corneto.methods.compute_phonemes_scores#

corneto.methods.compute_phonemes_scores(pvalues, *, fold_changes=None, pvalue_threshold=0.05, fold_change_threshold=None, direction='both', scale=True)#

Compute PHONEMeS node scores from p-values and fold changes.

Flat mappings, pandas Series, and one-dimensional arrays represent one condition. Nested mappings, pandas DataFrames, and two-dimensional arrays represent sites by conditions. The returned object preserves the input container type for NumPy and pandas inputs; mappings return dictionaries.

Negative scores reward measurements satisfying the selected evidence criteria, while positive scores penalize unsupported measurements.

Parameters:

  • pvalues – P-values indexed or ordered by phosphosite.

  • fold_changes – Optional fold changes with the same structure and labels.

  • pvalue_threshold (float) – Significance threshold in (0, 1].

  • fold_change_threshold (float | None) – Optional nonnegative absolute fold-change threshold.

  • direction (str) – Fold-change direction: "both", "up", or "down".

  • scale (bool) – Scale positive and negative scores separately per condition.

Returns:

Scores with the same labels and container type as pvalues.