FAF4LeadlikeFilter#

class skfp.filters.FAF4LeadlikeFilter(allow_one_violation: bool = False, return_indicators: bool = False, n_jobs: int | None = None, batch_size: int | None = None, verbose: int = 0)#

FAFDrugs4 Lead-Like Soft filter.

Designed as a part of FAFDrugs4 software [1] [2]. Based on literature describing physico-chemical properties of lead drugs. Designed to keep starting point molecules that can be further optimized, i.e. relatively small, with low logP, and that can be “decorated” further to increase affinity and or selectivity without becoming very ADMET unfriendly).

Basically a more restrictive variant of FAFDrugs4 Drug-Like Soft filter.

Molecule must fulfill conditions:

  • molecular weight in range [150, 400]

  • logP in range [-3, 4]

  • HBA <= 7

  • HBD <= 4

  • TPSA <= 160

  • number of rotatable bonds <= 9

  • number of rigid bonds <= 30

  • number of rings <= 4

  • max ring size <= 18

  • number of carbons in range [3, 35]

  • number of heteroatoms in range [1, 15]

  • non-carbons to carbons ratio in range [0.1, 1.1]

  • number of charged functional groups <= 4

  • total formal charge in range [-4, 4]

  • number of stereocenters <= 2

Note that the FAF4Drugs uses ChemAxon for determining functional groups. We use their publicly available CXSMARTS list of functional groups [3]. Phosphine and sulfoxide patterns could not be parsed by RDKit, so we manually fixed them.

Parameters:

  • allow_one_violation (bool, default=False) – Whether to allow violating one of the rules for a molecule. This makes the filter less restrictive.

  • return_indicators (bool, default=False) – Whether to return a binary vector with indicators which molecules pass the filter, instead of list of molecules.

  • n_jobs (int, default=None) – The number of jobs to run in parallel. transform_x_y() and transform() are parallelized over the input molecules. None means 1 unless in a joblib.parallel_backend context. -1 means using all processors. See Scikit-learn documentation on n_jobs for more details.

  • batch_size (int, default=None) – Number of inputs processed in each batch. None divides input data into equal-sized parts, as many as n_jobs.

  • verbose (int, default=0) – Controls the verbosity when filtering molecules.

References

Examples

>>> from skfp.filters import FAF4LeadlikeFilter
>>> smiles = ["C", "CC(=O)Nc1ccc(O)cc1"]
>>> filt = FAF4LeadlikeFilter()
>>> filt
FAF4LeadlikeFilter()

>>> filtered_mols = filt.transform(smiles)
>>> filtered_mols
['CC(=O)Nc1ccc(O)cc1']

Methods

fit(X[, y])

Unused, kept for Scikit-learn compatibility.

fit_transform(X[, y])

The same as .transform() method, kept for Scikit-learn compatibility.

get_metadata_routing()

Get metadata routing of this object.

get_params([deep])

Get parameters for this estimator.

set_output(*[, transform])

Set output container.

set_params(**params)

Set the parameters of this estimator.

set_transform_request(*[, copy])

Request metadata passed to the transform method.

transform(X[, copy])

Apply a filter to input molecules.

transform_x_y(X, y[, copy])

Apply a filter to input molecules.
fit(X: Sequence[str | Mol], y: ndarray | None = None, **fit_params)#

Unused, kept for Scikit-learn compatibility.

Parameters:

  • X (any) – Unused, kept for Scikit-learn compatibility.

  • y (any) – Unused, kept for Scikit-learn compatibility.

  • **fit_params (dict) – Unused, kept for Scikit-learn compatibility.

Return type:

self

fit_transform(X: Sequence[str | Mol], y: ndarray | None = None, **fit_params)#

The same as .transform() method, kept for Scikit-learn compatibility.

Parameters:

  • X (any) – See .transform() method.

  • y (any) – See .transform() method.

  • **fit_params (dict) – Unused, kept for Scikit-learn compatibility.

Returns:

X_new – See .transform() method.

Return type:

any

get_metadata_routing()#

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns:

routing – A MetadataRequest encapsulating routing information.

Return type:

MetadataRequest

get_params(deep=True)#

Get parameters for this estimator.

Parameters:

deep (bool, default=True) – If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:

params – Parameter names mapped to their values.

Return type:

dict

set_output(*, transform=None)#

Set output container.

See Introducing the set_output API for an example on how to use the API.

Parameters:

transform ({"default", "pandas", "polars"}, default=None) –

Configure output of transform and fit_transform.

  • ”default”: Default output format of a transformer

  • ”pandas”: DataFrame output

  • ”polars”: Polars output

  • None: Transform configuration is unchanged

Added in version 1.4: “polars” option was added.

Returns:

self – Estimator instance.

Return type:

estimator instance

set_params(**params)#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:

**params (dict) – Estimator parameters.

Returns:

self – Estimator instance.

Return type:

estimator instance

set_transform_request(*, copy: bool | None | str = '$UNCHANGED$') → FAF4LeadlikeFilter#

Request metadata passed to the transform method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to transform if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to transform.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

copy (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for copy parameter in transform.

Returns:

self – The updated object.

Return type:

object

transform(X: Sequence[str | Mol], copy: bool = False) → list[str | Mol] | ndarray#

Apply a filter to input molecules. Output depends on return_indicators attribute.

Parameters:

  • X ({sequence, array-like} of shape (n_samples,)) – Sequence containing RDKit Mol objects.

  • copy (bool, default=False) – Copy the input X or not.

Returns:

X – List with filtered molecules, or indicator vector which molecules fulfill the filter rules.

Return type:

list of shape (n_samples_conf_gen,) or array of shape (n_samples,)

transform_x_y(X: Sequence[str | Mol], y: ndarray, copy: bool = False) → tuple[list[str | Mol], ndarray] | tuple[ndarray, ndarray]#

Apply a filter to input molecules. Output depends on return_indicators attribute.

Parameters:

  • X ({sequence, array-like} of shape (n_samples,)) – Sequence containing RDKit Mol objects.

  • y (array-like of shape (n_samples,)) – Array with labels for molecules.

  • copy (bool, default=False) – Copy the input X or not.

Returns:

  • X (list of shape (n_samples_conf_gen,) or array of shape (n_samples,)) – List with filtered molecules, or indicator vector which molecules fulfill the filter rules.

  • y (np.ndarray of shape (n_samples_conf_gen,)) – Array with labels for molecules.