

PARAMETER	DESCRIPTION
`pssm`	PSSM DataFrame with at least columns `A`, `C`, `G`, `T`. TYPE: `DataFrame`

RETURNS	DESCRIPTION
`ndarray`	Array of shape (L, 4).

spatial_dataframe ¶

spatial_dataframe(bins: ndarray, factors: ndarray) -> pd.DataFrame

Create a spatial model DataFrame.

PARAMETER	DESCRIPTION
`bins`	1-D array of bin start positions. TYPE: `ndarray`
`factors`	1-D array of spatial factors (same length as bins). TYPE: `ndarray`

RETURNS	DESCRIPTION
`DataFrame`	DataFrame with columns `bin` and `spat_factor`.