load_pcba#

skfp.datasets.moleculenet.load_pcba(data_dir: str | PathLike | None = None, as_frame: bool = False, verbose: bool = False) → DataFrame | tuple[list[str]] | ndarray#

Load and return the PCBA (PubChem BioAssay) dataset.

The task is to predict biological activity against 128 bioassays, generated by high-throughput screening (HTS). All tasks are binary active/non-active.

Note that targets have missing values. Algorithms should be evaluated only on present labels. For training data, you may want to impute them, e.g. with zeros.

Tasks

128

Task type

multitask classification

Total samples

437929

Recommended split

scaffold

Recommended metric

AUPRC, AUROC
Parameters:

  • data_dir ({None, str, path-like}, default=None) – Path to the root data directory. If None, currently set scikit-learn directory is used, by default $HOME/scikit_learn_data.

  • as_frame (bool, default=False) – If True, returns the raw DataFrame with columns “SMILES” and 128 label columns, with names corresponding to biological activities (see [1] and [2] for details). Otherwise, returns SMILES as list of strings, and labels as a NumPy array. Labels are 2D NumPy float array with binary labels and missing values.

  • verbose (bool, default=False) – If True, progress bar will be shown for downloading or loading files.

Returns:

data – Depending on the as_frame argument, one of: - Pandas DataFrame with columns “SMILES” and 128 label columns - tuple of: list of strings (SMILES), NumPy array (labels)

Return type:

pd.DataFrame or tuple(list[str], np.ndarray)

References