pydruglogics.model.Statistics

Provides statistical and evaluation functions for analyzing the performance of Boolean model predictions. Supports PR curve calculation, confidence intervals, bootstrapping, and comparison between model runs.

Statistics.sampling_with_ci(observed_synergy_scores, model_outputs, perturbations, synergy_method='bliss', repeat_time=10, sub_ratio=0.8, boot_n=1000, confidence_level=0.9, plot=True, plot_discrete=False, save_result=True, with_seeds=True, seeds=42)

Performs sampling with confidence interval calculation and plots the PR curve.

Parameters

  • boolean_models (list): BooleanModel instances.

  • observed_synergy_scores (list): Observed synergy scores.

  • model_outputs: Model outputs.

  • perturbations (list): Perturbations to apply.

  • synergy_method (str, optional): ‘hsa’ or ‘bliss’.

  • repeat_time (int, optional): Repeats. Default: 10.

  • sub_ratio (float, optional): Proportion to sample. Default: 0.8.

  • boot_n (int, optional): Bootstrap samples. Default: 1000.

  • confidence_level (float, optional): CI level. Default: 0.9.

  • plot (bool, optional): Plot the PR curve. Default: True.

  • plot_discrete (bool, optional): Discrete points. Default: False.

  • save_result (bool, optional): Save to .tab file. Default: True.

  • with_seeds (bool, optional): Use fixed seed. Default: True.

  • seeds (int, optional): Random seed. Default: 42.

Statistics.compare_two_simulations(boolean_models2, observed_synergy_scores, model_outputs, perturbations, synergy_method='bliss', label1='Models 1', label2='Models 2', normalized=True, plot=True, save_result=True)

Compares ROC and PR curves for two sets of evolution results.

Parameters

  • boolean_models1 (list): First set.

  • boolean_models2 (list): Second set.

  • observed_synergy_scores (list): Observed scores.

  • model_outputs: Model outputs.

  • perturbations (list): Perturbations.

  • synergy_method (str, optional): ‘hsa’ or ‘bliss’.

  • label1 (str, optional): Label for first.

  • label2 (str, optional): Label for second.

  • normalized (bool, optional): Normalize first. Default: True.

  • plot (bool, optional): Show ROC/PR. Default: True.

  • save_result (bool, optional): Save results. Default: True.