smftools.tools.position_stats

smftools.tools.position_stats#

Functions

calculate_relative_risk_on_activity(adata, sites)

Perform methylation vs.

compute_positionwise_statistics(adata, layer)

Compute per-(sample, ref) positionwise matrices for selected methods.

plot_positionwise_matrices(adata, methods[, ...])

Plot grids of matrices for each method with pagination and rotated sample-row labels.

random_fill_nans(X)

Fill NaNs with random values in-place.

tqdm_joblib(tqdm_object)

Context manager to patch joblib to update a tqdm progress bar.
smftools.tools.position_stats.random_fill_nans(X)#

Fill NaNs with random values in-place.

Parameters:

X (ndarray) -- Input array with NaNs.

Returns:

Array with NaNs replaced by random values.

Return type:

numpy.ndarray

smftools.tools.position_stats.calculate_relative_risk_on_activity(adata, sites, alpha=0.05, groupby=None, index_col_suffix=None)#

Perform methylation vs. activity analysis within each group.

Parameters:

  • adata (AnnData) -- Annotated data matrix.

  • sites (Sequence[str]) -- Site keys (e.g., ["GpC_site", "CpG_site"]).

  • alpha (float (default: 0.05)) -- FDR threshold for significance.

  • groupby (Union[str, Sequence[str], None] (default: None)) -- Obs column(s) to group by.

  • index_col_suffix (str | None (default: None)) -- If set, use adata.var[f"{ref}_{index_col_suffix}"] for Genomic_Position instead of var_names (e.g. "reindexed"), so reindexing_offsets/reindexing_invert are reflected in the reported position.

Returns:

Mapping of reference -> group label -> (results_df, sig_df).

Return type:

dict

smftools.tools.position_stats.tqdm_joblib(tqdm_object)#

Context manager to patch joblib to update a tqdm progress bar.

smftools.tools.position_stats.compute_positionwise_statistics(adata, layer, methods=('pearson',), sample_col='Barcode', ref_col='Reference_strand', site_types=None, encoding='signed', output_key='positionwise_result', min_count_for_pairwise=10, max_threads=None, reverse_indices_on_store=False, min_position_valid_fraction=None, index_col_suffix=None)#

Compute per-(sample, ref) positionwise matrices for selected methods.

Parameters:

  • adata (AnnData) -- AnnData object to analyze.

  • layer (str) -- Layer name to use for statistics.

  • methods (Sequence[str] (default: ('pearson',))) -- Methods to compute (e.g., "pearson").

  • sample_col (str (default: 'Barcode')) -- Obs column containing sample identifiers.

  • ref_col (str (default: 'Reference_strand')) -- Obs column containing reference identifiers.

  • site_types (Optional[Sequence[str]] (default: None)) -- Optional site types to subset positions.

  • encoding (str (default: 'signed')) -- "signed" or "binary" encoding.

  • output_key (str (default: 'positionwise_result')) -- Key prefix for results stored in adata.uns.

  • min_count_for_pairwise (int (default: 10)) -- Minimum counts for pairwise comparisons.

  • max_threads (Optional[int] (default: None)) -- Maximum number of threads.

  • reverse_indices_on_store (bool (default: False)) -- Whether to reverse indices on output storage.

  • min_position_valid_fraction (Optional[float] (default: None)) -- If set, exclude positions where the {ref}_valid_fraction var column is below this threshold (same filtering as spatial clustermaps).

  • index_col_suffix (Optional[str] (default: None)) -- If set, use adata.var[f"{ref}_{index_col_suffix}"] for DataFrame labels instead of var_names (e.g. "reindexed").

Return type:

None

smftools.tools.position_stats.plot_positionwise_matrices(adata, methods, cmaps=None, sample_col='Barcode', ref_col='Reference_strand', output_dir=None, vmin=None, vmax=None, figsize_per_cell=(3.5, 3.5), dpi=160, cbar_shrink=0.9, output_key='positionwise_result', show_colorbar=True, flip_display_axes=False, rows_per_page=6, sample_label_rotation=90.0, n_ticks=10, tick_fontsize=7, tick_rotation=90.0)#

Plot grids of matrices for each method with pagination and rotated sample-row labels.

Parameters:

  • rows_per_page (-) -- how many sample rows per page/figure (pagination)

  • sample_label_rotation (-) -- rotation angle (deg) for the sample labels placed in the left margin.

Returns:

dict mapping method -> list of saved filenames (empty list if figures were shown).