drugforge.data.util.utils.parse_fluorescence_data_cdd

drugforge.data.util.utils.parse_fluorescence_data_cdd(mol_df, keep_best_per_mol=True, assay_name='ProteaseAssay_Fluorescence_Dose-Response_Weizmann', dG_T=298.0, cp_values=None, pic50_stderr_filt=10.0)[source]
Filter a dataframe of molecules to retain those specified. Required columns are:

  • “name”

  • “assay_name: IC50 (µM)”

  • “assay_name: IC50 CI (Lower) (µM)”

  • “assay_name: IC50 CI (Upper) (µM)”

  • “assay_name: Curve class”

Columns that are added to the dataframe by this function:

  • “IC50 (M)”

  • “IC50_stderr (M)”

  • “IC50_95ci_lower (M)”

  • “IC50_95ci_upper (M)”

  • “pIC50”

  • “pIC50_stderr”

  • “pIC50_range”

  • “pIC50_95ci_lower”

  • “pIC50_95ci_upper”

  • “exp_binding_affinity_kcal_mol”

  • “exp_binding_affinity_kcal_mol_stderr”

  • “exp_binding_affinity_kcal_mol_95ci_lower”

  • “exp_binding_affinity_kcal_mol_95ci_upper”

  • “exp_binding_affinity_kT”

  • “exp_binding_affinity_kT_stderr”

  • “exp_binding_affinity_kT_95ci_lower”

  • “exp_binding_affinity_kT_95ci_upper”

Parameters:

  • mol_df (pandas.DataFrame) – DataFrame containing compound information

  • keep_best_per_mol (bool, default=True) – Keep only the best measurement for each molecule (first sorting by curve class and then 95% CI pIC50 width)

  • assay_name (str, default=”ProteaseAssay_Fluorescence_Dose-Response_Weizmann”) – Name of the assay of interest

  • dG_T (float, default=298.0) – Temperature in Kelvin for converting pIC50 values to delta G values

  • cp_values (Tuple[int], optional) – Substrate concentration and Km values for calculating Ki using the Cheng-Prussoff equation. These values are assumed to be in the same concentration units. If no values are passed for this, pIC50 values will be used as an approximation of the Ki

  • pic50_stderr_filt (float, default=10.0) – Max allowable standard error in pIC50 units. Overly large errors lead to rounded values that don’t make sense, so set anything that will cause issues to NaN

Returns:

DataFrame containing parsed fluorescence and binding affinity values

Return type:

pandas.DataFrame